Thinking for a Living: How to Get Better Performances And Results from Knowledge Workers [1 ed.] 1591394236, 9781591394235

Citation preview

Library

AW 420

Spears Research Center N.

English Street

Greensboro,

NC 27405

Digitized by the Internet Archive in

2016 with funding from

Kahle/Austin Foundation

https://archive.org/details/thinkingforlivinOOdave

Thinking for a Living

"ujiuson), uJhat

cxAcru-r is a KhJowi-Eo&s worker apJp oo U)E

HAVE:

© 2005 Charles Barsotti from

AW

o/0 "THE

STAFF?"

cartoonbank.com. All Rights Reserved.

HD 8039 .K59

D38 2005

How to Get

Better

Performance and Results from

Knowledge Workers

v€lw

Thomas

H.

0enter

ubr2ry

Davenport

Harvard Business School Press Boston, Massachusetts

Copyright 2005 Thomas H. Davenport All rights reserved

Printed in the United States of America

09 08 07 06 05

5 4 3 2

1

No part of this publication may be

reproduced, stored in or introduced into a retrieval

system, or transmitted, in any form, or by any

means

(electronic, mechanical,

photo-

copying, recording, or otherwise), without the prior permission of the publisher.

Requests for permission should be directed to [email protected], or

mailed to Permissions, Harvard Business School Publishing, 60 Harvard Way, Boston, Massachusetts 02163.

Library of Congress Cataloging-in-Publication Data

Davenport,

Thomas

H.,

1954-

Thinking for a living workers p.

/

:

how

to get better

performance and

results

from knowledge

Thomas H. Davenport, cm.

Includes index.

ISBN 1-59139-423-6 1.

Knowledge workers.

2.

Knowledge management.

3. Intellectual capital.

I.

Title.

HD8039.K59D38 2005 658.3—dc22 2005002390

The paper used

in this publication

meets the

National Standard for Information Sciences Materials,

ANSI Z39.48-1992.

minimum

requirements of the American

—Permanence of Paper

for Printed Library

Contents

Preface

and Acknowledgments

1

What’s a Knowledge Worker, Anyway?

2

How Knowledge Workers Differ, and the Difference

3

It

1

25

Makes

Interventions, Measures, in

vii

and Experiments

39

Knowledge Work

4

Knowledge Work Processes

61

5

Organizational Technology for Knowledge Workers

85

6

Developing Individual Knowledge Worker Capabilities

7

Investing in

8

The

Knowledge Workers’ Networks and Learning

Physical

Work Environment and

1 1

141

165

Knowledge Worker Performance 9

Managing Knowledge Workers

187

Notes

211

Index

217

About

the Author

227

and Acknowledgments

Preface

I’ve

been planning

this

intellectual interests

areas in

which

1980s

late

I

I’ve

book

for a long time,

and background.

done

By 1993 or so

I

it

goes deeply into

at the intersection

a lot of work: processes

started to research the

ness processes.

It’s

and

of two

and knowledge. In the

improvement and management of busi-

had

knowledge-

realized that to address

intensive processes required a different set of approaches than

companies were using Then, around 1994,

management. For

for process reengineering

started researching

1

several years

I

and writing about knowledge

argued that knowledge management

attention to the process side of this

Because

I

knew that I would

a series of studies involving

combo

knowledge workers and

I’d like to

didn’t pay

I

until the late 1990s.

eventually write this book,

hoped would become key building blocks or involved collaborators, and

most

and improvement.

ought to involve “people, process, and technology,” but

much

my

their

chapters.

thank them

I

now

undertook

work

that

Most of

I

these

for their help

with these research projects and the ideas that resulted from them. The first

was undertaken with Sirkka Jarvenpaa of the University of Texas

and Mike

Beers, then of Ernst

Management Review cesses,” I

article called

Young.

It

resulted in a 1996 Sloan

“Improving Knowledge Work Pro-

and many of its concepts can be found

became the Director of the Accenture

Change that

&

in 1998,

made

their

and

way

I

worked with

into this

book

in chapters 2

4.

Institute for Strategic

several colleagues there in

and

on

projects

some form. The matrix of four

Vlll

Preface

and Acknowledgments

knowledge work types that edge management that

I

use here

worked on

I

Donoghue. Then Bob Thomas, Sue project

we

of this book. Sue and

on

is

on knowl-

a bit with Jeanne Harris

and

Cantrell,

I

also

It

led to

I

and Leigh

collaborated

on

most of the ideas

in chapter 8

this research

networks of high-performing

formed the

basis of chapter 7.

the result of an attempt at coauthoring an article with

Chap-

Warren

Bennis on the management and leadership of knowledge workers. never really emerged as a coauthored piece, but ideas

Warren and I had discussed

celebration

While

and

still

at

a

worked with Rob Cross of the University of

a project assessing the social

knowledge workers; ter 9

a project

“The Art of Work: Improving the Performance of

called

High-End Knowledge Workers.”

Virginia

came from

tium of IT firms

I

was able

Hewlett-Packard,

working with

Work

started the ball rolling,

Intel,

1

a consor-

Productivity Council

and Accenture, Cisco,

SAP, and Xerox also got behind

Accenture for Babson in 2003,

to use the

volume.

to begin

called the Information

(IWPC). Microsoft

was happy

as the basis for a talk at his Festschrift

a chapter in his Festschrift

Accenture,

I

It

it.

When

I

left

became the Academic Director of this

consortium, and also led a particular research project on “Personal Information and Knowledge Management” that became the core of chapter 6.

On the IWPC project worked closely with Susan Conway of MicroI

soft,

Dan Holtshouse of Xerox, Mary Lee Kennedy

(then of Microsoft),

and Carla O’Dell of the American Productivity and Quality Center. Since to this

cess

coming

to Babson, I’ve

book within

the

two programs, so

managers who joined

a couple of projects related

Working Knowledge Research Center and the Pro-

Management Research

these

worked on

I

me

Center. There are about forty sponsors of

can’t

name them

in exploring issues

all.

But I’m grateful to the

around “High-Performing

Knowledge Worker Environments” and “Managing Knowledge-Intensive Processes.”

Brad Power, the Executive Director of the Process Manage-

ment program, gave me much

useful feedback

on

my ideas, and

Larry

IX

Preface

and Acknowledgments

Prusak, codirector of the Working Knowledge program, has been a

source of insights, inspirations, and lively gossip for I

and

make my living writing the book

this

is

no exception.

donated their experiences for are

many

Adams I’ve

at

stories of

BT, and Luke

had

more than

a great relationship

a decade,

and

thank

me to package up

I still

all

are

Health Care, Kevin

both typical and exemplary.

with Harvard Business School Press for think they are the classiest act

business publishers. Melinda Merino was a great editor to stimulate

more and

formal comments.

mous

I

who

the managers

into these pages. There

at Partners

Koons of Intel

a decade.

adventurous businesspeople,

I’d like to

of them, but John Glaser

more than

who

among

never failed

better authorial thinking with her informal

also got very helpful

reviewers, although

I

house and Sue Cantrell, so

and

comments from four anony-

deduced that two of them were Dan HoltsI

can thank them by name.

I’d also like to

thank Hollis Heimbouch for long-lasting support, Monica Jainschigg for skilled

copy editing, Marcy Barnes-Henrie for seeing

this

book

through production, and Zeenat Potia and her team for an appealing cover design.

My

reader, friend,

without

her.

and

lover.

I

someday

book

My wonderful sons Hayes and Chase remain largely obliv-

they’ll

I

write

them

write.

in part

because

my favorite

be proud of them. Bucky,

died during the creation of this book, and I

as agent, adviser,

couldn’t have written this or any other

ious to their dad’s books, but

feet as

my work

wife Jodi continues to help with

I

will

I

hope

dog of

miss having

all

him

that

time, at

my

t

1

What’s a Knowledge Worker, Anyway?

Robin

out of control. She’s the head of press and analyst relations

feels

for a large professional services firm. Eight people report to her,

them knowledge workers. Some

Her job

these external audiences

is

—even

of

deal with particular types of press re-

lationships, others with technology industry analysts.

relationships herself.

all

to

She manages some

improve her company’s image with

to her, a

somewhat vague and

difficult-

to-measure mission.

Only half of the people on her

New

staff are

York, and only one of the four in

They’re

all

pretty independent in their

most of them very

The team used restrictions feels that

often. Several of

with her

New

at

York are on her

work habits, and she

much

floor.

doesn’t see

them frequently work from home.

to try to get together monthly, but the

on

headquarters in

travel budgets, so they can’t

of the time she doesn’t

that

anymore. Robin

exactly

what her people

do

know

company has put

are doing. Sure, she hears about meetings they’re having in their weekly

conference

call,

but what are they doing when they’re not having a few

— 2

Thinking for a Living

meetings? She thinks she trusts them, but always has nagging doubts

about

how hard and how effectively they’re working. some measures

She’s put

such as numbers of

in place for her group,

media mentions or favorable ratings

But when

in analysts’ reports.

things don’t go well, her people always have a ready explanation

“One of our

strongest thought leaders

left

the practice,” or “That ana-

has always had a grudge against us.” But

lyst

how else can

she assess the

And

strength of such intangible qualities as “relationships” or “image”? in

many

people are right

cases, her

—they have

closer relationships

with the reporters, editors, and analysts they deal with than Robin does. She

knows

the business well from having

done

it

but she

herself,

day and

also realizes that the job varies a lot

from day

pendent on interpersonal

a result, she’s reluctant to criticize

what seems

like

factors.

As

poor performance too

to

directly; she can’t afford to lose

the relationships and insights her people have built

Some

highly de-

is

up over

time.

of her analysts are definitely more productive than others,

however. One, for example, generates twice the press coverage and

meets with twice as

Other analysts

many industry analysts as anyone else

in the

it’s

something

— “The than mine do” — but Robin

group discount her performance

lines she represents just

thinks

in the group.

have more to say

this analyst does.

Robin has

service

tried to figure out the

secret to her success, but has never directly discussed

it

with her. She’s

afraid that the high-performing analyst will ask for a raise

she ad-

if

dresses her sterling performance too directly.

Robin

is

always trying to think of ways to get

members, but

rarely convinced she’s

is

week’s conference to

call,

communicate her

analysts. She’s sure that

mind her through told

is

that she

vocal team

already overloaded;

In fact, the situation

by her boss

announce an

new marketing campaign

some of her more

that “the channel

it.”

found the answer.

for example, she plans to

firm’s

more out of her team

may

is

we

On

this

initiative

to the press

members can’t

and

will re-

push more

worse than they suspect;

she’s

been

have to lay off one employee due to a

3

What’s a Knowledge Worker, Anyway?

budget cut for next

year.

Everyone

will

have to do more with

everyone needs to be more productive. But Robin finds

imagine

how

and who know

sees,

she decides to just ship with

—

more

she can get

it’s

results

from people

less;

difficult to

it

whom

she rarely

their jobs better than she does. In the short run,

call a

reporter she’s been trying to build a relation-

do her own knowledge work than

easier to

to

improve

that of others.

Like Robin, you and

—and most of our

We

knowledge workers.

are

us

I

all

friends

and colleagues

think for a living. Like Robin,

manage other knowledge workers. We’re

all

many

of

doing our work the best

we can

—or

results

from knowledge workers? Most of us have never even analyzed

our us

Can we,

are we?

like

Robin, figure out a way to get better

own performance, or had much

more productive and

doing our jobs

—and

effective.

help from our employers in making

We want to become more efficient at

to help others

do so

as well

—but we

know how. We know more about our own work than one in

else,

so

it’s

like to

we

fact that

are

—and

be told what to do. We’ve never thought

knowledge workers, or about the implica-

how we carry out and improve our daily activities.

tions of that fact for

What difference does it make that we are knowledge workers? not a

about any-

hard for a manager to improve our performance

any case we don’t

about the

just

just don’t

It’s

certainly

—think — why

new thing. This category of work has existed for centuries

about medieval monks, or the write about

nothing

workers

them now?

else they’re

—probably

Well, as

first I’ll

professors at universities

argue a

little later

so

in this chapter, if

important because they are a large category of

larger than ever before as a percentage of the

work-

force in sophisticated economies.

But regardless of the

size

of this category of workers,

it’s

a very

im-

portant one. Knowledge workers are responsible for sparking innova-

your organization. They invent your new products

tion

and growth

and

services, design

gies. In

in

your marketing programs, and create your

the current economy, they are the horses that pull the

strate-

plow of

4

Thinking for a Living

economic progress.

If

our companies are going to be more profitable,

our strategies are going to be successful,

come more advanced work

in a

—

it

will

if

our society

is

if

going to be-

be because knowledge workers did their

more productive and

effective

manner.

The Growing Importance of Knowledge Workers Of

course, the rise of knowledge

years. Automation in factories

work has been foreseen

and farms more than

most of the workforce from having

freed

Over the

last half-century,

on

take action

it.

a century ago

perform physical

labor.

the advent of computers and the pervasive

presence of information created a

duce the information

to

many

for

demand

for

workers

in the first place, extract

The economist

Fritz

who

could pro-

meaning from

Machlup did much of the

spadework on knowledge and knowledge work

and

it,

roles; as early as

early

1958

he stated that knowledge workers comprised almost a third of the U.S. workforce, and that the knowledge work sector was growing twice as fast as

the rest of the economy.

2

In the early twenty-first century,

likely that a quarter to a half

it’s

of

workers in advanced economies are knowledge workers whose pri-

mary (see

knowledge and information

tasks involve the manipulation of

“How Many

Are There?”). Even

if

they’re not a majority of

workers, they have the most influence on their economies.

They

all

are

paid the most, they add the most economic value, and they are the greatest determinant of the a high proportion of

intensive States

—

worth of their companies. Companies with

knowledge workers

are the fastest-growing

—

let’s call

and most

them knowledge-

successful in the United

and other leading economies, and have generated most of these

economies’ growth in the past couple of decades. The market value of

many

knowledge-intensive companies

—which includes

the market’s

perception of the value of knowledge and knowledge workers their

book

values,

which includes only tangible

assets

—dwarfs

(and the ratio of

5

What’s a Knowledge Worker, Anyway?

market

to

book value

in U.S.

companies has doubled over the past

twenty years, suggesting a great acceleration of knowledge asset value).

Even

in so-called “industrial”

companies, knowledge

used to differentiate physical goods and to product-related services. As James Brian

is

increasingly

fuel diversification into

Quinn has pointed

out, high

proportions of workers in manufacturing firms (roughly 90 percent in

semiconductors, for example) never touch the manufacturing process,

but instead provide knowledge-based services such as marketing, tribution, or

customer

service.

dis-

3

already apparent that the firms with the highest degree and

It’s

quality of

knowledge work tend

profitable. Microsoft, for example,

to be the fastest-growing is

and most

one of the most profitable organi-

zations in the history of the planet. Pharmaceutical firms not only pro-

duce sophisticated and life-saving drug treatments, they also tend to have high profit margins. Growth industries generally tend to be those with a high proportion of knowledge workers.

How Many Are There?

G just

iven

the imprecise

definitions in the world of

knowledge work and knowledge workers,

how many there

knowledge workers

are in

in

it’s

impossible to specify

any particular country. But there are many

advanced economies such

as the

United States

minimum they comprise a quarter of the U.S. workforce, and at maximum about half. and Europe, no matter how they

The

U.S.

Bureau of Labor

workers, but

what I

it

are defined. At a

Statistics

(BLS) doesn’t

classify

knowledge

puts U.S. workers into categories that can be (some-

arbitrarily, to

be sure) defined as either knowledge workers or not.

would put the following categories

into the

knowledge worker camp:

6

Thinking for a Living

•

Management

•

Business and financial operations

•

Computer and mathematical

•

Architecture and engineering

•

Life, physical,

•

Legal

•

Healthcare practitioners

•

Community and

•

Education, training, and library

•

Arts, design, entertainment, sports,

The in the

and

social scientists

social services

media

above yields about 36 million knowledge workers

classification

United States alone, or 28 percent of the labor

classification

scheme

force.

While no

perfect (for example, professional athletes are

is

included in the knowledge worker group, because the U.S. govern-

ment

data lumps

workers),

it’s

them

clear that

Using somewhat

in with arts, design, entertainment,

most people

less

and media

in these jobs think for a living.

conservative classification criteria than

(they include clerical workers, for example), Rubin and

mine

Huber con-

cluded that there were about 45 million knowledge workers in 1980.

Perhaps the simplest measure fessional,

is

the

a

BLS category of “managerial, pro-

and technical” workers, which was about 34 percent of the

workforce in 2003. Another approach to determining the number of

knowledge workers focuses on the percentage of the workforce that actually possesses the skills to

do knowledge work.

A

U.S.

Department

of Education report suggests that about 25 percent of U.S. workers

have the ability to process complex or moderately complex information in mathematical or verbal form,

Using yet another

set

(every approach has

its

and turn

it

into knowledge.

of criteria to define “information workers” peculiarities

—

in this case, for

example, the

category includes 50 percent of “hucksters”), Marc Porat calculated in

1977 that these workers comprised about half of the workforce, and that their total at

compensation passed that of non-information workers

about that time.

b

7

What’s a Knowledge Worker, Anyway?

Countries other than the United States, of course, have different definitions of

knowledge workers and

different

numbers, although

they are in the same ballpark. Statistics Canada, for example, defines

knowledge workers nical occupations,

as including

management,

and concludes

that these

and tech-

professional,

comprised 25 percent of

the Canadian workforce in 2001 (up from 14 percent in

study of over 28 million jobs in the United

is

some evidence

also

to be important States, for

from

that

a purely

knowledge workers

—

d

will

continue

example, the BLS projects that ten specific occupations will

registered nurses,

engineers

that 32

numerical standpoint. In the United

be the fastest-growing in the current decade. At ten

A

c

97 1 ).

Kingdom found

percent were knowledge-based, requiring a college degree.

There

1

—

computer support

are pretty clearly

least three

specialists,

of these

and IT software

knowledge workers, and they

all fall

into

the top half of hourly earnings in the U.S. economy. This suggests that the quarter-to-a-third proportion of knowledge workers in sophisti-

cated economies will persist or

a.

Michael R. Rubin and Mary

T.

grow over

time.

Huber, The Knowledge Industry

in the

United States

(Princeton, NJ: Princeton University Press, 1986). b.

Marc

Porat,

The Information Economy: Definition and Measurement,

Department of Commerce, 1977), 104-134. Baldwin and Desmond Beckstead, “Knowledge Workers

OT Special Publi-

cation 77-12(1) (U.S. c.

John R.

1971-2001,” Statistics Canada Analytical Paper, catalogue

in

Canada’s Economy,

number 11-624-MIE

—No. 004,

October 2003. d.

The U.K. data

are reported in Phillip

ment of Talent (Oxford: Oxford University

Brown and Anthony Hesketh, The Mismanage-

Press, 2004).

Within organizations, knowledge workers tend

to

be closely aligned

with the organizations growth prospects. Knowledge workers in man-

agement

roles

come up with new strategies. Knowledge workers

and engineering

create

ing package products

new

and

in

R&D

products. Knowledge workers in market-

services in

ways that appeal to customers.

Without knowledge workers there would be no new products and vices,

and no growth.

ser-

8

Thinking for a Living

Knowledge Workers and the World Economy

P

who was

eter Drucker,

the

first

person to describe knowledge

workers to any substantial degree (in his 1959 book Landmarks of

Tomorrow ),

said as far

back

as

1969

that:

To make knowledge work productive

will

be the great management

make manual work productive was

task of this century, just as to

3

the great

management

Then

1997 Drucker went even further out along the knowledge

in

task of the last century.

worker limb:

The productivity of knowledge and knowledge workers the only competitive factor in the world economy. likely to

become

the decisive factor, at least for

the developed countries.

Why

did Drucker

will

It is,

most

not be

however,

industries in

1"

—and why should we—

believe that

knowledge

workers and their productivity were so important to the world econ-

omy? There

are a variety of reasons. First, they are a large

category of workers.

If

we

quarter of the labor force

can’t figure out

more

a

economic

to the

and other groups, they haven’t

We know

little

knowledge workers’ performance, which less

make more than

knowledge workers

success of countries, companies,

no

to

productive, we’re going to have prob-

Yet despite the importance of

ceived sufficient attention.

how

and growing

is

about

how

to

re-

improve

very unfortunate, because

an authority than Peter Drucker has said that improving

knowledge worker performance

is

the

most important economic

issue

of the age (see “Knowledge Workers and the World Economy”). So

chosen to write

this

I’ve

book about how we can make knowledge workers

more productive and

effective at their jobs.

9

What's a Knowledge Worker, Anyway?

lems with our economy overall. Second, they are the most expensive type of worker that organizations employ, so

it’s

doubly shameful

if

they’re not as productive as they could be.

Third, they are key to the growth of

many economies.

Agricultural

and manufacturing work have generally become commoditized, and are

moving

est cost.

to the

economies where they can be performed

The only forms of agricultural or economies are those

in sophisticated

edge has been injected

—

in

industrial

work

that survive

which the

manufactur-

and pesticides

fertilizer

administered to a given crop are carefully monitored using

and manufacturing

vices in tractors. If agriculture

countries with low labor costs (China

low-

which a high degree of knowl-

for example, in biotechnology

ing, or in “precision farming,” in

at the

is

a particularly

are

GPS

de-

moving

to

good example),

the jobs that remain in the so-called knowledge-based economies are particularly critical to these countries’

exactly

what workers

are going to vices),

but

do

it is

in the

United

b.

States,

survival.

It’s

not clear

Western Europe, and Japan

for a living in the future (other than provide local serclear that if these

many of the workers must be a.

economic

economies are to prosper, the jobs of

particularly knowledge-intensive.

Peter Drucker, The Age of Discontinuity Peter Drucker,

(New York: Harper & Row,

1969).

“The Future That Has Already Happened,” Harvard Business Review

(September-October 1997): 21.

I

have arrived

fifteen years I’ve

at the topic

from two

different directions. For

about

been doing research on business processes and

they can be improved

4 .

I’ve

come

to the conclusion that the

how

most im-

portant processes for organizations today involve knowledge work. In the past, these haven’t really been the focus of

most organizations

improving administrative and operational processes has been

easier

but they must be in the future. The other starting point has been

knowledge management

3 .

I’ve

worked with or studied many organiza-

tions that have built systems to capture

and

store knowledge, but the

10

Thinking for a Living

real

key to effective use of knowledge

is

to

knowledge workers. That, of course, leads

knowledge work can lead

Over written

to better

embed

topic,

performance and

and created or participated

panies or groups of knowledge workers.

focused on

how

into the

work of

to a broader interest in

in six studies

Some

how

results.

from what

several years I’ve gathered information

on the

it

little

is

of com-

of the studies were

to use technology to better the lot of the

knowledge

worker; others were focused on improving knowledge work processes or understanding the effect of the physical workplace on knowledge

work. Across these studies

companies and more than I’ve also

I’ve

six

analyzed or surveyed over a hundred

hundred individual knowledge workers.

gathered together a large

number of case

of organizations that are addressing this In this

book

I’ll

treat the issue

studies

and examples

issue.

of improving knowledge worker per-

formance from a wide variety of perspectives

—organizational

and

managerial, process, information technology, and even the physical

workplace. lection of

I

believe

it’s

by

far the broadest,

most comprehensive

knowledge on the topic of knowledge work and

ment. That’s not to say that there aren’t be written on

it

—and read—over

the

many other books

coming years. To use

its

col-

improve-

that

need to

a well-worn

but apt phrase, I’m just scratching the surface.

What What

is

a

Is

a Knowledge Worker?

knowledge worker?

in the following

I’ve

defined them for well over a decade

way:

Knowledge workers have high degrees of expertise, education, or experience, and the primary purpose of their jobs involves the creation, distribution, or application of knowledge.

Knowledge workers think heavy

lifting

on the job

is

for a living.

intellectual,

They

live

by

their wits

—any

not physical. They solve prob-

11

What's a Knowledge Worker, Anyway?

lems, they understand

and meet the needs of customers, they make de-

and they collaborate and communicate with other people

cisions,

own work.

the course of doing their It’s

easy to point to examples of knowledge workers: physicians and

physicists, scientists

and

sci-fi writers,

work

in knowledge-intensive industries

— managers of any company

knowledge workers, applying knowledge

are

and airplane de-

airplane pilots

We know them when we see them. They don’t necessarily have

signers.

to

in

best interests of their enterprises. Even the

make

to

most

decisions in the

industrial

company

has engineers, researchers, marketers, and planners. Knowledge work-

work

ers

in small start-ups

and

large global corporations. Outside of

work, they reside in tony, cool areas of wealthy suburbs; some have

work

For

virtually.

you deal with

in

many

moved

cities

and

in middle-class or

and do

to resort locations

their

of you reading this book, virtually everyone

your job and your

social life

could be another knowl-

edge worker.

What’s are not

difficult

is

pointing to people

who

clearly

number of jobs time. Even

if

by (London

I

it’s

probably also true that the

requiring no knowledge whatsoever has decreased over

drive a taxi,

I

need some geographical knowledge to get

taxi drivers, in particular,

have to possess “The Knowl-

edge” of its streets before getting their licenses). Even

movie

ity to

theater,

recognize

I

if

I

take tickets at

need both customer service knowledge and the

when someone’s

trying to sneak

need some knowledge of soil conditions and dirt

definitively

knowledge workers. Most jobs require some degree of knowl-

edge to perform them successfully, and

a

and

in.

abil-

Even ditch diggers

how to

lift

shovels

full

without hurting their backs. I’m sympathetic to the idea that

creasing

numbers of workers need knowledge

ever, that doesn’t necessarily

to

do

their jobs.

on the job

in-

How-

make them knowledge workers.

Definitions of knowledge workers that incorporate anyone uses knowledge

of

who

are also not very helpful. Peter Drucker, for

example, has defined a knowledge worker as “someone

about his or her job than anyone

who knows more

else in the organization.”

Drucker was

12

Thinking for a Living

knowledge work

certainly prescient about the fact that

more important, and stand their

own jobs

that taxi drivers, as

he’s right that

is

knowledge workers often under-

better than others. But this definition also

movie

knowledge workers

there

becoming

is

means

and ditch diggers could qualify

ticket takers,

to Drucker; his definition also implies that

only one knowledge worker per job per organization. By

definition, these types of workers don’t fully qualify as

knowledge

workers because creating, distributing, or applying knowledge the primary purpose of their jobs.

They only think

my

isn’t

for a small part of

their living.

Whether someone

is

a

knowledge worker or not

sometimes

a matter of degree

knowledge

in their jobs

tise,

and

interpretation.

is

admittedly

Many

people use

and have some degree of education or exper-

but for knowledge workers the role of knowledge must be central

to the job,

and they must be highly educated or

data or information alone

knowledge worker,

isn’t

enough

—

it

expert.

would be

Working with

difficult to

be a

for example, without having a college degree (col-

lege dropouts Bill Gates

and Michael Dell notwithstanding).

Despite a few necessary shadings in the definition of knowledge workers, tiveness

it’s

clear that organizational success

depends on the innova-

and productivity of these workers within

their organizations.

However, along with adding value, knowledge workers also pose challenges to conventional

management wisdom and organizing

princi-

ples:

they are mobile and concerned that their experiences position

them

well for future opportunities; they are dispersed across the orga-

nizational structure plexity of their

and the globe,

work

requires

yet the interdependence

them

to collaborate effectively with

others in different functions, physical locations, time zones, organizations; they

must command

a

be constantly updated; and their work

and com-

body of knowledge is

and even

that needs to

inherently emergent

—the im-

portant problems they solve and opportunities they capitalize on are novel and rarely,

if ever,

standard to the point that the work can be-

13

What’s a Knowledge Worker, Anyway?

come

routine. In short,

knowledge workers are

critical to

the success of

almost any organization, but they present unique challenges as well.

Knowledge Workers as a Just

how unique

might argue

knowledge workers present? Some

are the challenges

— and

at least

one person has

and knowledge work should be managed

work .

.

is.

—

knowledge workers

that

in the

same way

that other

This person recently wrote:

the old dichotomy between

.

Class

workers

is

not very meaningful

manual workers and knowledge .

.

.

fewer and fewer workers

perform routine tasks that do not draw upon accumulated

knowledge and

expertise.

To paraphrase Richard Nixon, “We

knowledge workers now.”

are

all

He

argues instead that knowledge workers should be treated like

any other workers

in business processes,

and

that process

approaches apply just as well to knowledge workers I

won’t mention that person’s

because a

I

want

to use

him

management guru, and

to have

worked with him

as

a

name

as to

I

have a

lot

to follow

it.

an example to prove that

knowledge worker par

closely for several years.

Like

many knowledge

had an employer. He

He was

picky, quirky,

every time

I

he’s

wrong. He’s

excellence.

I

happen

While he argues that just like

improved by handing him a process

a very high degree of independence sibly

else.

anyone

of experience with this fellow’s work habits, and they

certainly couldn’t be

him

anyone

or cite the source of the quote,

knowledge workers’ work processes can be addressed else’s,

improvement

map and

telling

workers, this gent worked with

and autonomy

resisted schedules

—even when he

osten-

and often missed deadlines.

and cantankerous much of the time. Virtually

met with him, the

first fifteen

be about gossip or politics or books

minutes of discussion had to

—anything but the

task at hand.

Thinking for a Living

Yet

high

when he

finally

—sometimes

produced, the quality of his work was usually very

brilliant.

puts were sufficiently important that criticizing his

productive.

most matters

workers being an exception)

I

participation

him

him by

that they could be

made more

(the proper treatment of

knowledge

had considerable respect

and dealing with him was generally worth the quite ironic for

and out-

couldn’t risk alienating

I

work habits, or suggesting

On

work

Certainly his

to argue that

for his intellect,

trouble.

However,

it’s

knowledge workers are no different

from manual, administrative, or production workers, because he em-

make them

bodies everything about knowledge workers that

and

difficult to

manage.

example

If this

isn’t sufficient,

throughout the

and occasionally through the book,

ter

what

I’ll

of this

first

chap-

how knowledge

argue that they don’t

work

to do, that the flow of their

rest

describe

I’ll

workers are different from other workers. to be told

different

is

like

difficult to structure

work best when working with other people

and

predict, that they

cial

networks, and that they are better led by example than by explicit

in so-

management. It’s

some other

certainly true that

types of workers

—

say,

produc-

tion workers in a manufacturing plant, or a checker in a grocery store

chain

—share some of

these attributes to

some

degree. Virtually

no

one, for example, likes to be told what to do. Every day, however, pro-

duction and

retail

workers are told precisely what to do (“Tighten those

bolts with twenty foot-pounds of torque, not ten”;

“You can take your

break as soon as you change the register tape”), and their managers

and companies generally plicit

get

away with

it.

If

managers gave similar ex-

instructions to their knowledge workers (“Sharpen your pencil

before you start that financial plan”), however,

it’s

employees would stay with the company for long. they tolerated being job their

full

difference in

managed

commitment and autonomy

is

this way,

it’s

intellectual

unlikely that their If

by some chance

unlikely that they’d give the

horsepower. This substantial

only one of the key attributes of knowledge

15

What's a Knowledge Worker, Anyway?

workers, but by

itself

it’s

enough

to justify treating

them

as a separate

of workers deserving the separate approaches to performance im-

class

provement and management that I’m putting forth

in this

book.

Common Attributes of Knowledge Work and Knowledge Workers The

fact that

knowledge workers primarily

rely

on

their brains rather

than on their bodies in their jobs means that they have some attributes in

common. These

and they

aren’t terribly surprising,

all

follow from a

few basic principles and observations, but they need to be derive

haps

from the

fact that

less structurable,

knowledge work

less structured,

is

defined; areas

life is less

linear; inputs

and information

where making

is less

sense, interpreting,

meaning and knowledge

ison to

is

its

The

are at a

—

are, rather,

and understanding

areas where, above

premium

are

all,

6 .

predecessor forms of toil.

basic principles

edge workers

is

and observations

like

autonomy.

follow.

One important

that they don’t like to be told

what

aspect of knowl-

to do.

Thinking for

engenders thinking for oneself. Knowledge workers are paid

for their education, experience,

and

expertise, so

that they often take offense

when someone

their intellectual territory.

Of

their

These

are less well

an excellent characterization of knowledge work in compar-

Knowledge workers

a living

and per-

this distinction:

and outputs

“targeted.”

both problematic and highly valued

This

Most

than administrative or production work. As

John Seely Brown and Paul Duguid have described In such areas,

stated.

work

to be ignored,

it

is

else rides

not surprising

roughshod over

course, knowledge workers don’t like

and there are some things they

like to

be

16

Thinking for a Living

told,

and

such as the broader significance and implications of their tasks

jobs.

This

autonomy

edge work. Since

word

in part a natural result of the nature of

its difficult to tell

tually thinking at

take their

is

it.

The outputs of knowledge work

to specify in great detail, so that

Autonomy for the

is

whether knowledge workers are ac-

also

is

generally

left

up

to the workers.

ability to

And with

manage

why they worked hard

tailed processes they follow in

need to get done and when

comes

and schedules.

If a

like

autonomy

a greater

doing their work.

Tell

as to the de-

them what they

needs to be finished, and they

it

stances in which they think best.

They

details.

will, if

they

They know the circum-

also like to decide their

computer programmer

most productive working from

would

and grad-

which knowledge workers

in

would

have their preference, figure out the

is

re-

oneself.

prefer autonomy. In particular, they

he

Some

autonomy on

in college

increased education presumably

There are several different domains

locations

exchange

fair

training they have received.

search has suggested that scientists and engineers view the job as the major reason

have to

are also difficult

viewed by knowledge workers as a

amount of education and

uate school.

much

any given moment, supervisors pretty

for

knowl-

tells

own work

the boss that

smart boss

8 p.m. to 4 a.m., a

try to facilitate that arrangement.

For the most part, knowledge workers have gotten the autonomy they want. Since they

and knowledge

—

it

own

their

would be

means of production

—

their brains

difficult for organizations to

deny them

that

autonomy. Professional associations of knowledge workers have

also

sometimes

resisted attempts to apply too

much

control in the

workplace. For example, the American Medical Association has often struggled in the courts with hospitals over physician and medical staff

autonomy. There has also generally been

knowledge workers, so they have been able

autonomy

if

necessary.

a

to

good labor market

for

change jobs to find more

17

What’s a Knowledge Worker, Anyway?

However, because knowledge workers prefer autonomy doesn’t mean they should always be given the tion throughout this book,

maximum amount of

some

As

I

men-

will

improve knowledge worker

efforts to

performance

may

involve removing

edge worker.

Still,

organizations must be careful

some

it.

discretion

from the knowl-

when implementing

any new process or technology that significantly reduces the autonomy of their knowledge workers.

Specifying the detailed steps is less

valuable

corollary of

and more

series

down

The

is

a

to

improve performance we begin by

the structure of the task into

idea

is

processes can be

that

However,

resist

if

not be-

knowledge work

and measured, and unnecessary

this

approach often doesn’t work

very well for knowledge work and workers. In

edge workers will often

constituent elements.

into piece-parts,

easily followed

steps eliminated altogether.

its

Frederick Taylor’s day,

at least since

when broken

more

like to see

of boxes and arrows.

when we want

This has been the case fore.

than for other types of work. This

be told what to do, and they also don’t

like to

reduced to a

Typically,

breaking

difficult

my first generalization about knowledge work. Knowledge

workers don’t their jobs

and flow of knowledge-intensive processes

my experience,

knowl-

describing the steps they follow to carry

out an assignment. The more complex and knowledge-intensive the

work, the more

likely this will

tions that describing the typical flow of work

work

Even process,

if

you can

it

may

get a

impossible.

same

may be

difficult to describe

knowledge worker

Knowledge

I

talk to

among

or model.

to describe his or her First,

the

work

work flow

to another worker’s description of the suppos-

process. Second, the steps

cient: “First I

this

not be a very helpful description.

may not be very similar

Then

is

also often involves a high degree of iterative collaboration

knowledge workers, and

edly

many varia-

be true. Perhaps there are so

come up with an

idea.

my lab partner about

may seem maddeningly

Then

it.

I

think about

Then I think about

it

ineffi-

for a while.

the reactions

18

Thinking for a Living

me.” Such a process would be anathema to a stopwatch-

she’s given

packing Taylorist, but

it’s

how knowledge

often

those involved in knowledge creation activities

So does

this

mean

that

we have

to give

workers

—work.

—

up on taking

a process per-

on improving knowledge work? No. Despite the

spective

from analyzing process workflows

particularly

in detail, there are

lack of value

plenty of

still

other process tools available to potential improvers of knowledge

work.

I

will describe these in chapter 4.

eralization

I’ll

make about

“You can observe a right

—

is

you have

the next gen-

type of work.

by watching .” Lawrence Peter Berra had

that

if

you

to observe

can’t get

it

them

to describe their

how knowledge workers do

—

is

often an effective

—

it

knowledge

work

in detail. Systematic observation

“shadowing” or “ethnography”

stand

is

a natural follow-on to the previous attribute of

workers

as

lot

this

One, however,

in detail,

also

way

known

to under-

their work.

Observation can be undertaken through a variety of means. The

most or

common

is

to have

an observer “shadow” or follow around one

more knowledge workers.

ticipate in the

work process

It’s

—

also possible to have the observer par-

called “participant observation.”

server can obviously learn a great deal about the

work by doing

participation can also increase trust levels. Finally, takes place through videotaping

though

this

approach

is

and

later analysis

is

standing knowledge work. One, since thinking out” with knowledge workers as they do their server a better sense of what

comments, the

firsthand,

is

jokes,

and use them

and may lead

to

and

some observation

of the work process,

really

well suited to underis

invisible,

work

“hanging

will give the

and the complaints of knowledge workers

to piece together a story of the work.

more

ob-

going on. The observer can hear the

spending time with knowledge workers increases their trust server,

it,

hardly conducive to trust.

There are several reasons why observation

casual

The ob-

Two,

in the

ob-

disclosure about the nature of their work.

19

What’s a Knowledge Worker, Anyway?

But

this

approach to understanding knowledge work has significant

drawbacks. Paying for a

human

observer

is

expensive,

and

it

may

re-

quire substantial time for an observer to gain the trust of knowledge

workers and

fully

understand their work. Nevertheless, without obser-

vation, its unlikely that the

Of

work can be

penetrated.

course, there are also ethical considerations to observation.

group of knowledge workers shouldn’t be observed unless

and intentions of the observation

jectives

that

to help

the ob-

are fully disclosed. I’ve

knowledge workers are usually quite willing

cuss their

all

to

them, but nothing makes them clam up

found

welcome and

work with an observer if they feel the observation faster

is

A

dis-

intended

than a suspicion

that the observation will have negative consequences.

Knowledge workers usually have good reasons for doing what they In the days of business process reengineering, analysts could quickly figure out better fact,

often true.

Nobody had

we assumed

do.

that smart

ways of doing work. This was,

ever thought about

in

many administrative and

operational processes before, and improvements were easily identified. It’s

not so easy with knowledge work, which

we have

to observe

it

closely.

one of the reasons why

Knowledge workers have

about

why and how they do

many

of the obvious improvements to

their

work, and it.

improvements

typically

thought

may have themselves made There

behind almost everything they undertake (or tionalization). If

is

at

is

probably a reason

minimum

are going to be identified,

a logical rait’s

probably

who

only after serious and deep study. Julian Orr, an anthropologist

has studied technical service representatives, argues that analysts of

such work are rarely sufficiently “concerned with work practice

do not focus on what

is

done

in

accomplishing a given job .”

So for knowledge work, we need to take workers

more

importantly, their deed.

chatting with other workers

is

.

.

.

they

7

at their

word

—or

We must be careful before assuming that wasted time, or that time spent cogitating

can be eliminated without consequences.

It’s

usually safer to assume

20

Thinking for a Living

that

work

is

done

way

a certain

good reason than

for

to

assume

can

it

be quickly and easily changed.

Commitment

job with one’s body even to the job.

But

when

knowledge work.

It’s

a

unlikely that

performance out of a knowledge worker

if

he or she

mentally and emotionally committed to the job.

This fact has a

number of ramifications. Chief among them

knowledge workers need some say

do

one could do

the brain and heart weren’t committed

this isn’t the case for

you’ll get great isn’t

In the industrial economy,

matters.

it.

There

is

in

mous 3M approach

This factor

else.

is

like

being told what to

behind, for example, the

is

important to the com-

pany. Obviously knowledge workers are generally willing to things that others ask (or even

tell)

them

to do,

do some

but a degree of volun-

lot.

Another factor affecting commitment process.”

fa-

of giving researchers 15 percent of their time to

work independently on something they think

tarism helps a

that

what they work on and how they

nothing that limits commitment

work on by someone

is

is

a perception of “fair

As the strategy academics Chan Kim and Renee Mauborgne

have pointed out, workers

— and

particularly

knowledge workers

care not only about the fairness of outcomes, but also about the fair-

ness of the process used to arrive at outcomes: Fair process turns out to be a powerful

companies struggling

to

make

management

tool for

the transition from a produc-

tion-based to a knowledge-based economy, in which value creation depends increasingly

on

ideas

and innovation.

Fair

process profoundly influences attitudes and behaviors critical to high performance.

If

It

builds trust

and unlocks

ideas.

8

high levels of commitment are going to be maintained, managers

of people

who

think for a living need to acknowledge the fact that

workers can sometimes think better than their bosses. The manager or

21

What's a Knowledge Worker, Anyway?

may

executive

have a better grasp of the strategy or the big picture of

may know

the business, but the knowledge worker

her

much

field

better.

Having

a

manager pretend

the details of his or to

know

can be very damaging to an experts commitment and

Knowledge workers value

Knowledge

means of

trade, the

would have

,

their production.

It’s

—

it’s

it easily.

the tool of their

therefore natural that they

difficulty relinquishing or sharing

own jobs might be

loyalty.

knowledge and don’t share

knowledge workers have

that

is all

their

everything

it

in

such a way that their

threatened.

In the early days of

knowledge management, when companies were

beginning to talk about sharing knowledge within and across organizations,

I

used to

mentioned

that,

Companies

just

“Sharing knowledge

say,

“Of

needed

act.”

committed every

to put the necessary incentives

also

I

day.”

and assurances

people were willing to share their knowledge.

anything, a knowledge worker’s concerns about sharing have be-

come even more fears

an unnatural

course, unnatural acts are

in place to ensure that If

is

of

layoffs,

few years. In addition to the usual

justified in the last

some

clueless

companies have been asking

their

ers to train their offshore successors before they lose their jobs.

every knowledge worker the next to

move

is

wondering whether

to India or China.

It’s

work-

Almost

his or her job could

enough

to give

be

anyone pause

about contributing knowledge to some other worker or a knowledge repository. Again, this doesn’t

and processes

We

just

it

that

we

such a way that knowledge

can’t design organizations

will

flow across organizations.

have to acknowledge that knowledge workers will view their

knowledge share

in

mean

as a highly valuable asset,

and

that they will be reluctant to

without rewards and/or guarantees of continued employment.

There are other precedents for years companies have tried to

this

make

problem

in organizations.

For

the customer information of their

salespeople an organizational, rather than an individual, asset. Sales forces have

worked just

as diligently to

withhold what they

know about

22

Thinking for a Living

their customers, often resisting

management” systems

adding

it

“customer relationship

to the

their organizations have

implemented.

Many

organizations have added policies and procedures that reassure the sales force that they’ll

still

own

the customer information

and knowl-

edge while they work for the company. Smart organizations will put

knowledge

similar approaches in place for the

assets of their

knowl-

edge workers.

Implications of These Attributes The implication of these “managed”

attributes

knowledge workers

that

is

be herded.

9

Some

be

word. As Warren Bennis has

in the traditional sense of the

pointed out, they bear a peculiar resemblance to can’t

can’t

cats,

which

as

we know

of the attributes of knowledge workers sound

like T. S. Eliot’s descriptions

of the contrary

cat,

the

Rum Turn Tugger:

For he will do

As he do

And

do,

there's

When you

no doing anything about

let

him

He's always in the

For he only

The

Rum

likes

in,

.

then he wants out, ,

.

what he finds for himself ,

Turn Tugger

is

this

artful

.

.

.

and knowing

10 .

.

.

type of individual can only be led, not

aged, by visionary, inspirational leaders spect in their organizations. This

tise

.

wrong side of the door

Bennis argues that

some

it!

is

man-

who build trust and mutual re-

no doubt the

and

case,

I

discuss

aspects of this leadership in chapter 9 (no doubt with less exper-

and grace than Bennis would). But visionary leaders are often

are other ways in

in short supply,

which we can intervene

in

and

I

believe there

—and improve—knowl-

23

What’s a Knowledge Worker, Anyway?

edge work. As with any other type of work, there are

many

potential

avenues to betterment. In addition to better leadership and manage-

ment, in people

this

who

book

I

describe at least five other approaches to helping

think for a living.

What’s Coming Thus

far I’ve

chapter

described what knowledge workers are,

and some of the things they have

portant,

I’ll

Book

in This

how

describe

they

differ.

in

Both the

why

common.

they’re

im-

In the next

similarities

and

ences form an important set of baseline factors to consider

differ-

when

at-

tempting to improve knowledge work. That subject

—improvement

forms the basis for the other chapters. In chapter

3, I’ll

improvement or interventions

subject of

tervening in the

mon

first

place,

how

mistakes that organizations

subsequent chapter

I’ll

to

—the

case for in-

measure improvement, and com-

make

in their interventions. In each

describe a particular approach to intervening in

and improving knowledge work. Chapter 4

and measurement. Chapter tions as a

in general

describe the

talks

about the roles of process

5 describes technology provided

by organiza-

means of improving knowledge work, while chapter 6

on technology, information, and knowledge management sonal

level.

Chapter 7 describes the

social

focuses

at the per-

networks of high-performing

knowledge workers. Chapter 8 describes the physical workspace

knowledge work. Chapter 9 focuses on how

to

be an

effective

for

manager

of knowledge work and workers.

At the end of each chapter subject treated for

and

results

summary

I

will

point out the implications of the

managers wishing

to extract better

performance

from knowledge workers. These sections provide

a quick

of the recommendations in the chapter, but of course they

don’t provide the context, the examples, and deeper analysis.

24 Thinking for a Living

Recommendations for Getting Results from Knowledge Workers •

All jobs involve

knowledge

to

some

degree, but knowledge

workers are those whose jobs are particularly knowledgeoriented.

The recommendations

in this

book apply to those

expert workers in jobs whose primary purpose distribute, or apply

•

•

to create,

knowledge.

Knowledge workers their

is

differ

from other kinds of workers

autonomy, motivations, and

attitudes.

Knowledge workers enjoy their autonomy, so be

careful

about improvement approaches that impinge upon although sometimes •

this

may be

it

necessary.

Knowledge work tends

to

detailed flow of work

sometimes

is

in

be unstructured. Specifying a possible, but

is

probably

not the best way to improve a knowledge work process. •

Knowledge work often needs and

•

at

some length before

Knowledge workers

it

to be observed in

•

are usually intelligent, so be careful

thing to

is

work process can be improved upon

Commitment

detail

can be truly understood.

about assuming that a particular work task or that a

some

unnecessary, easily.

matters to knowledge work. Don’t do any-

damage

the knowledge worker’s

the job and to the organization.

commitment

to

2

How Knowledge Workers

Differ,

Difference

As

I

all

knowledge workers

stated in chapter

1,

It

despite having a aren’t alike.

A

and the Makes

number of traits

in

common,

computer programmer and

a

physician, for example, are both knowledge workers, but they have very

backgrounds, working conditions, business

different

educational

processes,

and measures of effectiveness and

proach to classifying knowledge workers

mine how

best to manage, measure,

classification serves a useful

purpose

success. Therefore,

may help

organizations deter-

and improve if it

an ap-

passes

their work. This

two

tests. First, it

should be easy to apply; putting any given type of worker into one of the classes should not require great tions.

amounts of work or mental gyra-

Second, the classification should be useful in improving the per-

formance of knowledge workers when some intervention Otherwise, the distinction makes no difference.

is

attempted.

26 Thinking for a Living

Among

Distinguishing

Workers One important purpose is

to

make

Your Organization

for distinguishing

improve

that organizations can’t

They have

in

Kinds of Knowledge

all

among knowledge workers

knowledge worker

roles at once.

choices about which knowledge-oriented jobs to

address at any given time. Understanding the distinctions

knowledge workers begins

when some

prioritization

to provide a basis for is

necessary.

the “perfect matrix” for distinguishing

have come to the conclusion that

many important ways variations to

sions are

it

I

choosing

among knowledge workers, and

doesn’t exist. There are simply too

two dimensions. However,

more important than

among them

have tried for years to develop

which knowledge work

in

among

others,

I

differs to

do think

some dimen-

that

and the matrix

reduce the

I’ll

describe next

has been very handy in terms of shaping an intervention for different types of knowledge work.

Distinguishing on

The matrix it

Judgment and Collaboration

in figure 2-1

uses the level of

is

an example of a

work complexity

classification

(the interpretation

approach;

and judgment

required in the process) and the degree of collaboration required as classification

dimensions.

1

These dimensions are important because

the level of collaboration often drives the degree of structure and

com-

puter mediation that’s possible in a particular job, and the level of

complexity of the work can dictate

perform

A An

it

how much knowledge is required to

successfully.

transaction worker, for example,

might be found

in a call center.

integration process

example might be an information system devel-

An

investment bank might provide an example for

opment

operation.

the collaboration model, and a primary-care physician

with the expert model.

I

practice

have found these distinctions useful in deter-

mining what kind of interventions make sense the matrix.

would

for the different cells of

27

How Knowledge Workers FIGURE

Differ,

and the Difference

It

Makes

2-1

A classification

structure for knowledge-intensive processes

Collaborative

Integration

model

Collaboration model

groups •

A

4)

O c

Systematic, repeatable work

•

on formal processes,

•

• Reliant

4) T3

Improvisational work

Highly reliant on

methodologies, or standards

c

Dependent on

•

4)

a

Dependent on fluid deployment of flexible teams

•

integration

across functional boundaries

4)

deep

expertise across functions

| 4>

Transaction model

*->

C

Expert model

o

•

Routine work

•

4)

•

on formal rules, procedures, and training

•

Dependent on low-discretion

•

«*-

Reliant

> 0)

•

Individual

Judgment-oriented work Highly reliant on individual

expertise and experience

Dependent on

workforce or information

star

performance

actors




a.

Process applications and work-flow

Decision

automation

+->

c

Data mining/ analytics

0)

> o

Transactional

Embedded knowledge

technologies

Individual

actors

Transaction

Expert model

model

interpretation/

Complexity of work

routine

where work involves

relatively

judgment

low amounts of collaboration and judg-

ment, the most appropriate technologies are those that automate structured transactions.

A

call

center system that brings calls and the

and knowledge

relevant information

to the

worker would be an exam-

ple of this type of system. BT’s AdvisorSpace system for advisers,

which

I’ll

describe later in this chapter,

its call

center

would be an example

of such a system.

As the degree of collaboration moves up into the integration model, applications that structure the process and the flow of the to

make

sense. In a

new product development environment,

lower-level engineers cycle

might have

management system

their

work begin for example,

work structured by a product life-

that keeps track of designs,

components, and

approvals for a major product design. Also within the integration

would be systems

for

knowledge reuse

—

again, for example, in a

cell

new

product development environment, where computer-aided design

88 Thinking for a Living

drawings might be reused. Reuse can also stretch into the collaboration environment, where an attorney, for example, might reuse a Decision automation, which this chapter,

I

detail later in

suited for job roles with a middle level of structure

is

expertise, such as insurance underwriting. roles

some

will describe in

legal brief.

can be automated, but experts are

The

still

and

lower-level jobs in such

necessary to build and re-

fine the system.

In the expert cell generally, the goal a

computer mediate the

I

Experts

ter.

some means of having

to find

expert’s work. If that’s possible, then

feasi-

it is

about embedding knowledge into the flow of the work

ble to think

process, as

is

will describe in a health care application later in this

may

from data mining and decision

also benefit

chap-

analysis

applications for jobs involving quantitative data. In the collaboration erative

cell,

as

I

argued in chapter

and unstructured. The only types of tools

2,

work

is

usually

it-

work

in

that typically

such environments are knowledge repositories and collaboration

which are used

voluntarily.

embedded knowledge, but and use about

There could possibly be systems involving these

in a highly collaborative

many of these

aids,

would be more

work

process.

I

difficult to

will

develop

have more to say

technologies later in the chapter.

Despite the power of technology to improve performance, not

all

forms of knowledge worker technology have been successful. There have been two pervasive dreams with regard to knowledge workers and technology. cess

all

One

of the data

necessary to itself

is

that

—

make

knowledge workers would be able

typically data arising

decisions.

to easily ac-

from business transactions

The second dream

is

that

knowledge

—typically unstructured, textual knowledge—could be

tured, shared,

and applied

to

knowledge work. Neither of these dreams

has been fully realized, but progress tion.

It’s

lows

is

taken

much

is

being

made toward

their realiza-

longer than anyone expected, however.

a brief history of organizational technologies for

workers.

easily cap-

What

fol-

knowledge

89 Organizational Technology for Knowledge Workers

Decision Support

new

In the late 1970s, a

idea

emerged from academics who worked

with applications of information technology. Called “decision support,” the concept involved the use of

make

data available for

human

complex computer algorithms

decision making.

2

The dream of

to

deci-

sion support proved to be a persistent one. IT academics in business

schools found subject

it

almost

writing thousands of papers on the

irresistible,

—perhaps many more papers than

vendors of software and services arose to ities.

The

original concept evolved into

sell

there were readers.

Many

decision support capabil-

Management Decision Support

Systems (MDSS), Executive Support Systems (ESS), Online Analytical Processing (OLAP), Relational Online Analytical Processing (ROLAP),

Multi-dimensional Online Analytical Processing

(MOLAP), and

so forth.

Despite subtle variations, the basic concept remained the same, however: a

computer program would churn through data and with human

interpretation

would

reveal previously

would allow an executive

to

hidden trends and patterns that

make smarter and

faster decisions.

However, the notion of computer-augmented decisions never exploded

like

some other categories of computer

and database management.

First

of

decision rules or algorithms from

computer

—and the

artificial intelligence

task didn’t

all, it

human

software, such as

was often

ERP

difficult to extract

experts and put

become much

really

them

in the

easier with the advent of

and knowledge engineering. Second, many knowl-

edge workers didn’t trust a computational black box to help them with decisions, preferring to rely to decision

making.

A

on gut

third issue

feel is

process to analyze the data needed to

support tool that did take off was a

or other traditional approaches

that

make

it

can be a time-consuming

a decision.

The only decision

relatively primitive

one

—

the Excel

spreadsheet. Even with this simple tool, however, a high degree of human

labor

is

often required to create

and

interpret the spreadsheets,

created by inexpert managers and analysts are often

rife

and those

with errors.

90 Thinking for a Living

Artificial Intelligence

Another idea that originated

and Expert Systems

in the 1970s,

and reached

its fullest

flower-

ing in the 1980s, was the implementation of “artificial intelligence” and

“expert systems.” These technologies were supposed to eliminate or re-

duce the need for knowledge workers by extracting their knowledge and having a computer

itself

make important

This was an important idea, and as are

still

attempting to realize

it.

But

decisions or judgments.

I’ll

describe

later,

organizations

at least in the first set

of attempts,

automating knowledge-based decision making didn’t work out very well.

A number

field

encountered numerous problems; for example:

•

The knowledge intended from the

•

of expert systems were developed, but pioneers in the

for the system

was

difficult to extract

expert’s brain

The knowledge

in a

system generally needed to change more

rapidly than the system designers anticipated, revising for such change •

The

was

difficult

tial

and expensive

best systems proved to be those that

experts, rather than replacing

them

and constantly

augmented human

—which lowered the poten-

economic returns from expert systems

Knowledge Management The

first

largely successful generation of organizational technology for

knowledge work was known nology began to appear Notes and

later the

as

in the

“knowledge management.” 3 This tech-

mid-1990s with the

availability

Web, and became quite popular

of Lotus

until the general

retrenchment of information technology in the early 2000s. Knowledge management technology generally involved the creation of repositories

—

essentially databases

—of

knowledge. Organizations stored

91 Organizational Technology for Knowledge Workers

almost every imaginable variety of knowledge, including best practices,

competitive intelligence, observations about customers, learnings

from previous

projects,

and so

forth.

The most knowledge-intensive and

industries, including professional services, pharmaceuticals,

functions within manufacturing, were

all

development of knowledge management But even

this

technology had

characterized by extensive

repositories.

problems.

its

It

was expected that

knowledge workers would find or contribute information time.

much

The problem, of

course,

spare time. As firms

became

increasingly lean

increasingly engineered,

it

became

difficult

large

and

—

rarely

had

and work processes

became impractical

contribute to knowledge repositories ries

in their spare

was that knowledge workers

became

R&D

particularly as

to consult or

some

reposito-

and unwieldy. As economic conditions became more

repositories

became

less

valuable in the early years of the

new century, knowledge management

retreated in

many firms.

However, repositories should not be entirely dismissed. There are circumstances in which they are probably the only feasible approach to

supplying knowledge workers with the data they need to do their jobs. If a

knowledge workers job process

orative,

it is

highly unstructured and collab-

is

very difficult to determine in advance what knowledge and

information a particular worker or position requires.

And

if

there

is

no

technological application that can mediate the worker’s job, repositories

may be

the only alternative. This

is

the case, for example, with the

collaborative jobs described in chapter

2.

Jobs such as consulting or in-

vestment banking, for example, meet

all

the criteria described above.

It’s

almost inconceivable that there would be an “investment banker’s

workstation” that would guide a banker through

all

the steps of the job,

supplying information and knowledge as required. In any case, no one has yet built such a system, although such tools for similar jobs have been

envisioned

if

not actually implemented.

4

Therefore, organizations in

such industries will have to free up the time to enable their employees to seek

and share knowledge from

repositories.

Thinking for a Living

Integrating

Knowledge into the Job

What’s the alternative to repositories as a knowledge management tool?

One answer

process

itself

— the

embed knowledge

to

is

embedded knowledge

into the flow of the job

applications that

I

men-

tioned above for expert knowledge workers. Under such an approach,

knowledge workers don’t have

them

at the

time of need. In

to seek out knowledge;

fact,

it is

“integrating knowledge

into business processes”

was selected

knowledge management

in a

as the

delivered to

management

most important

issue of

2002 survey of experts and practitioners/

Are organizations flocking, then, to embed knowledge into the

work processes of it’s

their

do

quite difficult to

knowledge workers? No, unfortunately this (I will describe

some of

below). There are a few good examples, however.

comes from Partners Healthcare, tals in

Boston.

Some

a

One

of

knowledge workers use for physicians.

to

and very

to

do

bake knowledge into knowledge is

to

embed

their jobs,

it

which

into the technology is

what Partners has

When knowledge supports the primary technology-

enabled transactions used in day-to-day work, activity requiring slack

There are

early

6

work, the most promising approach

done

favorites

other health care institutions are pursuing the

While there are several ways

that

my

group of Harvard-affiliated hospi-

same technology, but the Partners approach was both well executed.

the obstacles

a variety

it is

no longer

a separate

time and the motivation to seek knowledge. of ways to bring knowledge to physicians in the

course of their work, and Partners Healthcare employs several of

them. Knowledge used by

its

record, logic

tially

embedded throughout

physicians.

refers a patient to ical

is

When

the information systems

a doctor prescribes a drug, orders a test,

another physician, or even

modules and

a

calls

up the

patient’s

knowledge base are invoked

intervene in the care process. Referrals

may be

med-

to poten-

suggested by the

system to be incorrect or unnecessary. Calling up a medical record

may

Organizational Technology for Knowledge Workers

lead to a

recommendation

that certain follow-up tests or

recommen-

dations are desirable.

At the heart of

its

approach, however,

is

a

computerized physician

order entry system with trusted knowledge built

inform the physician that the patient

is

in.

The system may

taking a drug that interacts

with the drug being prescribed, or that the drug prescribed effective or

orders,

it

is

not

economical for the indicated disease. In the case of

test

may

the disease or

note that the

symptoms

A

not generally useful in addressing

identified, or that the test has already

performed on the patient treatment.

test is

sufficient times to indicate a diagnosis or

physician would use different systems (a referral system

or a computerized medical record system) for actions. a

However,

common

been

all

some

different trans-

of these systems are integrated and

all

leverage

common

database of patient clinical information and a

logic engine.

The order entry system because ordering tient care;

it is

is

is

key to the delivery of quality medical care

where physicians execute

the point at which knowledge

the system, there

and place where

would be no easy way it is

their decisions

most

is

to apply

about pa-

valuable.

knowledge

Without

at the

time

needed most. Such an order entry system may

crease both efficiency

and

safety, in the latter case

in-

by avoiding misin-

terpretation of poorly written orders. But the primary value

is

surely

the ability to insert knowledge into the process.

There

are,

of course, times

when

physicians need medical knowl-

edge when they are not face-to-face with a patient. For these circumstances, Partners has developed a patient “event detection” system that

provides alerts to physicians via wireless pager tienfs

monitored health indicators

pected.

or

new

treatment.

a hospitalized pa-

significantly depart

The physician can then proceed

call in a

when

from those

ex-

to directly observe the patient

The computerized medical record system

can also generate reminders to physicians that a particular patient

should receive a

call,

or schedule an appointment for a follow-up.

Thinking for a Living

The power of knowledge-based order erized medical record in real time.

entry, referral,

and event detection systems

Knowledge

is

and comput-

that they operate

is

applied directly and immediately in the pa-

does not have to seek

tient care process; the physician

it

out. In

situations, physicians can also get real-time access to experts for line or

are not real-time; this in the logic

tion to search

it

other sources of knowledge that

knowledge can be more extensive that that

modules, but

out.

many

requires

it

some time and

the motiva-

Online knowledge repositories (called the Part-

Handbook) include online journals and

ners

an on-

telemedicine-based consultation.

Partners has also assembled

found

some

databases, care protocols

or guidelines for particular diseases, the interpretive digests prepared

by Partners physicians, formularies of approved drugs and

and even online textbooks.

their use,

All of these

While the

more

sources available to Partners physicians are perhaps at

on

knowledge resources

are accessible through an integrated intranet portal.

than those

details

re-

extensive

become widely

other hospitals, similar resources have

available to those other practitioners as well.

Most Partners physicians tively

research, so they

their time to consult such resources

result the online

about

do

a

Handbook

is

accessed, across

thousand times a day. Contrast

this

is

a rela-

certainly limited.

all

situation of these physicians

is

the

As a

Partners institutions,

with the thirteen thousand

orders a day at one Partners hospital, Brigham and

The

may have

high appetite for online knowledge. They are busy people, how-

and

ever,

also

same

Womens,

as that in

alone.

which many

businesspeople find themselves today: a wealth of online knowledge,

but

little

time or context in which to peruse

it

unless

it

has immediate

practical application.

While the Partners knowledge approach has been under develop-

ment

for over a decade,

it’s still

not complete. The online order entry

system and related knowledge are only accessible within the organization’s

two

flagship hospitals, Massachusetts General

and Brigham and

95 Organizational Technology for Knowledge Workers

Womens. Medical knowledge

has not yet been codified for

eases Partners physicians treat.

The knowledge

several different information systems still

and

all

embedded within

is

difficult to access.

is

the dis-

There

is

plenty of work to be done.

Yet there

is

clear evidence that the

trolled study of the system’s

approach

beneficial.

A

con-

impact on medication errors found that

serious errors were reduced by 55 percent. that Partners experts

is

found particularly

increased from 12 to 81 percent.

When

The ordering of a new drug

beneficial for heart

the system began

problems

recommend-

ing that a cancer drug be given fewer times per day, the percent of orders entered for the lower frequency changed from 6 to 75 percent.

When the system began to remind physicians that patients prescribed a treatment of bed

rest also

needed

a prescription of the

blood thinner

Heparin, the frequency of prescription increased from 24 to 54 percent.

These improvements save not only ple,

lives,

but also

costs.

For exam-

recommendations from the system can point out drugs

cheaper as well as more

effective.

drug event (ADE) costs $2,000 other costs.

A

typical

Most

significantly, a single adverse

in repeat tests, extra hospital days,

seven-hundred-bed hospital

million per year in preventable

ADE

costs.

To

will incur

is still

yet seen their costs go

down, nor have national malpractice

some

rare

enough

and

about $1

date, order entry with

embedded knowledge

clined.

that are

that U.S. insurers have not

However, Partners, which insures

itself for

costs de-

malpractice, has

early data suggesting that malpractice reserves can be smaller be-

cause there are fewer drug-related claims. It isn’t

ally

—and

easy to develop such systems

—

either technically or manageri-

there are few off-the-shelf packages for knowledge-intensive

business processes that would allow individuals and organizations to

own knowledge

When

embed

their

create a

complex information and technology infrastructure

pull together the

into systems.

knowledge base and

logic

Partners needed to that

would

modules with an integrated

patient record system, a clinical decision support system, event

man-

Thinking for a Living

agement system capabilities,

hospitals

for alerts,

had

it

now

an intranet portal, and several other system

to develop

have some or

most of the systems

all

itself.

Other leading

of these capabilities, but Partners’

time knowledge approaches are certainly

real-

at the leading edge.

However, the nontechnical and managerial aspects of the overall approach are

just as

important to

its

success,

and perhaps harder

—each of which would be work— bake knowledge

implement. Several of these issues to

any organization seeking to

to

relevant

into

are de-

scribed below.

Motivation.

velop the

work

is

How did

Partners executives

become motivated

embedded knowledge approach? Embedding knowledge

into

time-consuming and expensive, so substantial motivation

necessary in order to undertake such an initiative.

motivation

at

One key

is

source of

Partners was research that found surprisingly high levels

of medical errors and preventable adverse drug events pitals.

to de-

at Partners

hos-

That these leading institutions could be unconsciously acting in

was troubling, and

direct opposition to their healing mission

vated action.

The

CEO

of one hospital

at

it

moti-

which these errors were ob-

served and measured, H. Richard Nesson of Brigham and

Women’s

Hospital in Boston, insisted that there had to be a solution.

The drug

errors

problem was the

knowledge

gram

in this area

first is

that Partners took on, because

relatively straightforward

into an order entry system.

Partners executives

moved on

to

As

trust

more

among

difficult

much

and easy

of the

to pro-

physicians increased,

and complex domains

such as patient care protocols.

Establishment of a credible and up-to-date knowledge base. the knowledge of an organization it

is

embedded

has to be of high quality and currency.

cratic, obsolete,

If

or untested knowledge in

When

into critical processes,

Partners employed idiosynits

medical care processes,

it

97 Organizational Technology for Knowledge Workers

would

subject

its

patients

and

itself to

high

The organization has

risk.

addressed this issue by forming several committees, and empowering existing committees, to identify, refine,

and update the knowledge used

domain. Medication recommendations

in each

from Drug Therapy Committees. Care protocols are designed

take

on the

for particular diseases

task of developing logic to guide radiology test ordering.

their particular

viewed

system come

by teams of specialists. Radiology Utilization Committees

Leaders of clinical service lines

is

in the

(e.g.,

cardiology) are recruited to

knowledge into the system. Participation

as prestigious,

and busy physicians

new umbrella

groups

are willing to devote extra

time to codifying the knowledge within their has been creating

in these

embed

fields.

Of

late,

Partners

organizations to oversee the knowl-

edge bases across the organization, and to put the knowledge in a more accessible

and manageable form.

Prioritizing which processes

these

embedded knowledge

and knowledge domains

initiatives are difficult

should only be undertaken for truly

critical

to address.

Since

and expensive, they

knowledge work processes.

There are many different kinds of knowledge work jobs

in hospitals

and

other knowledge-intensive organizations. Organizations should generally

look

first at

knowledge worker

organization’s mission

and

that

roles that are critical to achieving the

may

important knowledge. At Partners, ical

care processes as the

cisions to be

patient to a specialist)

more are

and multiple difficult to

still

(e.g.,

and

in

was

critical,

made about which

subprocesses to address

ations

most

it

be bottlenecks for large bodies of relatively easy to identify

but there were

disease

still

med-

important de-

domains and which medical

ordering medications versus referring a

what

order. Fields with

many disease vari-

alternative treatment protocols (e.g., oncology) are

include in the knowledge systems. Partners executives

trying to determine

what types of knowledge resources would

be most useful for other workers providing patient care

(e.g.,

nurses).

Thinking for a Living

Leaving the final decision up

end knowledge workers such

move

the

human

might either

as physicians,

was

it

would be

a mistake to re-

experts from the decision-making process.

system

reject or resent the

patients. This

knowledge worker. With high-

to the

a mistake

if it

made by some

They

challenged their role with

organizations, such as those

implemented medical expert systems attempting

to take over diag-

nosis, over the last couple of decades. Overreliance

on computerized

that

knowledge can

also lead to mistakes.

to present the physician with a

The approach taken by Partners

recommendation.

physician will then combine his or her system.

On an average day at Brigham

It is

hoped

own knowledge with

When

detection system generates

tics

alert,

that of the

computer

medication allergy or conflict warnings are gener-

ated, a third to a half of the orders are cancelled.

of an

that the

and Womens, out of 13,000 or-

ders entered by physicians, 386 are changed as a result of a suggestion.

more than 3,000

The hospitals event-

alerts

per year; as a result

treatment was changed 72 percent of the time. These

are an indication that the hybrid

at Partners is

is

working. Partners

ber of alerts to the

minimum

the need to routinely override

is

statis-

human/computer knowledge system

now working on

reducing the

num-

necessary, so that physicians won’t feel

them

for low-level concerns.

Developing a measurement and improvement culture. In order to justify

both the

effort

system, and to assess

on

this

kind of

and spending required

how well

initiative

it’s

for

working, an organization embarking

needs to have a measurement-oriented cul-

ture in place. Recall that the Partners effort

measurement of medical

an embedded knowledge

errors.

was

motivated by

initially

The tracking mechanisms within

order entry system can detect whether physicians use the

knowledge and change to

know that

their treatment decisions,

the approach

is

is

embedded

the only

way

working. Partners has always had a strong

measurement culture because most senior

which

the

it

is

an academic medical center, and

clinicians are also researchers.

But the development of this

99 Organizational Technology for Knowledge Workers

knowledge management approach has both enabled and required a greater emphasis

on measurement of key

processes.

The measures

are

used as justifications and progress reporting tools for efforts to reengineer and continuously improve care processes.

Whenever

Putting the right information and IT people in place.

knowledge technologies are applied to attribute

much

to business problems,

of the benefit to the technology

Partners case, and in

many others

I’ve seen, the

it’s

tempting

But in the

itself.

improvement

based on the employment of talented humans. Certainly

it

is

heavily

requires an

IT organization that knows the business and can work closely with key executives

and knowledge-rich

professionals.

could never successfully build a system of

management of measures, quires a staff that this discipline

is

is

this type.

It

For Partners, the also re-

information management. In health care

called “medical informatics,”

leaders in this field.

IT group

and knowledge

patient records,

skilled in

A “back room”

and Partners has recruited

has several medical informatics departments, in-

cluding Clinical and Quality Analysis, Medical Imaging, Telemedicine,

and

Clinical Information Systems

R&D. The

leaders of each of these

departments are MDs, but they also have advanced degrees fields as

computer

science, statistics,

in

such

and medical informatics.

Performance Support The Partners is

new from

not

new

case,

and the general idea of baking knowledge

the standpoint of knowledge

at all

into work,

management. However,

it’s

from the perspective of organizational learning ap-

proaches. Well over a decade ago, for example, leading thinkers in the learning and training fields began to notice that training given substantially before a task

performance of the

was performed was not

effective in

task. Gloria Gery, in particular,

improving

wrote a book on

100 Thinking for a Living

this topic entitled Electronic

Performance Support Systems in 1991/

It

argued for just-in-time learning provided through electronic technologies

—

a vision that

is

remarkably similar to the just-in-time provi-

sion of knowledge I’ve just described at Partners.

Gery and her colleagues who advocated what has come as

“performance support” were correct,

to be

known

ahead of their time. They

if

were confident that the concept would penetrate industries during the next few years and change the

way work and

organizations. Unfortunately, however,

all

learning are performed in

too few of these integrated

work and learning environments have been

some

Certainly there were port, but even

technological barriers to performance sup-

more problematic have been

cation, lack of understanding

traditionally

however,

ish,

that

is

minded it is

trainers.

likely to

actually implemented.

economic

and sponsorship, and

When

look very

embedded within work

issues of

justifi-

resistance

from

performance support does flour-

much like knowledge management

processes.

Role-specific Portals

The Partners example and performance support technology

how powerful edge baked

in.

it

illustrate

can be to build customized IT applications with knowl-

But there

knowledge workers that

is

is

another approach to delivering knowledge to

halfway between a knowledge management

repository and a customized application: the role-specific portal. portal

is

a

Web-based information delivery application

range of information and knowledge on one

site.

A

that provides a

A role-specific portal

narrows that range by attempting to provide only the information and

knowledge required

for a particular role or job. Like a repository,

quires that the user search for the information, but so that the search

is

not

may

re-

limits the scope

difficult.

The information and knowledge specific portal

it

it

accessible to the

worker

in a role-

be a mixture of transactional information, textual

101 Organizational Technology for Knowledge Workers

knowledge, multimedia educational content, and links to

by the

user.

The

screen should provide

much

take too

Not

time.

all

created

the information and knowl-

all

edge necessary to do the job, and no more

would

sites

—otherwise

the search

of the information on the portal

need be unique, but views of commonly available information should be specific to the job.

A great example of such a role-based approach is

telecom firm BT. The role on which

at the global

siderable effort

is

fifteen

transaction knowledge less

work

thousand. While this

ability

(typically

more on improved customer

Centers. BT’s goal

is

to

make

available

in real time while the

key design

criteria for

that focuses

on

an example of a

was

in call-handling

service through better avail-

BT implemented

“BT AdvisorSpace,” within

knowledge

tal’

measured

of relevant information and knowledge.

role-specific portal,

is

process, the focus for these workers

on increased productivity

times) and

has focused con-

the “customer contact” worker or “advisor,” of which

company employs

the

BT

all

customer

is

its

a

new

Customer Contact

needed information and

on

One

the phone.

of the

AdvisorSpace was to create an interface or ‘por-

delivering the information

and functionality the

advisor requires, as opposed to forcing the advisor to find the content via help is

files,

intranet

and paper documents. Eventually the goal

sites,

to bring the relevant information to the screen automatically based

on the context of the current customer transaction

(i.e.,

to

move more

in the direction of the Partners order entry system).

The new system has already crease in the

number of customers

and knowledgeable in the

led to a several-percentage-point in-

(it’s

at

97 percent now). The advisors’ confidence

information they use

improvements

is

in call-handling times.

mation resources on a particular As with Partners, BT focused

The BT example it

focuses

been

illustrates

what

select a role that

its

efforts

and

infor-

role. its

efforts

on

a single job.

transform every knowledge work role

need either to

also

up by 23 percent. There have

an organization can accomplish when

sible to

was helpful

feeling that their advisor

is

critical to its

at

It’s

not pos-

once. Organizations

mission (physicians

at

102

Thinking for a Living

Partners, for example) or very

numerous and expensive

center

(call

representatives at BT).

Automating Decisions The shortage of managerial time and

analytical expertise that hindered

may be behind

the rise of decision support

the rise of a

new trend

that

holds the promise of realizing that dream, at least to a greater degree.

With

today’s lean organizations, few

knowledge workers have the time

to delve deeply into data analysis or learn the intricacies of a decision

support system (DSS). Instead of employing a DSS, tions are beginning to ask the system to

Automated decision-making systems industries

made

and applications, and

decisions at least

up

to the

make

many

organiza-

the decisions for them.

are penetrating a

wide variety of

human-

are taking over previously

middle management

level.

As

I

men-

tioned above, they also tend to be appropriate for middle levels of expertise

and collaboration. With

this

speed decision making, and lower the

number of highly educated and

expensive decision makers needed. This hold, for example, in “yield

expanding

is

not a

new idea

management” systems

automated pricing decisions for the idea are

approach, organizations can

in the early 1980s

significantly.

—

it

first

in airlines that

and

made

—but the applications

Sometimes

called “in-line” or

“embedded” decision support, the concept might be described intersection of decision support

took

artificial intelligence,

as the

or the “in-

dustrialization” of decision support.

After the success of yield

management, automated decision making

then became pervasive in the financial services industry and

most

common

there.

An

increasing

cial services is available online,

amount of information

which makes

it

is still

in finan-

possible to integrate

and

analyze the information in more-or-less real time. In investment banking, these systems

and online information are behind the

rise

of pro-

103 Organizational Technology for Knowledge Workers

gram trading of

and other

equities, currencies,

financial assets. For

most consumers, the primary impact of automated decision making

realm of credit approval. Credit scores, such as those from Fair

in the

(known

Isaac Corporation

deny

is

as the

FICO

score), are

used to extend or

credit to individuals applying for mortgages, credit cards,

is-

suance of telecommunications accounts, and other forms of debt. Al-

though FICO-based simplistic,

hind the

it

credit scoring has

has certainly

made

been

the process

more

efficient,

availability of near-instant credit decisions.

tion information

is

home

and

is

be-

Housing valua-

making

also increasingly available online,

online mortgage and

being overly

criticized for

possible

equity loans in near-real time.

For example, Lending Tree, a marketplace of lenders for mortgage

and other types of loans, uses automated decision-making technology to decide

which lenders might be best suited

mortgage. Using seventeen different

based on the likelihood they either their

licensed five

criteria,

will close

on

from Lending Tree

four banks are selected

a loan.

own automated decision-making

Then

the banks use

technology or software

make an immediate

to

consumers a

to offer

decision (within

minutes) on whether to offer a mortgage to the consumer and

what

rate

sumer

and terms

to

do

so.

Lending Tree guarantees that the con-

will receive all offers in

one business

day, but they typically

come within minutes. Not only is the process much more that used that the

at

efficient

than

by the typical mortgage broker, but Lending Tree has learned

consumer

is

10 percent

more

likely to accept a

loan

when

it is

offered immediately. In financial services, automated decision

making

is

being used for a

broader variety of applications than just credit decisions. Citibank, for example, uses the technology for automated dispute resolution of credit card accounts.

to calculate

Mortgage banks have created automated systems

nonstandard loan terms; a schoolteacher, for example,

could get a loan that

is

only repaid during the school year. Most large

insurance companies use the technology to underwrite most

life

insur-

104 Thinking for a Living

ance policies, and some are beginning to employ insurance as well. Other firms have begun to use

ance to the mix of an investment portfolio.

tomated system to In

consumer

assess the risk of

credit

and

IBM

collections decision

no

bill

worthy of

is

using an au-

making, several firms

—

are beginning to use

—or whether customer with be —should not a

criteria, a cell

a past

phone com-

customer with dubious credit

for example, that a

a

a yes-or-

really

With more complex decision

pany could decide,

manage compli-

Credit

should be denied further service

decision.

is

a pay-in-advance account, if not a regular credit account.

Similarly, a

good

to

move beyond binary decisions. Whether

tools to

person should be extended credit

due

it

for small-business

entire credit portfolio.

its

most notably in the telecommunications industry automated decision

it

customer

who

misses a single payment with an otherwise

credit history should be treated differently than a

customer with

a history of difficult collections.

Now automated of other industries.

decision

Some

making

is

penetrating into a wide variety

of the U.S. Middle Atlantic-area

utilities that

avoided the

summer 2003

to avert the

problem through automated decision making.

trial

electrical

equipment manufacturer

tax implications of various

is

blackout claim that they were able indus-

using the technology to determine the

equipment contracts and

vices bills for maintenance. In insurance,

are being used to process claims

An

to calculate ser-

automated decision systems

and underwrite insurance

policies. In

health care, they are being used to determine reimbursement levels. In travel

and transportation, where

large U.S. airlines such as

yield

management once helped

American fend off

less technically sophisti-

cated discount airlines such as People Express, automated pricing sys-

tems have become pervasive (no longer conferring advantage on the financially-hurting large carriers).

the pricing of hotels

employed ers

and

zon),

and

The same

rental cars.

for other types of products

electronics (at Dell

new car promotional

tools are also

Automated pricing and

Computer,

is

used in

also being

services, including

for example),

offers (Ford),

now

comput-

books (on Ama-

and even apartment

rents.

105 Organizational Technology for Knowledge Workers

Often these automated decisions are made within the context of a broader business process that

is itself

automated. “Decision engines” or

“business rule engines” for automated decision making are increasingly being

embedded within business

process

management (BPM)

technology that orchestrates the entire workflow for a business process.

Some

observers

call this

process “smart

BPM.”

If a

bank, for example,

were using the technology for credit card dispute resolution,

it

could

not only manage the process from cardholder to bank to merchant and

back again, but also embed automated decision making about

much and when Of

to bill the customer.

involve

some

decisions, or a sampling of them. In

many

course, these systems

human

—

review

either of

all

and processes can

cases, particularly difficult cases are kicked

tem

human

to a

expert,

and experts

mated decision systems and constraints of time will

probably

how

mean

out of the automated sys-

needed

to help build auto-

refine the rules they use.

and expertise that few

are also

still

But the same

that limited decision supports rise

humans will be looking over the

shoulders

of automated decision systems. This will undoubtedly lead to considerable changes in ties,

far

and

how

in the labor

organizations view knowledge-intensive activi-

market for analysts and midlevel managers. Thus

automated decision making has been

but also

it

may

is

largely invisible to the public,

lead to a quiet revolution in organizations

not without

risk:

societies. It

automating poor decision processes can

quickly get a firm into trouble, and managers

problem

and

may

not recognize the

until there are substantial losses.

Other Types of Knowledge Worker Software In addition to

embedding knowledge

in

work

processes, performance

support, and automated decision making, there are a variety of IT applications that are intended to

mance. These

fall

improve knowledge worker perfor-

into a few specific categories, however,

and

are

106 Thinking foT a Living

unlikely to be applicable to a broad range of knowledge worker perfor-

mance

issues.

One

category

is

software for knowledge workers.

role-specific

These applications support a particular role that cuts across several different industries. Call center workers, for example, have at their dis-

posal a broad range of technologies, though they are likely to have been

chosen and implemented by others, not by the workers themselves. This little

call

is

one aspect of low-level, transaction-oriented work: there

just

discretion about

what

tools to use in

performing the

center agents, for example, don’t have access to e-mail

ternet

from

The

Most

and the

In-

their office computers.

applications for call centers include customer relationship

management

software, tools for scripting conversations with customers,

knowledge tools

for solving

ing customer feedback. crease the

volume of

somewhat

less

customer.

Some

humans

job.

is

customer problems, and tools for captur-

The goal of these applications calls that a call

is

typically to in-

center agent can handle,

and

often to increase the quality of service provided to the

organizations want to go even further and eliminate

altogether

from

call centers,

hence the

rise

of interactive voice

response and other customer self-service technologies.

At the other extreme of role-specific technologies are tools for scientists in

icals,

pharmaceutical, medical equipment, chemicals and petrochem-

and environmental

firms.

Such tools

as electronic lab

notebooks

(not necessarily notebook computers, but rather software for capturing the results of laboratory experiments)

management systems have been

and laboratory information

available for

many

years,

but these

high-discretion workers have generally been given latitude as to whether

and how the applications are used. lab notebook, this

was

largely tolerated.

edge gathered were viewed as the didn’t matter in

If a scientist

what format

was generally productive.

it

wanted

to use a paper

The information and knowl-

scientist’s

was gathered

personal property, so

—

it

at least if the scientist

107 Organizational Technology for Knowledge Workers

More

recently,

however, companies have begun to

As laboratory documents become

tools be used in a consistent fashion. legal

documents, and

comes more

critical to

insist that these

as laboratory

information and knowledge be-

R&D and regulatory processes, firms are discov-

ering that they can’t leave the use of laboratory applications to the scientist’s discretion. Infinity

Pharmaceuticals, for example, a

Cam-

bridge, Massachusetts, drug development start-up that employs

approaches to chemistry and genetic screening, mandates that entists use electronic lab

notebooks, and that they

tion available to everyone in the

make

company. These

new

its sci-

their informa-

tools,

along with

other scientific and analytical applications, have been combined into the “InfiNet Knowledge Platform,” which

intended to provide a

is

broad knowledge capture and knowledge sharing capability for the

company and ductivity

its

research partners. As the importance of scientific pro-

and knowledge sharing increases

likely to see

in this type

of firm, we are

more mandated use of the previously voluntary

solutions.

Other technologies are more experimental, and not yet of

clear

value in the goal to increase knowledge worker performance. However,

they offer the promise of enabling

and

new knowledge worker

functions

applications.

Social networking software,

which

I

will describe in chapter 7,

one such category. This technology is intended

to

is

enhance the function

of social networks both within and across organizations. Certainly these tools

remind us that knowledge worker performance

an individual

effort; ideas

and

working together. However,

their execution

if it’s difficult

the performance of individual workers,

termine

how well

social

way from knowing how

to

it’s

all

derive

is

not only

from people

measure and understand

even more

networks are performing

difficult to de-

overall.

to assess the productivity of

We’re a long

networks and the

value that networking technologies bring to them.

There are other forms of “socialware” cial

relationships

—

that

some view

as

—software

that supports so-

important to the future of

108 Thinking foT a Living

knowledge work. Academics have studied

this category for years, in-

cluding technologies for finding people sharing

common

interests, for

enabling a virtual conversation or discussion, or for group decision

support and decision making. Most such

have proven stub-

activities

bornly resistant to any sort of automation, although occasionally a technology gets broader

One example

is

Web

and acceptance.

visibility

logging or “blogging,” which

is

a

means

for in-

dividuals to record their opinions for others to access. Partisans of

blogging argue that there are

many

potential business applications of

the technology (and they discuss these applications in their blogs!).

But

I

believe that blogging

falls

knowledge worker performance cations are largely hypothetical uals to express their

into the is

—

unproven category

concerned.

at

the

as far as

the business appli-

First,

moment

8

it’s

a tool for individ-

somewhat random musings. Second, I know of no

organization in which the benefits of blogging have ever been measured. Perhaps the biggest to write

and read

for blogging

is

we

it. I

am

all

for

phenomenon with

shouldn’t confuse the

that his

own

managing

his

my friend and

blog

former colleague

—and those of some others—

own

caught on broadly,

personal knowledge. it

could represent a

it

takes

from pro-

freedom of expression and

knowledge worker performance. One potential point ging comes from

the time

blogs. If anything this tool has detracted

ductivity, not increased

publishing, but

problem

increased

in favor of blog-

Bill Ives,

is

self-

who

argues

really a vehicle for

If this particular

new approach

use of blogs

to organizational

knowledge management.

Summary Organizations need to strike a balance with

knowledge workers. They need technologies,

and learn what

to

new

technologies for

experiment and tinker with new

their potential benefits

might be for en-

109 Organizational Technology for Knowledge Workers

hancing performance. But

if

they are to be used for business, a hard-

nosed attitude should be adopted

fairly quickly.

What’s the value?

should any improved performance be measured? the cost

— not

hardware and software, but

just in

and

to learn, tinker with,

Is

fix

How

the payoff equal to

in the

time required

the technology? Ultimately, any evaluation

of knowledge worker technologies will require close observation of

how

the technology

new technologies

is

fits

labor-intensive,

performance payoff is even more

and understanding

their value

and

ganizational level, for organizational processes is

been

at the or-

objectives.

subject of chapter

knowledge worker

level,

and

this

is

6.

Recommendations for Getting

Results

from Knowledge Workers Information technology for improving knowledge work can be divided into two types: organizational and personal. •

Under the banner of “knowledge management,” many organizations have created knowledge repositories to aid knowl-

edge workers in jobs, but few workers have the time needed to

•

browse and learn from repositories.

An important alternative to embedding knowledge process. This if

But

another world of technology and performance interventions

that operate at the individual

•

and

so.

All of the technologies discussed in this chapter have

there

and using

into the context of the job. Learning

done

well

is

—

repositories

into the

difficult to do,

as in

some

is

the idea of

knowledge worker’s job

but can be very rewarding

health care processes.

the

110 Thinking for a Living

•

Performance support and role-specific portals are ways to reduce the amount of searching and browsing for timepressed knowledge workers.

•

Well-structured decision processes are increasingly being

automated. In edge workers often

•

still

many cases,

is

the

work of entry-level knowl-

being done by computers, but experts are

needed.

There are many other categories of organizational applications for

knowledge workers, including

social

networking

software and blogging tools. These should be the subject of

experimentation within organizations, but they have yet to

•

demonstrate clear business value.

Motivation to pursue knowledge technologies (such as the prevention of medical errors)

is

an extremely important

factor in successful implementation.

6

Developing Individual

Knowledge Worker Capabilities

Most interventions

to

improve performance

ganizational or process level, but also

improve individual

it

in business are at the or-

doesn’t have to be that way.

capabilities. Ultimately,

We can

knowledge worker

performance comes down to the behaviors of individual knowledge workers. process,

If

we improve

their individual abilities to create, acquire,

and use knowledge, we

are likely to

improve the performance

of the processes they work on and the organizations they work

Of course, some also can

I’ve

described

improve individual performance, so what’s the difference? In-

dividual knowledge First,

of the organization-level initiatives

for.

work improvement

initiatives

have two attributes.

they are directly focused on improving performance of knowl-

edge worker employees as individuals, not as members of a larger group.

A CRM a

program

for

number of people

customer service workers doesn’t

in that function

would use

it,

qualify,

because

and the system

is

not

112

Thinking for a Living

(or at least rarely) customized to individual needs. Second, individu-

oriented initiatives are targeted at improving

ally ity,

rather than instituting a

new

process. Giving

piece of hardware or software

phone

cell

new

—wouldn’t

qualify,

—

say, a

some

skill

or capabil-

knowledge workers

a

personal digital assistant or

but teaching them

how

to use these de-

vices effectively would. I

became persuaded of the

when doing some work with

capabilities at the personal level

ware Engineering its

knowledge worker

virtues of improving

Institute at Carnegie-Mellon.

The SEI

is

the Soft-

famous

for

Capability Maturity Model, an assessment tool for software engi-

neering processes that

I

described in chapter

evaluates firms or

3. It

business units on their overall approaches to software development.

But Watts Humphrey, the developer of the sight.

He

realized that

move up through what might

it

was taking too long

five stages

of the

accelerate the process.

would probably improve much and

CMM, had another key in-

CMM,

for

many organizations to

and began

He concluded

faster if they

to think

about

that organizations

were to develop team-

individual-level capabilities in addition to those at the organiza-

tional level. SEI’s research has

borne out

this hypothesis.

Companies employing

the personal software process (PSP) and the team software process (TSP)

have been

known

to

move from

development maturity

in

the lowest to highest levels of software

about a year

—versus an average of

close to

ten years for this journey using only organization-level approaches.

What

1

Kinds of Capabilities Do

Knowledge Workers Need? The

lessons of the

edge workers.

PSP

Some

aren’t,

of course, relevant to

all

types of knowl-

individual capabilities are process-specific. In the

113 Developing Individual Knowledge Worker Capabilities

PSP, for example, software developers are taught and assessed

should focus on such capabilities as interview-

Similarly, a consultant

and

analytical skills.

But there are also generic knowledge worker

skills

that almost

everyone employs, and that could benefit from improvement. all

their

measure, deal with data, and handle defects.

ability to estimate, plan,

ing, presentation,

on

knowledge workers do? They read and

write, of course,

What do

and our ed-

ucational systems do a pretty decent job of inculcating these

skills.

Even that doesn’t stop some assiduous knowledge workers from taking courses in speed reading, business writing, or the

like.

No doubt more

of this should be done, but given the available resources for building this capability, there

is

no need

Knowledge workers

also

to discuss

spend a

ganizations, of course, don’t

do

lot

a very

ployees run meetings effectively.

A

it

further in this book.

of time in meetings. Most or-

good job

at

helping their em-

few, like Xerox, have instituted

organizationwide programs focused on maintaining a high quality of meetings. Again, there are plenty of written materials and educational

options for people

ment, so

I

who want

and knowledge

and

writing,

about

more about meeting manage-

won’t say anything more about

Increasingly, however,

sages,

to learn

—on

it

here

knowledge workers

2 .

also process information

paper, in telephone conversations

electronically. This subject

and meeting, and there

is

is

and voice mes-

much newer

relatively little

than reading,

information available

how to do it well, or how organizations can help their knowledge

workers do

it

well. In this chapter

better understand this subject.

companies seeking

to

I’ll

report

on

three research efforts to

Two were undertaken by

also report

on more

detailed interviews of individuals

to be very effective in their

management.

group of

understand information work; both corporate

and individual-level research projects were undertaken by I’ll

a

own

this group.

who

claim

personal information and knowledge

114 Thinking for a Living

Personal Information

and Knowledge

Management for Knowledge Workers In 2003 the Information

Work

Productivity Council (IWPC), a con-

sortium of technology and IT services vendors, funded a number of research projects

management

on the

topic of personal information

(see “Research Approaches”).

searchers participating in this project ficient

knowledge on how

knowledge workers, and

The companies and

realized that there

was

that this topic

all felt

3

One compelling

that information workers (people

who

a

major topic

reason for this was

use technology and

information in the context of their jobs

and

—personal information

—would very soon become

by businesses.

re-

insuf-

to get productivity out of information

and knowledge management to be addressed

had

and knowledge

—perhaps

a

work with

somewhat

larger

category than knowledge workers) were spending ever larger amounts

of time (more than three hours per day by our data) messaging, creat-

and knowledge, and other

ing documents, searching for information

information-intensive

activities.

Despite this large time commitment, information workers have

been mostly

left

to their

their organizations in

tasks effectively

and

own

devices, so to speak, with

little

help from

how to perform key information and knowledge

efficiently.

And

those devices, or the technologies

used for handling personal information and knowledge, have been largely separate

wireline

and

and unintegrated. Thus

cell

still

our desktop PCs, laptops,

phones, PDAs, handheld communicators, and other

— not mention the paper-based many employ— have been unconnected. At the same

assorted technologies dividuals

far

to

tools

in-

largely

time that we face increasing technological challenges in managing personal information and knowledge, few individuals can be said to be well educated their jobs in

and well informed on how

an optimum fashion.

to use the tools to

perform

Developing Individual Knowledge Worker Capabilities

Research Approaches chapter

his

d r aw s from three studies of personal

infor-

mation and knowledge management. Two were performed under the auspices of the Information a

Work

Productivity Council (IWPC),

consortium of technology firms founded

to carry out research

and

educational initiatives in the areas of understanding, measuring, and

enhancing information work productivity. In 2003, the year the

IWPC

was founded, sponsoring firms included Accenture, Cisco Systems, Hewlett-Packard,

mic

director,

In the

first

in large

Microsoft, SAP, and Xerox.

and collaborated with

of these studies,

I

was the acade-

company representatives.

several

we interviewed twenty-one managers

companies and two government agencies who were

ested in the issue. cally

Intel,

The

particular

inter-

managers we interviewed were

typi-

knowledge managers, managers of new technologies, and IT

managers who

dealt with personal productivity tools for their orga-

nizations.

The second study was focused on how

individuals were dealing

with personal information and knowledge, and involved just over

500 U.S. -based information and technology

users.

These individuals

We then reduced this whom had access to a

volunteered to complete a Web-based survey.

sample to 439 qualified respondents,

computer and e-mail

at

who

of

work, spent some time during the week pro-

cessing work-related information,

The

all

and used e-mail

at least weekly.

third study involved hour-long interviews with ten individuals

reported that they were highly effective managers of their per-

sonal information and knowledge environments.

of jobs across

many industries.

They had

a variety

116

Thinking for a Living

Working with these devices

to

manage

formation and knowledge, however,

personal, work-related in-

increasingly

is

what people do

within organizations. Its not hard to believe that with better technology, better education,

and

better

management, the key tasks

that infor-

mation workers perform within organizations could be done with greater speed

tion have

them can

and

become

quality,

and

at

lower

cost.

Technology and informa-

so closely integrated with

easily create

more

effective

and

work

that better use of

efficient organizations.

Information Manager Findings Information managers showed considerable variation in their orientation to personal information

and knowledge management, with some

companies already treating

as

erable attention, in

some on

it

the road to that status, and

roughly equal proportions.

will

an important issue worthy of consid-

I

some unaware

suspect that the adoption of the idea

mirror that of other business and management innovations, and a

focus

on personal information and knowledge management

tually spread

learn

and become much more pervasive. But there

from organizations that are addressing the

The leading-edge companies

— found

is

much

information and

knowledge-intensive information technology, pharmaceutical, and nancial services industries that they

—exhibited

fi-

a variety of traits suggesting

were focused on personal information issues (see figure 6-1

for a graphic display of the orientations to these issues).

ready actively dealing with personal information specific initiatives to address productivity

ogy. Cisco Systems, for example,

Work”

to

issue today.

the

in

will even-

initiative for

behavior changes for

how to

al-

management with

through the use of technol-

had begun

a

“Change the Way

employees, which involved a

technologies, education in

Some were

recommended

We

set

of

use them, and a set of recommended

optimum information-processing

effectiveness.

117 Developing Individual Knowledge Worker Capabilities

FIGURE

6-1

Company orientations to knowledge management

personal information and

address productivity in corporate initiatives Go beyond just technology to usage and behavior Provide an integrated approach to support

• Explicitly

Already there

• •

•

On

the road

•

•

No awareness or activity

•

Rely primarily on technology Address knowledge, but not holistically Focused on other issues

Not aware of the issues technology Fragmented support

• Little

•

Capital One, the financial services firm, had a broad initiative under-

way

to

improve individual-level productivity with technology

tioned their approach to experiments in chapter in this category

had similar programs under way,

men-

(I

Other companies

3).

either for

all

employ-

ees or for a particular subset.

One

of the earliest adopters of these approaches

created an “eWorkforce” initiative

is Intel,

which has

composed of three previous

separate

groups addressing knowledge management, collaboration, and personal productivity.

The eWorkforce group has determined

use of these technologies ers are aggressive users

is

a pressing

and spend

problem

large

that better

for Intel, since

its

work-

amounts of time doing

Sixty-three percent of Intel employees participate in

more than

three

teams; 62 percent routinely collaborate with people from different or regions; 40 percent regularly

work with people who use

collaboration technologies and tools, and

people

who

use different

work

processes.

based collaboration meetings and ferences every week. laptops, cell phones,

more than

half

so.

sites

different

work with

Employees conduct 8,300 Web-

dial in to

roughly 19,000 audio con-

The group supports knowledge worker use of PCs, and PDAs, and

is

developing integrated solutions

118

Thinking for a Living

for “generic”

knowledge worker processes

—

and

tasks like arranging

conducting an asynchronous meeting or managing a project. Deliverables

from the

first

phase of eWorkforce

(just

concluded

as

I

write)

included a consolidated collaboration platform, a standard project-

management

platform, next-generation meeting management,

generation presence availability)

and

(graphically indicating presence

and

instant messaging. In this phase, Intel also addressed

the development of

methods

management

first-

common

for collaboration

processes and the use of best

known

and group work.

These leading-edge organizations were making heavy use of emerg-

commu-

ing technologies such as instant messaging, PDAs, handheld nicators,

and shared document

repositories.

wasn’t just on technology, but also on

hind the success or erally

failure

use and the

human

issues be-

of technologies. These companies were gen-

making some attempt

to

Informatics and Knowledge vartis’s research

its

However, their focus

change user behaviors and cultures

Management

organization within

Knowledge Culture.” Others were using technology

itself to

specialists

No-

guide the

changes in behavior. The support groups for individual users

were not

the

Head of

group, for example, had created a “Global

firms, like Intel’s,

—

at these

by type of technology, but had

a

holistic focus.

Other companies we interviewed were facing challenges with personal information and knowledge and were aware of them, but hadn’t yet formulated a holistic response. to a focus

I

view them

as

being “on the road”

on personal information and knowledge management. They

were using some of the same emerging technologies edge organizations, but the usage was

less

as the leading-

monitored and managed.

There was a strong orientation to technology products

as a

dealing with personal information (“Our major project

from Lotus Notes

to Microsoft Outlook”), but less focus

is

means of changing

on the use of

those tools. There was generally no holistic support group for users of

personal information, but in several cases a

community was beginning

119 Developing Individual Knowledge Worker Capabilities

to

emerge across the relevant functions. In

several cases,

some major

technology or business issue seemed to be preventing a focus on individual productivity, but discussion of productivity at a broader level

was taking place within the company

The companies

in the third

group were somewhat interested

in the

topic (or they presumably wouldn’t have taken the time to participate in the interview process), but

sues

had

—economic

some were primarily focused on other

survival, for example.

None of

is-

these organizations

important enough to

really identified individual productivity as

address with any seriousness, and generally did not recognize

as a

it

corporate issue. They had no formal group to support even the basics

of knowledge management or individual information use.

What

sup-

port they did provide to individual users was very fragmented by tech-

nology type.

Little training

or education was offered to users, and what

was offered was product-specific. These organizations made

little

use

of emerging technologies for personal information and knowledge or

even discouraged their use; several specifically banned instant messaging, for

example. Several stated apologetically that “we

be doing more in

made

this area,

but there

is

just too

much

know we should

else

going on,” or

similar remarks.

Measuring and Increasing Individual Performance The

early adopters’ focus

on

individuals wasn’t just a technology issue,

but also involved an emphasis on productivity. ductivity tools

and approaches, however, were

nomic motivation was surveyed

—even

certainly present.

the successful companies

Some

necessary pro-

lacking, although eco-

Many

—were

organizations

we

facing lower margins,

reduced headcount to do the same work, and so forth. Most had some corporate orientation to increasing knowledge worker productivity,

and technology and messaging

tools

were seen

as vehicles to that end.

120 Thinking for a Living

However, several firms complained that there were no good approaches to the measurement of productivity Several managers ductivity

(e.g.,

commented

implement

a

at the individual level.

that the “time saved”

new

tool,

approach to pro-

and save thirteen minutes per

day per knowledge worker) was no longer credible in their organizations. Productivity

as yielding

improvements

were viewed

at the individual level

only incremental benefits, and for companies not on the

leading edge, such investments couldn’t compete with other categories

of needed spending.

had

Several firms

a strong “self-service”

approach

to getting func-

whether

tional tasks accomplished, but a few respondents questioned

that strategy

was

work

just shifting

measure the impact. For example, perform

ers to

all

their

to points

if a

own human

where

it

company begins

is

difficult to

to ask

its

work-

resources transactions (choosing

benefits providers, changing addresses, checking vacation balances, etc.),

but

HR

the cost reductions in the

function can easily be measured,

how does it affect the worker’s productivity? This problem

be studied in

much

greater detail to determine whether

it’s

needs to

really help-

ing organizational performance. In order for firms to begin improving the

management of personal

information and knowledge, they also have to begin to change the behavior of users. There was considerable variation, even leading-edge organizations, about best be created.

As might be expected,

proach to behavior change

through the software ration

how

it

is

for example, Microsoft’s apelicit

workers, for example, should be handled by inter-

this as a

eWorkforce group, however, takes an ap-

proach centered on customized is

more human

some Microsoft people would view

fault in the software. Intel’s

Cisco

needed behaviors

the marketplace. All aspects of collabo-

Microsoft’s SharePoint collaboration software. If

ventions are necessary,

the

behavior change might

generally to try to

sells in

among knowledge

this

among

tools, process consulting,

primarily focused on training as a

and job

aids.

means of creating behavior

121 Developing Individual Knowledge Worker Capabilities

change. ferences

It is still

too early to understand the implications of these

and what might comprise

behavior.

The approaches of

been driven thus

far

a “best practice” in

really

There

is

more by the

company and

culture of the

by an empirical

also

little

consensus on

how

to

analysis of

segment information and

for differential treatment.

Most leading-edge

however, seem to recognize that they cant treat alike,

mentioned

the expe-

works.

knowledge workers

workers

changing user

these leading-edge organizations have

rience of the offering group, rather than

what

dif-

and are beginning

Intel’s

all

firms,

their information

segmentation approaches.

to create

segmentation approach in chapter

At

2.

I

least three

other organizations have an implicit or explicit segmentation by role identifying particular roles

enough

to justify

and jobs

an aggressive

that were

effort to design

an information and

Some

other organizations,

knowledge environment around the

role.

including information storage leader

EMC

pany

and the engineering com-

MWH Global, are creating a taxonomy of roles within their orga-

nizations so that information role-specific basis with

Xerox, ties

numerous or important

among

others,

some

is

and knowledge can be

precision.

A

delivered

third approach,

on

a

employed by

the development or recognition of

communi-

of practice, and the creation of information and knowledge envi-

ronments that support those communities. As with behavior change, segmentation

proach works

is

in the early stages,

some

it’s

manager surveys confirmed

management

and those

attention. Firms that

in industries in

agement and knowledge

which

real,

and worthy of con-

make and

effective personal

sell

technology,

information man-

are critical to success, are believers in the idea

and are already addressing

moving

that at least

organizations, the problems and opportunities of personal

information and knowledge management are certed

not yet clear which ap-

best.

Overall, the information for

and

in that direction, or

it

with

initiatives.

Other firms are either

not doing anything

at

all.

No

matter

how

122 Thinking for a Living

advanced on the

issue,

almost

all

companies are encountering

issues of

productivity measurement, behavior change, and user segmentation.

Information User Findings information manager survey showed that companies vary

Just as the

widely in their approaches to personal information and knowledge

management, our Web survey of information users

also revealed a high

degree of variance with regard to these issues. The intent of this survey

was

to discover the behaviors

and

attitudes of typical users of informa-

tion technology at work, with particular emphasis

information distribution technologies. These

on messaging and

activities are

obviously

of importance to individuals and firms, since, as already noted, the av-

more than

erage person interviewed spent

— and

day

There

it is

is

likely that these

also

no doubt

that

numbers

will

some people

of several questions in the survey

—saw

three hours

on them each

only increase over time.

—about 20 percent on each

a substantial

problem with

their

personal information and knowledge management. This fraction of individuals

felt

overwhelmed by

their information flow,

saw too much

use of e-mail in their organizations, and viewed e-mail and other technologies as hindering rather than helping their productivity.

On

each

of these issues the remaining 80 percent saw no real problem, although there were considerable differences in

how much

information they

ceived and the media they used. Overall, few respondents

up

their It

give

messaging technologies, but some were frustrated with them.

may be

nology

would

is

surprising to see just

how much

information and tech-

used by the respondents of our survey. The average user in

the survey: •

re-

Spent 3 hours and 14 minutes a day using technologies to process work-related information

—

just over

40 percent of an

8-hour workday (see figure 6-2 for allocation)

123 Developing Individual Knowledge Worker Capabilities

•

Devoted

1.58 hours/day to e-mail (45 percent of the information

processing [IP] time, and 20 percent of an 8-hour day) •

Spent 47 minutes, or 24 percent of IP time, on telephone, conference

•

calls,

and voice mail

Received 44 e-mails daily (4 people said they received around

500 a day!) •

Sent 17 e-mails daily

•

Had more

•

Received 16 instant or text messages a day (for respondents using this

than 3 e-mail accounts

technology) got 7.6 voice mail messages

•

Received 18

•

Participated in 2.75 conference calls a

FIGURE

calls,

placed 15

calls,

week

(if

any)

6-2

Average percentage of information processing time spent on each medium (based on average hours spent)

124 Thinking for a Living

FIGURE

6-3

Percent of typical users (N=439) using each

All

medium weekly*

use e-mail on a weekly basis.

The respondents’ use of

different

messaging and information

dis-

tribution technologies also varied widely. Figure 6-3 displays the fre-

quency of use of the various media. User attitudes toward their personal information flow and various

media revealed further their information

differences. 17 percent felt

management and

flow, but this

appeared to be more

an issue of attitude than of volume of information. tion, this

overwhelmed by

On closer examina-

group didn’t receive more e-mails than average and didn’t

spend more time processing work-related information during the day. However, they felt

that e-mail

felt

was

they were less

less likely to believe that

managing information. in control

less effective at

managing information,

valuable to their work, and were substantially

they received help from their organization in Just

under half of the sample, 49 percent,

of their information

management and

flow.

felt

125 Developing Individual Knowledge Worker Capabilities

Despite their complaints about personal information management,

sample

this is,

many

illustrates

felt

less effective.

but also to the

The

—

that

more

informa-

effective at personal

others they knew; only

1

1

percent

felt

they were

This finding speaks not only to the self-confidence of

tion environments.

invisibility

of

how we manage

our

own

informa-

We simply don’t know how other people do

it.

overall lack of orientation of users to personal information

management

We

that they were

management than

users,

effect”

believed they were better than average. Forty-one percent of

our respondents tion

Wobegon

Garrison Keillors “Lake

is

suggested by an open-ended question in our survey.

asked respondents what one thing they would change in their per-

sonal information environments, and then categorized the responses.

The most common responses were “nothing,” with 16 “don’t know,” with 13 percent.

percent would reduce

Among more

spam or pop-up

ads,

The

large

and

number of uninformed responses

viduals have not thought very

much about

The survey tions helped

them manage

most

this issue thus far

own

indi-

—and

personal informa-

suggested.

what degree

4

their organiza-

their personal information flow. Forty-one

percent said that they received tions in

limit

idiosyncratic to report.

—which other researchers have

also asked respondents to

would

7.5 percent

suggests that

that they have probably underinvested in their

tion environments

and

substantive responses, 11

amount of e-mail. Other answers were too

the

percent,

little

or no help from their organiza-

managing personal information; only

3 percent felt that their

organization had completely mastered the problem of personal infor-

mation management. This confirmed nizations have a long

of

way to go before they have

may

issues.

However, individuals

and

since they aren’t getting

can,

in the

my expectation feel that

much

that

fully dealt

with

they are doing

this set all

they

help from their organizations,

absence of any direction or contrary evidence, they

are doing fine.

most orga-

may feel they

126 Thinking for a Living

Attitudes Toward Specific

Media and Technologies

E-mail was one of the most frequently used media in the study, and

one of the most problematic

also

sample, 26 percent 10 percent

felt it

felt

received

level, 21

and

very or extremely valuable to their

their

percent

felt

it

was being

overwhelmed by

sent, 41 percent felt that e-mail

was

work performance, and only 4 percent

was not valuable. Although 15 percent

work

their organizations;

was underused, and 64 percent believed

amount of e-mail

felt it

terms of negative attitudes. In our

was overused by

that e-mail

properly used. At the individual the

in

productivity, 53 percent

felt

that

felt

that e-mail diminished

increased

it

it.

On balance,

the responses of the users were positive about e-mail, but less positive

than for some other messaging and information distribution technologies.

For the roughly 20 percent reporting a significant problem with

e-mail, the key question

or they are just

is

whether they are

more conscious of its

ineffective at

titudes than about e-mail,

—

elicited

amount of telephone 49 percent, to their

felt

calls

—

actual telephone calls,

somewhat fewer negative

and somewhat more

Only 12 percent (versus 21 percent

it,

negative effects than others.

Survey questions involving the telephone voice mail, and conference calls

managing

for e-mail) felt

and messages they

at-

positive attitudes.

overwhelmed by the

received.

Almost

half,

or

telephone information to be very or extremely valuable

work performance

—

8 percent

dents assessed the impact of telephone

more than calls

for e-mail.

and messages on

Respon-

their pro-

ductivity similarly to that of e-mail, with 15 percent concerned that the

telephone diminished their productivity, and 50 percent believing

enhanced felt

their productivity.

About the same percentage (14 percent)

that voice mail diminished their productivity; 40 percent (10 per-

cent less than telephone

calls) felt

it

enhanced productivity. Only 4 percent

were overwhelmed by the number of conference in,

it

but a relatively low percentage (35 percent)

were very or extremely valuable to their work.

calls

felt

they participated

that conference calls

127 Developing Individual Knowledge Worker Capabilities

was apparently

Instant messaging (IM)

still

an emerging

medium

Among our respontheir organizations. Among

for this sample, with lower overall perceived value.

dents, 56 percent didn’t yet have

those

who

IM

use in

used the technology, 29 percent

felt

that

sages were very or extremely valuable to their

IMs and

text

mes-

work performance

substantially less than for e-mail or telephone-based technologies. slightly higher percentage

than for other technologies (18 percent of

those

who use IM)

felt it

enhanced productivity

felt

A

that

it

diminished their productivity; 35 percent

—

again, lower than for other technologies.

Perhaps the most popular technologies were those involving infor-

mation and knowledge “pull” rather than “push.” These are corporate

Web sites, information portals, and document sharing systems. Those who used these tools reported low levels of being overwhelmed (4 percent)

and diminished productivity

(4 percent), with high levels finding

these technologies very or extremely valuable (47 percent for corporate

Web

sites,

tivity

enhancement

enhanced

and 64 percent

for

Web

document sharing

for

document sharing systems) and produc-

(4 percent report sites;

diminished productivity, 50 percent

3 percent diminished, 67 percent

systems).

The

positive reaction to

enhanced

document

ing was the highest for any technology in the study. As the

for

shar-

amount of

information in organizations continues to proliferate, these technologies are likely to

become even more

Strategies for The survey

popular.

Managing Information

also included questions

on how

these information users

were coping with the types and amounts of information they received.

The majority of respondents reported using

specific

approaches to limit

or control their personal information. These varied by technology. For

example, the strategies respondents employed for dealing with e-mail are illustrated in figure 6-4.

*

128

Thinking for a Living

FIGURE

6-4

Percent of typical users (N=439) employing e-mail

management

strategies

Check

e-mail frequently

Skip/delete generic

c/>

messages