Hargadon-1997-Technology Brokering and Innovation in A Product Development Firm

Technology Brokering and Innovation in a Product Development Firm
Author(s): Andrew Hargadon and Robert I. Sutton

Source: Administrative Science Quarterly, Vol. 42, No. 4 (Dec., 1997), pp. 716-749
Published by: Sage Publications, Inc. on behalf of the Johnson Graduate School of Management,
Cornell University
Stable URL: http://www.jstor.org/stable/2393655 .
Accessed: 14/01/2015 03:46
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating
with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.
http://www.jstor.org
This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

All use subject to JSTOR Terms and Conditions
Technology Brokering We blend network and organizational memory perspec-
and Innovationin a tives in a model of technology brokering that explains
ProductDevelopment how an organization develops innovative products. The
model is grounded in observations, interviews, informal
Firm conversations, and archived data gathered during an eth-
nography of IDEO,a product design firm. This firm ex-
Andrew Hargadon ploits its network position, working for clients in at least
StanfordUniversity 40 industries, to gain knowledge of existing technological
Robert 1. Sutton solutions in various industries. It acts as a technology
Universityof California, broker by introducing these solutions where they are not
Berkeley known and, in the process, creates new products that are
original combinations of existing knowledge from dispa-
rate industries. Designers exploit their access to a broad
range of technological solutions with organizational rou-
tines for acquiring and storing this knowledge in the or-
ganization's memory and, by making analogies between
current design problems and the past solutions they
have seen, retrieving that knowledge to generate new
solutions to design problems in other industries. We dis-
cuss the implications of this research for understanding
the individual and organizational processes and norms
underlying technology and knowledge transfer more
generally.'
Knowledge is imperfectlyshared over time and across
people, organizations,and industries. Ideas from one group
might solve the problems of another, but only if connections
between existing solutions and problems can be made
across the boundariesbetween them. When such connec-
tions are made, existing ideas often appear new and creative
as they change form, combiningwith other ideas to meet
the needs of different users. These new combinationsare
objectivelynew concepts or objects because they are built
from existing but previouslyunconnected ideas. This paper
presents an ethnographicstudy of a productdesign firmthat
routinelycreates new products by makingsuch connections.
The role these connections can play in the innovationpro-
cess is evident in inventions by Thomas Edison's laboratory.
? 1997 by Cornell University. Edisonand his colleagues used their knowledge of electro-
0001 -8392/97/4204-071 6/$1 .00.
magnetic power from the telegraph industry,where they
first worked, to transferold ideas that were new to the light-
We are grateful to Stephen Barley, Beth ing, telephone, phonograph,railway,and miningindustries
Bechky, Dennis Boyle, Kathleen Eisen-
hardt, Herminia Ibarra,Linda Johanson,
(Hughes, 1989; Millard,1990). Edison's productsoften re-
Christine Oliver, David Owens, Marc Ven- flected blends of existing but previouslyunconnected ideas
tresca, and three anonymous reviewers that his engineers picked up as they worked in these dispa-
for their contributions to this paper. We
are also grateful for the support provided
rate industries.The phonographblended old ideas from prod-
by the Center for the Advanced Study in ucts that these engineers had developed for the telegraph,
the Behavioral Sciences, Hewlett-Pack- telephone, and electric motor industries,as well as ideas
ard, the Stanford Integrated Manufactur-
ing Association, the National Science
developed by others that they had learnedabout while work-
Foundation (SBR-9022192), and the Cen- ing in those industries.Edison's inventionswere not wholly
ter for Innovation Management Studies at
Lehigh University. We especially appreci-
original.Likemost creative acts and products,they were ex-
ate the time and effort that Paul Barsley, tensions and blends of existing knowledge (Merton,1973).
David Blakely, Gwen Books, Dennis As Usher (1929; quoted in Petrovski,1992: 44) argued, "in-
Boyle, Sean Corcorran, Tony Fields, David vention finds its distinctivefeature in the constructiveas-
Karshmer, David Kelley, Tom Kelley,
Chris Kurjan,Bill Moggridge, Chuck similationof pre-existingelements into new syntheses, new
Seiber, LarryShubert, Peter Skillman, patterns, or new configurationsof behavior."
Roby Stancel, Rickson Sun, Scott Under-
wood, Don Westwood, Jim Yurchenko,
and many others at IDEO Product Devel- Social networktheory suggests that Edison's laboratory
opment devoted to this research. could innovate routinelybecause it occupied a "structural
716/AdministrativeScience Quarterly,42 (1997): 716-749

Technology Brokering
hole" (Burt,1992a, 1992b), a gap in the flow of information

between subgroups in a largernetwork. For Edison,these
gaps existed between industrieswhere there was and was
not knowledge about the newly emerging electromagnetic
technologies. Actors fillingthese gaps are brokerswho ben-
efit by transferringresources from groups where they are
plentifulto groups where they are dear (Marsden,1982;
Gouldand Fernandez,1989; Burt,1992a; DiMaggio,1992).
Brokershave an advantage over competitors because "non-
redundantcontacts are linkedonly throughthe central
player,[so brokers]are assured of being the first to see new
opportunitiescreated by the needs in one groupthat could
be served by skills in another group" (Burt,1992a: 70). Edi-
son's laboratoryacted as a brokerof technological ideas be-
cause it had connections to many industries,ratherthan be-
ing central in one, and it linkedindustriesthat had few other
ties (DiMaggio,1992).
By highlightingthe structureof resource flows across group

boundaries,researchers have shown that brokersbenefit
from disparitiesin the level and value of particularknowl-
edge held by differentgroups, but they have not explicated
the process by which informationis transformedor com-
bined within these flows. Valuablesolutions seldom arriveat
the same time as the problemsthey solve, they seldom ar-
rive to the people workingon those problems, and they sel-
dom arrivein forms that are readilyrecognizableor easily
adaptable.Edison's laboratorydid more than just transfer
knowledge from groups where it was plentifulto groups
where it was dear;this organizationacquiredsuch informa-
tion, stored it, and retrievedit to create new combinationsof
old ideas. Walsh and Ungson (1991: 61) described these pro-
cesses (i.e., acquisition,retention,and retrieval)as routines
supportingan organization'smemory, which they defined as
"stored informationfrom an organization'shistorythat can
be broughtto bear on present decisions." This perspective
suggests that a technology brokerdepends on both its net-
work position as a brokerand on an organizationalmemory
that allows it to acquire, retain,and retrieve new combina-
tions of informationobtained throughsuch a position.
The notion that brokerstransformand blend informationis

implicitin DiMaggio's(1992) descriptionof how Professor
Paul Sachs used his strong connections to the previously
weakly connected worlds of museums, universities,and fi-
nance to help create New York'sMuseum of ModernArt.
This notion is also implicitin writings on technology transfer
(Rosenberg, 1982, 1994; Rogers, 1983), which recognize
that existing technologies are often adapted and transformed
before they become usable in a new field. But these writ-
ings do not focus on the role that individualactions and orga-
nizationalroutines play in recognizing,storing, blending,and
transformingthose technologies to make diffusionpossible.
Except for Attewell's (1992) descriptionof how consultants
facilitatedthe diffusionof a business computingtechnology,
we don't know of any empiricalor conceptual work that
weaves together macro perspectives on external networks
with micro perspectives on internalroutinesto describe the
role of brokeringin innovation.
717/ASQ, December 1997

This paper develops such an integratedperspective in a pro-
cess theory of how one productdesign firm acts as a tech-
nology broker.FollowingWeick's (1992) approachto theory
building,we develop a relativelyfull explanationof brokering
in a small region, which is then used to guide general dis-
cussion about brokeringin other settings. We use an eth-
nographyof a productdesign consultingfirmto develop a
local theory of how this organizationacts as a technology
broker.This firm has designed productsfor several hundred
differentfirms in over 40 industries,rangingfrom pagers to
closet-size medical analysis products.The quantityof new
productdesigns (the firm works on between 60 and 80 prod-
ucts at a time), together with the tangible natureof the me-
chanicalsolutions that usuallymake up those designs, al-
lowed us to observe how this firm recognizes, blends, and
transformsexisting ideas into new and innovativecombina-
tions. By havingstrong connections to many industriesbut
not being central in any one, the engineers in this firm have
constant opportunitiesto learnabout technologies from a
broadrange of industries.The firm exploits its network posi-
tion with internalroutines that help its designers create prod-
ucts for currentclients that are new combinationsof existing
individualtechnologies that these designers have seen be-
fore. Manyof these products reflect the transferof ideas to
industrieswhere they have not been used before and the
creation of combinationsof ideas that no one in any industry
has seen before.
METHODS
Research Setting
This ethnographywas conducted at IDEO,the largest prod-
uct design consultingfirm in the United States.1 IDEOwas
co-founded by the currentCEO DavidKelleyin 1978. It em-
ploys over 125 designers who develop productsfor other
companies. Headquartersare in Palo Alto, California,with
smaller offices in Boston, Chicago,GrandRapids,London,
San Francisco,New York,and Tokyo.The bulkof IDEO's
work is in mechanicalengineering and industrialdesign. Me-
chanicalengineers design productsfor physicalperformance
and ease of manufacturing;industrialdesigners use artistic
skills (with an appreciationof engineering)to design prod-
ucts that are attractiveand easy to use. Our study focused
on the 45 or so engineers in Palo Alto who do mechanical
engineeringand (to a lesser extent) electrical,software, and
humanfactors engineeringand on the 35 or so managers
and staff who supporttheir work. We follow our informants'
usage and describe IDEO'sengineers as "productdesign-
ers" or "designers" most often, but we (andthey) some-
times use "engineers." Most designers are 25 to 40 years
old, male (about80 percent), white (about80 percent), and
usuallyhave a B.S. or an M.S. in engineering. Managers
1
have a similarprofilebut tend to be older (35 to 50 years
Sutton and Hargadon (1996) also used
old). Supportstaff also have a profilesimilarto designers,
this ethnography as the basis for a paper but a higher proportion(approximately50 percent) are
on the effectiveness of brainstorming women.
sessions. That paper contains additional
information about the research setting for Clients typicallyhire IDEOto design partor all of a product
this ethnography and the methods used,
as well as about IDEO's structure, work that they would like to manufactureand sell but lack the ex-
practices, norms, and values. pertise or staff levels to design. Clients range from Fortune

50 to start-upcompanies. IDEOusuallycharges clients for

the time and materialsrequiredto design a productbut oc-
casionallyworks in exchange for a percentage of sales or
profitsfrom the finished product.Design projects last from a
few weeks to three years, with an average of about a year.
Results range from sketches of productconcepts to crude
workingmodels, to complete new productdesigns. IDEO
has contributedto the development of over 3,000 products.
Widely known products includethe originalApple computer
mouse, a Microsoftcomputer mouse, Smith ski goggles,
AT&Ttelephones, Oral-Btoothbrushes, Crest toothpaste
tubes, Steelcase furniture,Sega game controllers,Hewlett-
Packardprinters,rechargersfor GeneralMotors' electric ve-
hicles, laptopcomputers for such firms as Apple Computers,
Dell, and NEC,the Macintosh DuoDock, Reginavacuum
cleaners, and a life-sized,functioning,mechanicalkillerwhale
used in the film "Free Willy."Less widely known products
include surgicalskin staplers, a combinationbeach chairand
cooler, a coin sorter, a blood platelet function analyzer,a toy
guitar,and the Enormetelephone. IDEOis widely praised in
the business press for its innovativedesigns; for instance,
IDEOwon more Business Week Design Excellence Awards
in 1993, 1994, and 1995, and over the last decade, than any
other productdesign firm.
Method
Each of the two authors spent six to eight hours per week
doing an ethnographicstudy of IDEOfrom March1994
through May 1995. Fieldworkcontinued at a less intensive
level through February1996, with at least one of us visiting
IDEOeach week. We wrote field notes after each visit or
meeting. Each of us also visited IDEOat least once a month
through December 1996, often to collect more evidence or
to check the accuracyof facts that appear in papers about
IDEO.Any visit to IDEOentailed unplannedconversations,
because many engineers, support staff, and managers were
curious about our research and because IDEOnorms sup-
port friendlytalk about the firm and the design process. The
buildingswhere most design engineers work have a modi-
fied open-office plan, which furtherencourages informaltalk.
In addition,most IDEObuildingsin Palo Alto (industrialde-
sign, administrativeoffices, two machine shops, and a joint
venture with a large corporation)are on the same street and
within a few blocks of one another, so many unplannedbut
enlighteningconversations occurredas we walked between
buildings.
We began this ethnographywith a vague research question:
How does IDEOinnovate routinely?Althoughwe often inter-
acted with senior managers duringthese visits, our data col-
lection focused on watching and talkingto productdesigners
and lookingat and gatheringthe drawingsand physicalarti-
facts that resulted from their work. We adopted this focus
because our primaryaim was to understandhow people do
and experience innovativework, not how it is viewed by
management or support staff. Followingguidelines for induc-
tive research, we were as descriptiveas possible untilmajor
themes emerged from the data (Glaserand Strauss, 1967;
Miles and Huberman,1994). When a promisingtheme like
technology brokeringemerged, we focused data collection

on it, read pertinentliterature,and did preliminaryanalyses
to decide if it was worth pursuing.Our interest in brokering
was sparked in the winter of 1995 when we noticed that
designers offered solutions to new problems by describing
similarsolutions they had seen in past products, a process
they called "cross-pollination."2The evidence guidingour
descriptionsof and inferences about technology brokeringat
IDEOis divided into seven general categories:
1. Tracking development projects. Each of the two authors
followed a development team as it designed a product.We
met with team members about once every two weeks, at-
tended design meetings, and were given sketches, reports,
and videotapes. The first authorfollowed a team for four
months untilit nearlyfinished designing a Reginavacuum
cleaner. A new CEOstopped work on this prototypeand
other designs being done by and for Reginato reevaluate
the firm's productstrategy. About six months after we
stopped trackingthe team, Reginadecided that IDEOshould
finish the design, and the productwas completed, manufac-
tured, and sold. The second authorfollowed a team for six
months while it worked on personalappliances. He followed
this team untiltwo prototypes and detailed drawingswere
completed. These completed designs were not manufac-
tured and sold because the firm changed strategic direction
after it was acquiredby a largercorporationand the CEO
stepped down.
2. Semistructured interviews with designers and manag-
ers. We conducted 60 semistructuredinterviews;37 were
tape-recordedand transcribed;we took notes duringothers.
We had multipleinterviews with some informants,so ap-
proximately35 people were interviewed. In initialinterviews,
we asked senior managers and designers general questions
about IDEO'shistory,clients, competitors, structure,human
resource practices, and work process. Subsequent inter-
views focused on themes like technology brokeringthat we
wanted to learnabout in detail.
3. Informal discussions. We had hundredsof informalcon-
versations with managers, designers, and support staff,
rangingfrom brief exchanges to long talks over lunch.We
talked with almost every employee at the Palo Alto head-
quartersand had dozens of conversations with the CEO.We
2 also had informalconversations with ten IDEOclients about
Various versions of this paper have been
described to and read by IDEO designers
the company. The content variedwidely, with designers of-
and managers since our ethnography ten gossiping about new clients, employees who had been
ended in March 1996. This conversation hiredor had left, the virtues or drawbacksof currentIDEO
and reading has led some designers to
adopt the term "technology brokering" prototypes, "cool" new technologies that they had seen or
as a synonym for "cross-pollination." heard about, or why they loved or despised existing prod-
ucts, rangingfrom toy Slinkiesto Harley-Davidsonmotor-
3 cycles. In addition,after we began asking questions about
The technological details of the products emerging themes, includingtechnology brokeringand brain-
and projects we observed at IDEO are
critical to presenting a theory of technol- storming sessions, designers often approachedus with com-
ogy brokering. For reasons of client confi- ments, questions, stories, prototypes, and sketches that
dentiality, however, we have had to dis-
guise approximately 10 percent of these
they believed would enhance our understandingof these
products. In doing so, we substituted topics. Conversationswith clients were equallydiverse, but
products of similar technological com- at least three of them talked with us about technology bro-
plexity and confirmed these selections
with the designers who were involved in
kering.Forexample, one client described how IDEOdesign-
the original projects. ers had introducedhis organizationto promisingtechnical

solutions that were new to that industrybut were used

widely elsewhere.
4. Brainstorming sessions. We observed 24 group brain-
storming sessions in which productswere designed, six in
person and 18 on videotape. Each meeting was initiatedby
members of a design team. They invited IDEOdesigners
who were not team members to generate possible design
solutions for the project. IDEObrainstormsare scheduled
meetings and are held in conference rooms. Five brainstorm-
ing rules are displayed in large letters in several locations in
each room: (1) defer judgment;(2) buildon the ideas of oth-
ers; (3) one conversationat a time; (4) stay focused on the
topic, and (5) encourage wild ideas. IDEO'sMethodology
Handbook,which outlines IDEO'stechniques for new de-
signers, contains 11 pages of instructionsabout how to fa-
cilitateand participatein brainstorms.Designers who lead
brainstormsare skilled and experienced facilitators;nearlyall
IDEOdesigners have extensive experience as participantsin
brainstorms.
The sessions we observed lasted between 45 minutes and
two hours. The topics rangedwidely: three about personal
appliances,three about furniture,three about video cameras,
two about surgicalskin staplers, two about medical devices
to aid healing,two about blood analyzers,two about laptop
computers, two about personal communication,one about
remote controls, one about ski goggles, one about vacuum
cleaners, one about faucets, and one about a portabletraffic
control system. Typically,projectengineers introducedthe
projectand described a design problemthey were facing,
then the other engineers offered possible solutions, often in
the form of solutions they had seen in other settings. Solu-
tions were sometimes found in similarproductsthat were
broughtto brainstorms(e.g., a designer suggested adapting
a design solution for a new skin stapler that was already
used in a competitor's product)or in productsthat were
broughtin from differentindustries(e.g., a designer showed
how a gas engine from a model airplanecould be used to
power a skin stapler). Designers also described and
sketched solutions on paper or on whiteboards in the room.
The visible and vocal natureof these meetings offered us
the opportunityto observe how new problems and existing
solutions were shared among the designers.
We wrote field notes about each brainstormand were given
"brainstormingreports"for nine of the 24. These reports are
preparedfor the client by the brainstormorganizers;they
summarizethe ideas generated and develop promisingideas
in greater depth. We also distributeda short survey about
productdesign brainstormsat IDEOto engineers in the Palo
Alto office. We distributed45 surveys, and 37 were re-
turned;27 includedwritten comments. This survey con-
tained 40 closed-ended questions, but in this paper we only
use written comments that designers made in response to a
request for "any other comments about brainstormingat
IDEO."See Sutton and Hargadon(1996) for a more exten-
sive discussion of brainstormingat IDEO.
5. Other meetings. We attended a session about how to
handle a majorclient, a session with that client, a meeting

with IDEOengineers who studied their firm's design pro-
cess, and about twelve "Mondaymorningmeetings." Most
Mondays, CEO Kelleymeets with the employees in Palo
Alto who do or support "engineeringdesign." They usually
sit on the floor in a circle. Meetings start with Kelleytalking
about pressing, interesting,or funny events and then turnto
new projects and progress on ongoing projects. "Show and
tell" is next, in which designers displayand describe new
products, prototypes, materials,and methods. We also at-
tended and participatedin three meetings about IDEO'sde-
sign process. The first and second of these were brainstorm-
ing sessions on how to describe and transfer IDEO'sdesign
process to other organizations.The thirdmeeting focused on
technology brokering;experienced designers talked about
how and when they had combined their diverse technical
knowledge to create new productsand things they did to
facilitatethis process.
6. Design team interviews. We did retrospective interviews
with four design teams, which were tape-recordedand tran-
scribed. The productswere a label-maker,a blood platelet
analyzer,a mechanicalkillerwhale, and a furnituresystem.
Each was a large-scaleprojectrequiringmultipleengineering
disciplines. Designers broughtprototypes and the final prod-
uct to two of the interviews. We asked the group to de-
scribe how the project unfoldedand the role that each mem-
ber played. We asked them to describe the technical details
of the project:the prominenttechnologies of the final de-
sign, how these were chosen, and how each team gener-
ated and explored alternativesolutions throughoutthe proj-
ect. Finally,we asked them to describe any interpersonal
and politicalissues that arose duringthe project.
7. Materials about the organization. We gathered several
dozen stories about IDEOfrom varioussources, including
Fortune,Business Week, Wired,ID, WallStreet Journal,and
popularbooks. We viewed approximatelyfifteen television
programsabout IDEOfirst shown on outlets such as ABC,
CNN, BBC, PBS, and the DiscoveryChanneland explored a
CD-ROM"tour"of IDEO.We gathered other materialspro-
duced by and about IDEO,includinga MethodologyHand-
book for new engineers and sketches of prototypes.We
also reviewed IDEO'scollection of approximately1,400 pho-
tographs of productsketches, prototypes,the design pro-
cess (e.g., pictures of brainstormingsessions), and com-
pleted products.
A PROCESSMODELOF TECHNOLOGY
BROKERING
These qualitativedata indicatethat IDEOlearns about poten-
tiallyuseful technologies by workingfor clients in multiple
industriesand finds opportunitiesto use that knowledge by
incorporatingit into new productsfor industrieswhere there
is little or no priorknowledge of these technologies. This
design process results in the movement of technologies be-
tween industries,reflectingthe technology transferand dif-
fusion that is recognizedas fundamentalto technological
evolution (Rosenberg, 1982; Basalla, 1988; Hughes, 1989).
Existingresearch has considered the social, economic, and
politicaleffects of this type of innovation,but little is known
7221ASQ,December 1997

about the nuances of how such processes unfoldwithin or-

ganizations.
Technology brokeringat IDEOentails more than just trans-
portingideas between previouslyunconnected industries;it
also means transforming,sometimes radically,those ideas
to fit new environmentsand new combinations.An innova-
tive productmight contain several components that are new
to the industry,blended with many old components that
continue to fit the industry'sneeds. Brokeringrequiresinte-
gratingthese new and old technologies in ways that allow
each to functionwell. Forexample, to develop the Cholestec
Home CholesterolTester, IDEOdesigners combined a com-
pact disk inject-ejectmechanism, a simple software inter-
face, and high-volumeproductiondesign principles,each of
which were relativelynew ideas to the medical products in-
dustry,with samplingand testing components and chemical
treatment technologies alreadyused widely and fairlywell
understood in that industry.Designing the productrequired
modifyingboth the compact disk inject-ejectmechanism to
fit the needs of the existing samplingtechnologies and
modifyingthe existing samplingtechnologies to fit the capa-
bilitiesof the inject-ejectmechanism. Manyof IDEO'sprod-
uct designs are, like the home cholesterol tester, new com-
binationsof existing components that reflect Weick's
(1979a: 252) definitionof creativityas "puttingold things in
new combinationsand new things in old combinations."4
This perspective on technology brokeringbegan to develop
when, early in our study, we noticed that many of IDEO's
designs contained innovativefeatures that engineers had
seen in previous products.We created a list of IDEO-
designed productsthat includedfeatures designers had
adapted from previous products, prototypes, or other
sources outside of the client's industry.We met with IDEO
designers individuallyand in groups to add to and refine this
list, which is shown in table 1. It contains 30 examples of
IDEO-designedproductsthat are new combinationsof old
technologies taken from both inside and outside the client's
industry.One example is a portablecomputer docking sta-
tion designed for Apple Computer.It consists of traditional
computer components that were combined with an insert
and eject design adapted from video-cassette recordersand
powered by an inexpensive motor found in toys.
Our data suggest that IDEO'sabilityto generate innovative
productsthat are new combinationsof existing technologies
4 can be understood by consideringboth the organization's
Thisdefinitionof creativityis not new. network position and the behaviorsof its designers in ex-
Schumpeter(1934:65-66) describedin-
novationas the "carryingout of new ploitingthat position. Figure1 summarizes the relationship
combinations"and Usher (1929: 11) de- between network position and internalbehaviorsin a four-
scribedtechnologicalinnovationas the
"constructiveassimilationof pre-existing step model of technology brokering.Access (step 1) de-
elements into new syntheses." A decade scribes how IDEOfills a gap in the flow of informationbe-
earlier,Ogburn(1922;quoted in Basalla, tween industriesand is able to see technologicalsolutions in
1988: 21) defined inventionas "combin-
ing existingand knownelements of cul- one area that are potentiallyvaluablein others (Burt,1992a,
ture in orderto form a new element." 1992b; DiMaggio,1992). But the way these technologies
And, even earlier,Ribot(1906; quoted in become innovativesolutions to currentproblems depends
Torrance,1988: 45), a psychologiststudy-
ing creativity,maintainedthat creative on how these potentialsolutions are shared within IDEO
thinkingproduced"unforeseenand novel across designers and over time. The remainingthree steps
combinations,"but "inequal measure
absurdcombinationsand very original of the model describe the role of IDEO'sorganizational
inventions." memory in turningtechnologies seen in past products into

Table 1
IDEO-DesignedProducts that IncorporatedTechnological Solutions from Outside Industries
1 Waterbottle:Combinesexisting body with leak-proofnozzle based on previousshampoo bottle design.
2. Bloodanalyzer:Combinesexisting analytictechnologies with computercomponents: printer,keyboard,display,
and circuitboard.
3. Portablecomputer:Hinge design in portablecomputerdisplayincorporatesa bailmechanismfound in
typewriters.
4. Whale special effects: Mechanicalwhale combines hardwareand software from the computerindustry,
hydraulicsand roboticsfrom designer's academic background,and latex skin and other existing special effects
techniques.
5. ComputerPCMCIAcardadaptor:Ejectmechanismcombines nitinol(memorymetal)from defense industry
technologyand existing circuitboardand connectortechnologies.
6. Vacuumcleaner:Combinesexisting components with new complex plastic partsdesigned utilizingprevious
CADexperience.
7. Home cholesteroltester: Existinganalyticcomponents combinedwith CD inject/ejectmechanismfrom
consumer products.
8. Portablecomputer:Retractingfoot design based on foot mechanism on slide projector.
9. Toy electric guitar:Incorporatestoy industrymaterialsand design with microprocessortechnologies from
previouscomputerprojects.
10. Inputdevice for kid'svideo games: Combinesoversizedtrackballfrom previouscomputerinputdevices and
existing toy industrycomponents.
11. Cosmetics product:Incorporatesflexibletubingfrom previoussurgicalproductand vacuumtechnologyfrom 2
previousvacuumcleaner projects.*
12. Bicycle helmet: Includessailclothstrengthenerfrom designer's sailingbackgroundas well as existing foam and
shell components.*
13. Labelmaker:Existinglabel makerenhanced with interfacedesign from computerprojectsand displayscreen,
printer,inputdevices from computerprojects.
14. Personalcomputer:New design for cooling computers based on design principlesin ceiling fans.
15. Surgicalskin stapler:Existingstaplercombinedwith ideas from model airplaneengines, office staplers, and
other medicalproducts.*
16. OriginalApple mouse: Mouse design trackingmechanismadaptedfrom giant trackballin video game machine.
17. Handheldcomputer:Hinge mechanism based on principlesfound in office binderclips.*
18. Personalcomputer:New computerdoor design based on idea from garage door via previouscomputer
projects.
19. Tirepressure monitorand valve: Pressure gauge based on bellows mechanismfound in stainless-steel fuel line
product.
20. Desk lamp:Uses articulatingball-and-socketjointdesign taken from principlesin humanhip-bonesockets.
21. Portablecomputerdockingstation:Uses an eject mechanism based on ideas from video-cassette recorders,
dockingconnectors from a previouscomputerdockingproject,and an inexpensiveelectric motorfrom toys.
22. Medicalanalysis product:Incorporatesa solid-statefluidwarmerfound in portablecoolers for automobiles.
23. Computermonitor:Existingmonitorincorporatesa clutch springdesign based on idea of leaf-springsin
automobileshocks.
24. Office chairspring:Seat springcombines existing seat springcomponents with design of rubberspringshocks
used in tool and die industry.*
25. Electriccar charger:Powered door opener uses gas piston from rearwindow of station wagon combinedwith
electricchargingcomponents.*
26. Portablecomputer:Displayfastened closed using bicycle spokes and existing displayhousingtechnologies.*
27. Waste papercollector:Take-upreel design based on ideas from continuoustowel dispensers and typewriter
printerribboncartridges.
28. Paperhandlingproduct:Papertray mechanismcombines parallelrulerdesigns in draftingboardswith existing
paper handlingcomponents.
29. Computermonitor:Conceptfor new mountingdesign based on "Monkey-on-a-tree" toy.
30. Slide printer:Motioncontrolsolutionbased on stainless steel, zero backlashmotion controlfound in early
disk-drivehead designs.
* Theearlyprototypesof these productsused off-the-shelfcomponentsfromthese differentindustriesintestingthe performanceof this
technologicalsolutionin its new combination.Laterdesigns became specificto the new product,but inthe case of the bicyclespokes,
a localwheel manufacturer actuallybecame a supplierof productionpartsfor a portablecomputer.
useful informationfor designing new ones (Marchand Si-

mon, 1958; Huber,1991; Walsh and Ungson, 1991). Acquisi-
tion (step 2) describes routinesthat IDEO'sdesigners use to
bringtechnologicalsolutions into the organizationalmemory,
where they are stored for possible use in future design pro-
jects. Storage (step 3) describes how these solutions remain
in memory untilthey are considered for use in future de-
signs. Finally,retrieval(step 4) describes how designers re-
trieve some of these old technologicalsolutions from the

organizational memory in forms that fit the new combina-

tions they are creating.
This process model presents access, acquisition, storage,
and retrieval as linear and distinct phases. We use this
model because it fits our data reasonably well and provides
a simple and analytically useful way of summarizing these
data. Nonetheless, the process was not always as neatly
linear as the model implies and the steps could not always
be cleanly distinguished. As Walsh and Ungson (1991: 82)
recognize, "Because the acquisition, retention, and retrieval
of memory is an ongoing process, it is difficult to pinpoint
the exact boundaries between these processes."
We used an iterative process to develop the inferences
about the process of innovation through technology broker-
ing at IDEO that are summarized in figure 1. Following Gla-
ser and Strauss (1967) and Miles and Huberman (1994), a
set of iterations usually began with a hunch inspired by the
data or literature (e.g., an informant mentioned that the origi-
nal idea for a water bottle valve came from another designer
who had worked on a previous shampoo bottle project,
which suggested that ideas from different industries pro-
vided IDEO with potentially valuable solutions in later pro-
jects). Then, to see if a hunch could be grounded, we com-
piled pertinent evidence from all seven data sources (e.g.,
we looked for evidence that IDEO's experience in a range of
industries provided its engineers with useful ideas). These
analyses led us to abandon, modify, or maintain each infer-
ence (e.g., we retained the inference that access to a range
of industries was an important aspect of IDEO's innovation
process). If the inference was retained, we summarized the
grounding for it in a within-site display reflecting how
strongly each inference could be grounded in each data
source. We then wrote up our inferences about each re-
tained consequence, weaving together conceptual argu-
ments, additional evidence, and citations to pertinent litera-
ture. Table 2 presents the evidence that grounds our
process model of technology brokering.
Access: IDEO'S Network Position as Technology Broker
Brokers derive value by enabling the flow of resources be-
tween otherwise unconnected subgroups within a larger net-
work (Marsden, 1982; Gould and Fernandez, 1989; Burt,
Figure 1. A process model of how innovation occurs through technology brokering.
t
epIndustries
Steop 1: Access Step 2: Acquisition Step 3: Storage Ste 4i in
Retrieval
in the flow of
luGaps IDEO'sdesigners These potential Designersworking
information bringthese
Out.ut technological on new products
between industries technological solutions remain in retrieve
provide IDEO's _solutions into the _memory until _technological _Design solutions
designers with organizational design projects solutions from the that are new
to
exp osure memoryand thus come up that organizational of
technological intopotentaluse mightbenefitfrom memoryin existingideas
solubions in one area in future design them. appropriate forms
that are potentially poet. to fit the new
_valuable yet combinations they
previously unseen in are creating.
others.NtokPsto DOs raiainlMmr

U _0
~~~
>~~U)U)
cnc/,-n 2~~~~~~~~
U c~~0cU
5)-C ~ ~~~
O~~~~~~~~c~~~~
U)~~~~~~~~~~~~~~~~CU 0
-~~~~~ ~- U)~~~~~-. co
L-
O I~~~~~~~~~~~'E~~~~~~o
U))04cn0' cn00
)
0 a)C/
~~~cuU) ~ ~ ~ ~~ 2 8-~~~ 0Co
cUO
C
U)
:5 ~~~~~~~~ a)0)L
~~~~~o-5co
Co 0)cn
uU) 0
U 20E
al)- -C
a) U) a) L-
co 0 U)E)0 , ~ U C.
C CC 0) C - C fl u
CU C -C a)
co
~ ~~ ~ ~
EC-)
co C x -C
C~~~~C
0 E 0~~~~~~~~~~~~~~~~~~~~~~~c
0l
C
a)E :3 U)
0) 0 0)
-0
L-
0
(D
co~~~
-
a) a) coU
0
0U)
-W
)
>
C)
ECl
0 L- co .- 0 :3
- _0~~cI) 0 C3
U) 0co ) -E c a)U)+ U) U)
.L Cufi.- ~U) Ccn U)0 0
o i-I aC~
'.
-' -0 co 0 -C
0
C~~~~~~~~~~~~~0 - E
-C/
c
0)cnc !E00 U)2coC cu

U) U1 ) E )CD co
U
E C= E - ( >2E, AeU0co0 ) -O 0.
U)Cu
o OO l0 cn - U
c
C
-0 CL(DCnC
L-0))>nECo
g
.-
co
-'
c Ae-CCC -&&
-C Ae
x0CC 0 cu -
a)_ 0( ( (DcoU) _00c a) a) a) 0
2? ~~~~~~~
~~~ U)~~~~~~0 C L- 0 Lr ) - C U)-
U E 0= >/ )Co C a)C;. Co o U c->

0
co E
U).~~~~~~~~~~~~ -t*~~~~~~~~~+ 3: C a) +U)
a) +0
U))
C
(D (D
0cEe
cnsn ~ ~ -C
c
~ U
E co _- .--
0-C ) C a)w u-
4 )+n,
_ c:0 =E c
U) ~Cl) 0
-':
Cl) C
U)
U~~~~cu.~~(D
~~~u,
u,~~~~,00 0
_ >C
O C-c 0 0) U) )EC >-Ou
)- o-~
_0 o a)~~~Q...U) 1 C U)U -C
o 0 ~ ~~~~~~~~~~~~~0
U)
C UU)
Cl)-~~~~~~~~~~~~~+
Cfl.-'U
U) -
Co 0
a'~ c)
U) --C U) a) 0 )
* EC ~~~~~~~~~~~~~~~~~~~~>
:C 1
0UC)c
C-~~~U)
?Cl) LU~~~~~~) 00
a) 0u U)C 0 Co
E~~~~~~~~
~nCU) C C~- 0 4 0+
Cu
Ot C o CQ5co0E ocn- / EcC coU) * C)

co 0) co) 0 U) C ~ Q_ C)
En U) ,5- U -C)
~~ ~
C) (DE ~-S U3 ) (D UaSAe0 U)U -0
~ c
o -ft~ o-~~2~ ~ -CuU) U)Q U)o El)
a)- _ c -- C UU .C >-- 0 o) a I
uo ~
4)
>
U) o-~~~~
j~~~~~~~3
U) ~~
CU) +3
U)
E-'.
o
C/ - 0U)
*~~~
~~,cuu,
0 0
D ' -U-C
x~~a
E-
-o
U)0
cn
a)Cl -C
M >
0 C29
>
co
coU
.0u
U (ut ~ C)Q co
a)' U))
C.) )a)) U
n
UCJ >~c )
-0-Cl a)jCu a nCU oOE'U~ 0UC)C) ,
(/)0 0 (1)~~~~~~~~~~
- - ( i
0
0)~~~~~~~~(D_ E( a,
4)
(Dcn
u cl)C~~~~~~~~)coc
2 3 C> 5: a,0 U)0Q > >
CD0
c
~a) O )C 3. ) co CCou
0 U) a) 0 4- -5~~~~~ ~ ,
- .~~~~o a~~~~aD
-oC ?c 0-E C: 0C Qc
L C.) +,
H- LU 0 ?~~~~~~n a-0 (1)>-0CY o ) - ) m-

Co
t E u5E . - _C
a) co
o~ E
_n a)
o a) VC~
0 cn~
E
C
a) 0
CD
0a)
CD M ~o ~ ~ 0~ c
ECo
~ C a C
C: -00 0' a >
U, ~ ~ ~~~~~~~~~~Y<~~~~~~~~~~C~~ __
c
Ca),
-) C _r_ -0 aC 0" CC Co0

,~_C
~ ~ ~a, ~~~~~~~~
C-C Co
cn a,
a,,.~~~~~~~~~~
a~~~~~o~~~~
C~~~~~~C-~~~
Ca,~~~~~~~~~~~a, ~ ~ ~ t~~~~~c
C-''-a,~~~~~~~~~:
-~~~~ 8a, )
_,Co
0 (n a) -j0O a -C W ,. C-' E , Co:0C~
c a) ,. - > a) : C CUE
V
Co
C
>
."o E d
a, -o -a.a~C aa, a, a)O~ :CoC a)-'
~~ ~ 0~ Va~E ~~ E
~O-o a,
a)~
a,
~ ~ ~~~~~ aj~~
L> C: cn
0 w a) 0a - C C a0)n0 LO
O~
a) a) Coaa
a) a) - a)
C-'
-M d M -o Co -0a C:
03 0a a)
0 -a
a)o CD0~ CD C.' - ( Ca E a) m
) a),C' aC-O a C= 3'
0
U) .~~~~~~~ ~~~~~~~k
a)
EESE a)-0
-.
Va
: +-
o
C)
cola,'-
V ~~~~~~~~~- S-
a,
C Q -
7,
aC
C'0o
a)
a)
C
a, a,.c o
V
C-
c-',na
C --' 50CC
~ 0w
a
0, a
a)
c~~~~~ a, C'- ~~~~~~~~~,5 a,. o CD.~a a)

-
CC.> >-
)a, W
C.Ca
- VC a) C)C. a) `0
InCo
c o Ea,.C5, -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~c
-
)'5 VcnC
C
cn
-
5 jCCD
.9
a EZO a, ~CL MC Co- a, CU0 a) a~.- a, :

0 Co C
CoTOa 0 0CC-1 0a), 0o, E
a, a, C~~~.C ~ ~~~~
-~~~~~~~~~ -og*~~~~~~~~.~~~a,
a,~~) -r Ca) a) M
oC
2~ C
C.O.a)
LO
Ea,2a
4E ~ ~ C/)
a)o0
C
C
C Oa a-0 CI a
C/)
C/.
~z c
0)-Co~~~~~~~~~~~a
>- -c C Co
a) A -- o _ a
Zfl V1 -C
0 +1 a,~ a,0
a) a) aE anC-.
a)C ~ 0 a,
C.~
a)
+-a)
0
=3
00 o a,
+- E a>-) a0) -C'0
a)
CV'-~~~~) 0. Co.C~~a)E
0 Co oEn M CD.V
C>
C E'aa,
E a, CLa)._ a) 0a )
_ -C a, a)C +-
a)-' -e 0C. . a) Co3 ,co
C-aC
aC0 CY) Co co 0) a, c00)r
(n n c m a
CYo -o
a, C- a, +1 >C- a) a)a0C
Co:
-X a,
o aC C : - CoCC. - o' C.C-
Coa, "'VtCa, a
a) C- -
0 (1) Co ~CY C.l) C
CL COO+ 0 C/t
'15 Ew 10 0 oo a) co co V
-C VCoa)a, C.0) a, a 0: =3) C --
C;) 0a) E
a,0 CO o C ti a
+-Cac
a))o a ..C>ca, a,
C.)a,.C 0 co0" -C-
~~ aC~~~~~I5 E~~~~~ECU>'- - C-CaoC
o ~~~~~ 0a,~~~~~~Co
Vn >C: -C)C4j coa V >a V 0tf -C > '~ a,0VQ-O
> a Of
.C
C) a, 0E 5a) aa, C) Co +, ,0 ,W
a,
F-
~ ~~~~~~~~~~O(
0
~ 0
0
a)0COc
C.n0
C-' U C.0?.' C.~~~~)
C.~~-'
C.- a)~ c V
0

1992a; DiMaggio,1992). Marsden(1982: 202) defined bro-
kers as intermediateactors that "facilitatetransactionsbe-
tween other actors lackingaccess to or trust in one an-
other." Considerablenetworkanalyticresearch has shown
the power that accrues to brokers.Fernandezand Gould
(1994) showed that organizationsoccupying brokerageposi-
tions in the nationalhealth policy domainwere more likelyto
have greater perceived influence. Padgett and Ansell (1993)
explainedthe rise to power of the Medicifamily in fifteenth
century Florenceas the result of a network position span-
ning otherwise unconnected subgroups. Burt(1983, 1992a,
1992b) described how the value of connecting differentsub-
groups depends on the relativelack of other ties between
those subgroups. By restrictingthe flow of informationbe-
tween subgroups, this lack of ties creates disparitiesin the
knowledge held by the differentsubgroups and enables bro-
kers to profitby providingaccess for each subgroupto the
ideas of the largernetwork. Such a disconnected network
structureallows brokersto benefit because they "are well
connected in several networks, ratherthan extremely central
in just one" (DiMaggio,1992: 130). When the ideas are tech-
nologicalsolutions, brokersbenefit by being well connected
to a range of disparateindustriesand enablingthe flow of
existing solutions between those that have such knowledge
and those that do not.
Social networktheory describes networks of individualor
organizationalactors and the relationshipsbetween them.
Withinthese networks, subgroups bound sets of actors that
"know one another, are aware of the same kinds of opportu-
nities, have access to the same kinds of resources, and
share the same kinds of perceptions" (Burt,1983: 180). An-
other network perspective, actor networktheory, has
emerged from studies in the social constructionof technol-
ogy and presents networks as comprisingnot only actors
but also the physicalartifactsand concepts with which those
actors relate (e.g., Callon,1980; Latour,1987; Law, 1987).
The relationshipsof this more diverse network arrange
physicalartifacts,individuals,and concepts into complex or-
ganizationaland technologicalsystems. Just as organizations
comprise networks of actors, products become "networks
of juxtaposed components" (Law, 1987: 113). This expanded
definitionof network elements may more accuratelyreflect
the technologicalenvironmentthat IDEOdesigners face,
where informationabout existing solutions resides within the
artifactsthemselves, and brokersneed not have close ties to
other actors to access that information.Forexample, in one
projectwe followed, designers learnedas much about de-
signing a new consumer productby studying the existing
(and related)productsas they did from talkingto the client.
5
Evidence suggested that IDEO also bro- Subgroups in this expanded network,then, reflect relatively
kered technological solutions between
groups within client organizations (as one
isolated sets of actors, technologies, and concepts. The
informant described, "we take your boundariesbetween these subgroups can exist at many dif-
watch and tell you what time it is") and ferent levels, between individuals,organizations,or indus-
organizations within a single industry
(though IDEO is careful to avoid intellec-
tries; we chose to draw them between industriesbecause
tual property issues). We focused on bro- technologies most clearlyemerge and evolve within particu-
kering between industries, however, be- lar industriesyet may have potentialvalue in other industries
cause conceptually and empirically it
offers the clearest perspective on the (Basalla,1988; Hughes, 1989).5The transferof potentially
process of technology brokering. valuabletechnologies to other industries,when it occurs,

can cause significanteconomic and competitive changes

(Schumpeter, 1934; Rosenberg, 1982), but gaps in the flow
of informationacross industryboundariesoften prevent this
diffusion.Organizationslike IDEO,by occupying positions
within multipleindustries,may bridgethese gaps.
IDEO'saccess to outside industriesoffers an advantageto
clients who want new productinnovations.IDEO'sdesigners
have generated partor all of over 3,000 new productde-
signs for clients since its formationin 1978. They have
worked most heavilyin the personal computer, medical
products, and office furnitureindustriesand have also de-
signed productsfor the toy, telephone, automotive, movie,
ski, bicycle, printer,and video game markets. IDEOdoes not
maintaina database of clients by industry,but our research
indicates it has worked in over 40 industries.Fromthese
industries, IDEO'sdesigners have typicallyseen a broader
range of technologies than clients with experience in only
one or a few industries.The network concept of range de-
scribes the extent to which an actor contacts a diversityof
other actors and can be measured in two ways, as volume
of contacts or as qualityof contacts (Burt,1983). Volume
measures the total numberof contacts an actor has; quality
measures the extent to which an actor's contacts provide
nonredundantinformationand support.The evidence sum-
marizedin table 2 suggests that IDEO'svalue as a technol-
ogy brokerdepends not only on the numberof clients and
industriesit works with (volume of contacts), but also on the
technologies in those industriesthat are potentiallyvaluable
yet previouslyunknown in others (qualityof contacts).
IDEO'sMethodology Handbookrecognizes this value:
"Workingwith companies in such dissimilarindustriesas
medical instruments,furniture,toys, and computers has
given us a broadview of the latest technologies, materials,
and components available."
Access to dissimilarindustriesalso describes Edison's labo-
ratoryin West Orange,which consulted to diverse clients.
Millard(1990: 48) described Edison's simultaneous pursuitof
electricalproductsfor clients in multipleindustries:"The ex-
tensive contract research carriedout by the laboratorystaff
opened up new areas of investigationand offered valuable
spilloversof informationthat Edisonwas waitingto exploit."
Millard(1990: 68) cited an example of this spilloverin Edi-
son's work for differentclients in sound recordingand in
telephones, both of which requiredtechnical knowledge
about acoustics: "The experiments in reproducers[for re-
cording]were paralleledby the continuingwork on tele-
phone transmittersand receivers; as usual he was hoping
that one series of experiments might turn up some informa-
tion useful in another." Millard(1990: 48) also implicitlyrec-
ognized the value of such access when he described the
purpose of Edison's laboratory:"to bringtogether flows of
informationat the rightmoment, providingthe basic raw ma-
terialfor the inventionfactory."
LikeEdison's laboratory,IDEOhas access to dissimilarin-
dustries that enables it to generate new productinnovations
throughtechnology brokering.In one case, a blood analyzer
was originallydesigned to be controlledby a separate per-

sonal computer. IDEO'sdesigners instead used their previ-
ous experience designing such computers to incorporatethe
necessary features-a circuitboard,printer,keyboard,dis-
play screen, and software interface-into the product.The
result was a new and more integratedblood analyzerthat
represented a relativelydramaticcombinationof existing so-
lutions in computer and medical producttechnologies. Bro-
keringalso provides innovativesolutions to more common
design problems. For instance, when designers became
aware that a portablecomputer display lacked the room nec-
essary for traditionalfasteners, they developed a solution
using modified bicycle wheel spokes as fasteners. In both of
these examples, IDEO'sdesigners were able to bringto-
gether technologies from within and outside of their client's
industryto generate innovativenew productsand solutions.
IDEO'sMethodology Handbookrecommends that designers
"lookfor opportunitiesto expand IDEO'snetwork and/orin-
dustry knowledge." This may be more easily done now that
IDEOis relativelylarge and alreadywell connected in a
range of industriesand thus has something to offer other
clients. When IDEOwas a small start-upcompany, access to
disparateindustriesand technologicalknowledge was seren-
dipitous. IDEOhas its roots in the SiliconValleyand the
computer industry,and one of the early IDEOemployees
described their originaladvantage:"At the time we were
reallynaive, but our customer base, when we all just
started, was just as naive. So, we knew just enough to be
ahead of them and it worked pretty well. They were all elec-
tricalengineers and software guys. They didn't know any-
thing about mechanicalengineeringor makingthings, and
we knew just enough to be able to be useful to them." As
the SiliconValleyand informationtechnologies developed,
IDEOwas able to continue creating innovativeproducts by
designing the mechanicaland electromechanicalcompo-
nents surroundingthe new informationtechnologies as they
diffused to other industries.CEODavidKelleydescribed this
strategy as "being the high-technologycompany to low-tech-
nology companies." And as IDEO'sconnections to different
industriesgrew, designers gained experiences with many
other technologies and have providedtheir clients in the
computer industrywith useful technologies, such as low-
cost electric motors, new hinge designs, or new materials
that were taken from these other industries.
Technology brokeringis visible at the level of firms and in-
dustries, but it takes place throughthe actions of teams and
people. IDEO'scontact with differenttechnologies comes
throughthe individualengineers' contact with the industries
where those technologies are used. For industriesin which
IDEOhas multipleclients, such as the computer industry,
many designers have worked with and understandthe tech-
nologies involved.For industriesin which fewer engineers
have participated,such as surgicalinstruments,only a few
might be familiarwith the prevalenttechnologies. Designers
also come into contact with potentiallyvaluablesolutions
throughthe technical trainingand jobs they had before com-
ing to IDEOand throughtheir hobbies and personal back-
grounds. IDEOhas hired particulardesigners, in part,be-
cause of their past experience in and knowledge of medical

products, manufacturing,and disk drives. In addition,IDEO

hires designers for their knowledge and interest in areas out-
side of their work, such as toys, bicycles, model airplanes,
sailing, sculpting,farming,woodworking,music, opera, cars,
motorcycles, skiing, and mountainclimbing.The broad
knowledge and interests of IDEO'sengineers result in ac-
cess to design solutions beyond the solutions that IDEOis
exposed to in its clients' industries.
Most social networkanalyses measure currentsocial rela-

tions between actors. In contrast, because an engineer's
knowledge of potentialsolutions represents past as well as
ongoing relations,his or her networkties to a range of in-
dustries accumulate over time. As a result, each engineer
has a distinct body of technological knowledge from working
with IDEOclients, from past technical trainingand work ex-
perience, and from his or her personal interests and back-
grounds. The role this diverse knowledge plays in creating
new products is evident in a descriptionof how one design-
er's personal backgroundwas the source of new solutions:
Everybodyseems to have a couple of kindsof interestingback-
grounds. Fred-because he used to buildmodel airplanes-was
very good at it. Model airplaneshave all these swivel and control
things, and he'd bringall this kindof technology to our prototypes:
little brass tubes and all these little hinges. He knew all that was
possible, and he would have a bunch of stuff in his garage and he'd
bringthat in and we'd make prototypesout of it. Otherpeople are
into a lot of stuff. The technology of bicycles, which is actually
quite developed and refined,can be appliedto so many things.
People have a universallove of toys here and I think I'm the
epitome of that. Toys have so many neat things to offer. They are
high volume, mass production,often plastic, and very clever be-
cause they're so cheap. I especially love Japanese toys. We will
bring'em out in brainstormersand kindof applythe ideas to skin
staplers or mechanisms, whatever.
At IDEO,designers view their communityas a valuable

clearinghousefor technologicalsolutions that they have ac-
cumulatedthroughyears of access to dozens of industries.
Intheir words, this communityexperience allows them to
"cross-pollinate"their ideas between productsand indus-
tries. A network perspective describes how IDEOis able to
exploit this cross-pollinationin its innovationprocess by de-
scribingthe structuralconditionsthat allow an actor linking
otherwise disconnected domains to have access to ideas
that are potentiallyvaluable,but unknown,to others. Yet
while this network perspective is necessary to explainthe
conditionsthat make technology brokeringpossible, it is not
sufficient to explainhow the innovationprocess occurs
through brokering.Much of this creative process occurs
within the firm. Solutions rarelycome to the firm at the time
they are needed, to the people who need them, or in the
exact forms necessary to solve the problems designers face.
To understandhow designers at IDEOmake connections
between existing solutions and new problems over time and
across people, we need to look within the firm at the rou-
tines that designers and teams use to create new products
by learningof possible solutions, rememberingthem, and
retrievingthem in new forms that fit in new combinations.

Acquisition, Storage, and Retrieval: IDEO'sInternal
Routines for Technology Brokering
Technology brokeringmeans that IDEO'sdesigners solve
currentdesign problems by drawingon technologicalsolu-
tions they have seen in the past. This use of shared knowl-
edge from past experiences is the focus of conceptual work
on organizationalmemory (Marchand Simon, 1958; Huber,
1991; Walsh and Ungson, 1991), which refers to the means
that organizationsuse to retainpast stimulus-response infor-
mation. Organizationalmemory becomes visible when indi-
vidualmembers react to new demands by drawingon an
organizationalpool of priorresponses to similarstimuli
(Walshand Ungson, 1991). Withinthe productdevelopment
process, past stimulus-response informationrefers to past
design problems and their technologicalsolutions. The
means by which IDEOretains these past solutions become
visible in the routines its designers use to draw on this orga-
nizationalpool of priorresponses to solve currentdesign
problems.
There are conflictingperspectives on the role of organiza-
tional memory in innovativeactivities such as new product
development. March(1972), Weick (1979b), and others have
described its programmedresponses as threats, because
they increase the organization'spotentialfor unconsciously
or mistakenlyinvokingingrained,but often inappropriate,be-
havior.Other scholars, sometimes the same scholars, have
argued the opposite: that organizationalmemory supports
organizationalinnovation(Cyertand March,1963; Neustadt
and May, 1986; Walsh and Dewar, 1987; Kantrow,1987). By
routinizingsearch activities in standardoperatingprocedures,
organizationscan become more efficient at performing
them. Organizationalmemory can also support innovationby
retaininga broaderrange of potentialresponses, providing
more options for organizationaldecision makers. March
(1972: 427) asserted that "for most purposes, good memo-
ries make good choices." The tension between these two
perspectives lies between the efficiency of "automaticre-
trievalprocesses" and the uncertaintythat these processes
will evoke responses that are "out of step with the [prob-
lems of the] present circumstances" (Walshand Ungson,
1991: 73). Whether an organization'smemory supports or
undermines its abilityto innovate depends on how well its
past solutions-and routinesfor drawingon those alterna-
tives-can be adapted to fit the problems of the present cir-
cumstances.
At IDEO,althoughthe design problems change constantly,
past knowledge remains valuableif designers can recognize
similaritiesbetween old solutions and new problems. IDEO's
network position provides its designers with access to a
range of disparateindustries.This vantage point enables de-
signers to see a continuingstream of new problems to
which their old solutions apply and a continuingstream of
new (to IDEO)solutions that may be useful for future prob-
lems. IDEO'sorganizationalmemory providesthe linkbe-
tween these solutions and problems, and thus between in-
dustries, by providingsearch routinesfor generatinga range
of alternativeresponses based on past experiences. IDEO's
organizationalmemory can be described with the three fun-

damental processes presented by Walsh and Ungson (1991):

acquisition,retention,and retrievalof informationfrom the
past.
Acquisition. Access to a wide range of industrieshas
brought IDEO'sdesigners into contact with a wide range of
technologicalsolutions; the acquisitionprocess bringsthese
solutions into the organizationalmemory for possible use in
currentand future designs (Walshand Ungson, 1991). IDEO
designers acquirethese solutions by talkingto and watching
new clients and others in the industry,by readingabout the
industry,by lookingat and takingapartproducts in and re-
lated to those in the industry,and, finally,by designing prod-
ucts for that industry.
Most projects at IDEObegin with clients describingtheir ex-
isting productsand their desires for the new product. IDEO's
Methodology Handbooksays: "As consultants we need to
quicklybecome experts in the client's productarea. We
want to acclimate ourselves to the market,the buzz words,
the competition.We need to orient ourselves to the major
pitfalls,alternatives,and opportunities."This expertise helps
IDEOdesigners to understandand work with a client's exist-
ing technologies. This ongoing process of learningfrom cli-
ents started in IDEO'searly days and remains partof all pro-
jects. A senior designer recalled:
We were learningas we went. If we didn'tknow how to do some-
thing we would never say to the customer, "we don't know how
do that!" We would learnhow to do it either on the job from
people withinthe customer's company or just by going out and
findingout about the stuff. Now, given our experience, there's
probablyvery littleabout standardmanufacturingprocesses that we
don't know. But we still have the sparkcoming in because clients
come in with a new technology and want to apply it and we have
to come up to speed on that particulartechnology. So the excite-
ment of keeping up on the learningprocess is there and there's
also the excitement that you get when you hireyounger engineers
who don't have the experience and they come up to speed very
quicklyon this mass of information.
IDEO'sengineers keep "a mix of new technology coming in
from the customer side" by actively searching for the poten-
tiallyvaluabletechnologies of new industries. Knowledgeof
these technologies resides in more than just the clients that
approachIDEO,so designers also look to industryconsult-
ants, users, and suppliersas sources of existing knowledge.
In one projectwe followed, the design team talked about
how much they learnedwhen they flew to Los Angeles to
meet with an expert on the intricaciesof the materialsand
physics involvedin the project. On another project,design-
ers hiredan outside consultant,who they described as an
"electromagneticinterferenceguru,"to help them with the
design of a new computer. A projectto improvethe 15-
minute oil change offers a typicalscenario:
Ourapproachwas to understandwhat's going on, to understand
and observe. So we talkedto the presidentof the client company
to see what is going on. The clients indoctrinatedus into the new
tune-up procedures.We talked to the head trainer,the woman who
does trainingfor all the trainersthat are trainingthe mechanics and
that give demonstrativetune-ups. We talked about all the design
methodology and guidelinesfor those types of products.We went
to visitthe training[centeriandjustwatchedandobservedthe en-

tire process. We went to some shops, watched cars being worked
on, talkedto managers, looked in the back rooms. We went to
trainingseminars. We just absorbed as much as we could as an
objective outsider. . . they gave us all their tools and they gave us
engine dummies, and so we were using the tools and doing all
sorts of stuff. We were reallygetting familiarwith what the tools
were liketo handle,just immersingyourself as much as you possi-
bly can into the whole realmof it. We broughtin a mechanic,we
had him go througha tune-upas a group demo because we had a
whole bunch of people who were going to be in the brainstorm.
There were about 15 designers in the room.
The knowledge acquiredduringthese broadindustry
searches was useful in this design projectand may lead to
innovativesolutions in future projects.
IDEOengineers also acquire knowledge by studying an in-
dustry's existing products. As the quotes in table 2 indicate,
designers read industrytrade magazines and productcata-
logs. They gather all availableproducts in the field, use and
sometimes abuse them, and take them apartto find out
how they were designed. We saw this approachby a team
that was designing a kitchenappliance.The projectmanager
explained:"The best way to come up with ideas is first of all
to go out and look at what's out there. So, look at the exist-
ing products, ripthem apart,then look for peripheralobjects,
like toasters, blenders, and mixers. When you find technical
problems you go out, look around[some more] and walk
aroundrippingapartpossibly relevantproducts."Designers
on this projectcollected over 100 appliancesfrom the client,
stores, catalogs, and fellow designers to learnall they could
about the nuances and possibilitiesof the technologies in-
volved.
The primarygoal of these learningactivities is to design an
innovativeproductthat performsas well as, if not better
than, what previouslyexisted in the industry.While much of
the knowledge acquiredduringthese activities remains in
the memories of IDEO'sdesigners after a project is com-
pleted, the act of designing that new productis also an im-
portantstep in bringingworking knowledge of these new
technologies into the organization.Rosenberg (1982) de-
scribed this experience as "learning-by-using," and Cohen
and Levinthal(1994) argued that it is criticalto a firm's ab-
sorptive capacity,or abilityto exploit emerging technologies.
In creating new products, designers acquire intimate knowl-
edge of the limitationsand possibilitiesof technologies be-
yond what they might have learned by only talkingabout,
lookingat, or readingabout those technologies. In addition,
projectteams will often develop alternativesolutions
throughoutthe projectthat, while not used in the final prod-
uct, represent viable alternativesfor subsequent projects. In
one case, a sliding door covering a computer front panel was
originallymodeled after a garage door mechanism. The de-
signer builta prototype,but that design was not ultimately
chosen. Anotherprojectteam was lookingfor a similarsolu-
tion and borrowedthe concept from the originalprototype.
Again it was not chosen for the final design, but a thirdproj-
ect team, years later, did use it and bringit to production.
To IDEO'sdesigners, existing products serve as records of
the technologies in an industry.By gatheringtogether, study-

ing, and ultimatelydesigning such productswhile workingon

a specific project,designers acquire knowledge of these
new technologies for use in both currentand future projects.
But they also constantly engage in less focused and often
haphazardsearches that, in additionto being fun, provide
them with knowledge for future projects. These searches
might be entertaining"fieldtrips," like when designers went
to an airplanejunkyardto buy a DC-3wing, to a "robot-
wars" competition,and to the BarbieHallof Fame; in each
case they described the fun they had, as well as some
"cool" design ideas they saw, to the rest of the company
during"show and tell" at the Mondaymorningmeeting.
These searches also occur through more mundane acts, like
collecting new materials,catalogs, or interestingproducts,or
just takinga walk to the hardwarestore to look around,
which might also yield "cool ideas" that are announced dur-
ing "show and tell" or at least are mentioned duringinformal
conversations. Designers also told us that they helped teach
and grade design classes in local schools, partlybecause it
was an opportunityto see new and interestingideas.
Storage. Writingson organizationalmemory describe stor-
age as how an organizationputs away informationuntilit is
needed (Marchand Simon, 1958; Huber, 1991; Walsh and
Ungson, 1991). At IDEO,the storage of technological knowl-
edge became visible only as we observed the retrievalpro-
cess in conversations, brainstorms,and other group prob-
lem-solvingactivities. Fromthese observations, however, it
was evident that much of the knowledge of potentialsolu-
tions resides in the minds of the individualdesigners as
productsthey had seen or used before, projects they had
worked on, or technologies they had read, heard, or talked
about. The evidence, summarizedin table 2, indicates two
types of routines at IDEOfor storing potentialtechnological
solutions: routinesfor storing specific knowledge and rou-
tines for maintainingand refreshingthat knowledge untilit
can be used.
Routinesfor storing specific technological knowledge at
1DEOplaced potentialsolutions in the memories of individual
designers and in the objects and productsthat designers col-
lected from their previouswork. One designer grew up on a
farm and, in two brainstormswe attended (one on a new
faucet design and another on cleaning carpets), he offered
potentiallyuseful solutions based on technologies from trac-
tors and combines. But designers do not consciously identify
themselves with particulartechnologicaldomains;this de-
signer, for example, also offered many solutions that did not
come from his farmingbackground.Anotherdesigner talked
about how his memory of design solutions was "one big pile
of tools," and, while "each industryhas its own set of tools,
I only remember the tools, not where they came from." De-
signers also keep written records of previous projects, such
as brainstormingreports and partdrawings.We went to one
meeting about flexible tubing in which another designer
(who was not in the meeting) overheardthe discussion and
interrupted.He had faced the same problemwhile on a proj-
ect in another industryand offered an old brainstormingre-
port that listed the potentiallyrelevantsolutions they had
found.

Designers augment their individualmemories and written
materialsby collecting, lookingat, and talkingabout products
or parts of products,which act as records of existing tech-
nologies. Designers stockpile old productsand parts in their
offices and hallwaysor hang them from the ceiling. Some-
times these are parts and prototypes of previous IDEOde-
sign efforts that act, as one designer described, like "con-
gealed process-a three-dimensionalsnapshot of the ideas
of a previous project."Sometimes they are toys, collections
of productsfrom industries IDEOhas previouslyworked in,
parts and assemblies collected from vendors, or objects that
reflect designers' personal interests or quirks.The shelves in
one designer's office held 23 battery-poweredtoy cars and
robots, 13 differentstyles of plastic hotel keys collected dur-
ing trips, a battery-lessflashlightpowered by squeezing the
handle, an industrialpump, 11 prototypes of a portablecom-
puter, 14 prototypes of a computer docking station, six com-
petitive computers in various stages of disassembly, 15 bind-
ers from past projects, a pile of disk drives, a collection of
toothpaste tubes he had designed, a toy footballwith wings,
a pairof ski goggles he had designed, four humorous
plaques awarded for past projects (e.g., "under-the-gun"
awardfor workingunder pressure), a Frisbee that flies under
water, and dozens of other productsand parts. This designer
was especially fond of toys and spent a lot of time telling us
about how his toys contained useful ideas in the form of dis-
tinctive hinges, materials,moldingfeatures, or assembly re-
quirements.
Designers routinelyloaned these objects to one anotherfor
brainstormsand other parts of the design process. A group
of designers took the notion of shared "cool stuff" a step
furtherby fillingseveral centrallylocated file cabinets with
hundredsof "cool design inspirations."This "technology
cabinet" was a collective ratherthan an individualgood. It
was started with "donationsfrom several designers' private
collections of cool stuff," and it soon became "cool" among
a wider set of designers to add new "neat and strange
things" and to tell other designers about what they added.
One of the cabinet's self-appointedcustodians told us, "ev-
ery time you look in here, something new shows up." The
contents include nitinol(a metal-thatpredictablychanges its
shape in response to temperaturechanges), tiny fans and
motors the size of a fingernail,magnetorheologicfluid (which
changes viscosity when magnetized),samples of carbonfi-
ber parts, an inflatabletoy gorilla,and samples of flexible
circuitboards, all or parts of which the designers hoped one
day to use in their design projects. There are also metal
cases in the cabinet so that designers can transportperti-
nent parts of the collection to meetings in IDEOand at client
companies. There are also largerproductson display,like
the DC-3airplanewing and an old Texaco gas pump placed
next to a new electric vehicle chargerthat IDEOdesigned
for Hughes/GeneralMotors. A similarlydiverse set of objects
was evident in Edison's laboratory.Millard(1990: 15) quoted
Edisonas saying, "the most importantpartof an experimen-
tal laboratoryis a big scrap heap," reflecting, Millardsaid,
"his relianceon a well-stocked store-roomand a collection
of apparatusand equipment left over from previous experi-
ments." LikeEdison's equipment and apparatus,artifacts

seem to lie aroundIDEO'soffices as remindersof interest-

ing and potentiallyuseful technologies, patientlyawaitingthe
appropriateproblem.
The technologicalsolutions in the minds, written records,
and productsof individualdesigners are valuableonly when
they can be retrievedeasily for use in currentprojects. IDEO
maintainsthese memories in ways that are easy to access.
Displayingobjects where other designers can see them
makes ideas accessible, as do the constant conversations
that IDEOdesigners have about who has what design
knowledge. As the evidence in table 2 indicates, designers
informally"catalog"this knowledge about one another. In-
formants told us that, to be valued by one's peers, it was
importantto establish a reputationfor havingexpertise that
is distinctivewithin the communityof IDEOdesigners. The
Methodology Handbookis explicit in this advice: "become
an expert in the productarea: learneverythingthat's out
there." IDEOdesigners act like "gatekeepers" (Allen,1977)
who bringknowledge from the outside world into organiza-
tions. Allen's gatekeepers represented a minorityof the
population,but most engineers at IDEOfill this role. Each
designer acts as a technology brokerwithin the internalnet-
work of IDEOdesigners, bringingthe experiences of his or
her unique backgroundto bear on the problems faced by
other designers and being rewardedfor doing so with re-
spect from peers, more responsibility,and more interesting
work. Forexample, one designer who had worked in the
medical industrywas respected for her knowledge of tech-
nologies involvingfluidtransportation,such as pumps, tub-
ing, or valves. When other designers believed that such
technologies might offer potentialsolutions to problems they
faced, they would try to involve her in their project.As a de-
signer put it: "The model is you become a real expert and
you're recognized aroundthe company as being an expert in
that particularfield. You can specialize in some design area
like materialsor motions or whatever and at least be recog-
nized in the company as being reallygood at that. So, people
come to you with their questions." In addition,upper-level
managers serve as quasi-librarians. By knowingwho had
worked previouslyon what projects, they could often direct
designers to individualswith relevant knowledge. One infor-
mant described the value of an upper-levelmanager:"Peter
was the best 'hub' of information;he is involvedin all of the
projects and knows what everyone is workingon." This in-
formal reference system equates individualdesigners with
families of technologicalsolutions. So beyond developing his
or her own focused technical expertise, each designer devel-
ops broaderknowledge about which designers have which
technical knowledge. This broaderknowledge grows in paral-
lel with the retrievalprocess in brainstormingmeetings and
other social interactionslike Mondaymorningmeetings, in-
formal lunches, and company parties. Designers displaytheir
technical knowledge duringthese meetings and, in the pro-
cess, teach others when it might (and might not) be useful
to ask him or her for assistance. Forexample, we saw one
brainstormingsession lapse into a five-minutelecture by a
designer on the mechanicalpropertiesof soap before the
group started generating ideas again about carpet cleaning.
In this way, the process of retrievalfacilitates storage be-

cause it bringstechnical knowledge back to the surface, re-
freshes the memories of the individualdesigners, and re-
minds everyone else of what that designer knows.
Retrieval. Retrievaldescribes those routinesthat support
the applicationof -stored informationto an organization's
present decisions (Walshand Ungson, 1991). At IDEO,re-
trievalentails bringingstored knowledge of potentiallyvalu-
able technologicalsolutions to bear on the design problems
of currentprojects. As the evidence in table 2 indicates,
IDEOdesigners retrievetechnologicalsolutions throughana-
logic thinkingand throughestablished routinesfor sharing
the problems of currentdesign projects with other designers
in the organizationwho have relevantand potentiallyvalu-
able knowledge.
Technologies, in an abstractand conceptual form, carrythe
potentialto address many differentproblems in many differ-
ent industries. In most cases, however, designers at IDEO
learnof possible technologies by seeing them in existing
products, in specific forms intended to serve particularindus-
tries. To recognize the potentialvalue of a product'stechno-
logicalcomponents, the designers must abstractthem from
their specific, past implementationbefore adaptingthem to
meet the needs of the currentproblem.One designer de-
scribed the simplicityof the idea behind this difficulttask: "If
you take all the existing productsor thoughts on existing
productsand gather them and then took the best partof
each one and combined them, you'd have a better product.
It is as simple as that." The way that IDEO'sdesigners take
all the existing products, recognize the best partof each
one, and combine them is a criticalaspect of the retrievalof
potentialsolutions from IDEO'sorganizationalmemory.
Neustadt and May (1986) argued that analogies play a critical
role in organizationalmemory because they allow individuals
to linkpast stimulus-response informationto currentstimuli.
Similarly,Schon (1993) described the use of analogies, or
"generative metaphors,"in creative problemsolving. In
productdesign, creative problemsolving draws on the orga-
nization'smemory by makinganalogies between past tech-
nologicalsolutions and currentdesign problems. Forex-
ample, a designer workingon the hinge mechanism for the
screen of a new portablecomputer might recognize the po-
tential value of a hinge design that he or she had noticed
holdingtogether the wing of a plastic toy dragon.The re-
trievalprocess involves generatingan analogy between this
particularsolution in its implementationas a toy's wing and
the specific requirementsof the new design problem,a por-
table computer.
Analogies allow productdesigners to see the portablecom-
puter screen momentarilyas a toy dragon's wing, to view
old technologicalsolutions from a new frame of reference
that allows them to recognize certain useful characteristics,
such as material,design, or flexibility,and to ignore other
less transferablefeatures, such as shape, size, or original
use. As a result, designers can recognize potentialconnec-
tions between technologies they have seen before and their
currentdesign problems. In one project,designers attempt-
ing to develop a spill-proofnozzle for a bicycle water bottle
described how they recognized similaritiesbetween the

problems of the water bottle design and a previous project

designing a shampoo bottle that could hang upside down. In
another, designers lookingto power a door opener on an
electric vehicle chargertold us they recalledan analogous
action in the pistons that open the rearwindow of a station
wagon. In both of these examples, the ultimatesolution built
on the principlesthat were identifiedthroughthese analo-
gies, but not all analogies become partof the finished de-
sign. Designers will use analogies to generate a wide range
of alternativesolutions to choose from. One brainstorming
session on "ways to deep-clean carpets" elicited analogies
to tank treads, street sweepers, tractorcombines, hairre-
moval devices, shavers, whips, vibratingcombs, squids, and
Velcro (this last suggestion seems obligatory,and even hu-
morouslyoffered, in all brainstormswe attended). Another
brainstorm,on designing a "portablekitchencounter," re-
trieved potentialsolutions by drawinganalogies to jet fighter
wings, plastic coolers, children'sfurniture,washing ma-
chines, bentwood chairs, surfboards,and skis. Elements on
these lists may appear unrelated,but each analogywas used
to turn aspects of existing technologies into potentialsolu-
tions for the problemat hand. Analogicthinkingis criticalto
the brokeringof potentiallyinnovativesolutions because it
allows for acquisitionand storage of technologies in their
originalimplementations,but for retrievalin forms adapted
to the needs of the currentdesign problem.
To recognize the potentialvalue of a technology and adapt it
to disparateproducts, designers must be familiarenough
with a technology to generate analogies appropriatefor cur-
rent designs. Thus, much of the retrievalprocess at IDEO
entails bringingdesigners with knowledge of potentiallyrel-
evant technologies into direct contact with the problems of a
new design project. Brainstormingsessions are one of the
most direct ways that such contact occurs. Brainstormsare
face-to-face sessions for generating ideas (Osborn,1957);
design teams convene them intermittentlythroughouta proj-
ect. Almost all IDEOdesigners participatein brainstorms,
which typicallyinclude six to twelve designers who are tar-
geted for their potentiallyrelevant knowledge in a range of
technologies or industries(Sutton and Hargadon,1996).
IDEO'sMethodologyHandbooktells project leaders: "Set up
at least two majorintroductorybrainstormersto get the best
minds in the company, the collective consciousness of the
office, workingon your problem."And a designer described
what people setting up brainstormingmeetings should do:
Lookfor others with relatedexpertise that might see the idea from
a differentperspective. The most fruitfulbrainstormsin these types
of areas are when at least one participanthas a good deal of spe-
cific, availableknowledge from a differentarea that is still very ap-
plicable.Inthese cases, the client is probablyunawareof this new
information,and we can transfera lot of detailed, implementable
solutions.
Forone brainstorm,a designer picked participantswith the
mix of skills needed for designing ski goggles; some knew
about foam materialsand their design requirements,some
knew clear plastics, some knew manufacturing,and some
knew skiing. IDEOuses brainstormsthroughoutthe develop-
ment process, and these meetings often bringover half of
IDEO'sdesigners into contact with a particularproject.

Designers create a visuallyrichenvironmentin brainstorms
to help them make connections between existing product
technologies and the design problems they currentlyface.
Designers and brainstormorganizerstypicallybringin "tons
of related hardware."One brainstormwas in a room filled
with dozens of pieces from four or five vacuum cleaners
plus anotherfour assembled vacuum cleaners. A designer
broughtabout twenty differenttelevision remote controls
into another brainstorm.One designer described his prepara-
tion for brainstorms:
When the brainstormis on a trickyproblem,I always set up what is
called a crash cart. I get one of these roll-aroundcarts and fill it
with anythingI can that is relevantand then some things that may
not be even remotely relevantso that you have this big playpenfull
of stuff that's sitting on the table or it's sitting on the cart:informa-
tion, devices, data. Inthis way, brainstormsare like a big open-book
exam where you're allowed to bringstuff in.
Brainstormsare not the only arena where potentialsolutions
are drawnfrom IDEO'sorganizationalmemory. Designers
routinelyask for technical assistance at the company-wide
Mondaymorningmeetings (e.g., "Who knows about sheet
metal fasteners?"). They use electronic mail to broadcast
questions to the firm, for example, about fasteners, materi-
als, adhesives. A designer explained:"Thereare often things
that people need help on. They put them on e-mail and
there's almost always responses. If you get a question and
know the answer, you just take the time and answer it and
that's partof the job descriptionbecause you know you're
going to get it back." Retrievalalso occurs duringinformal
conversations that follow meetings or e-mails, which are fa-
cilitatedby IDEO'sopen office plan and (in Palo Alto)en-
counters that occur as designers walk between the seven
IDEObuildingswithin a three-blockradius. Informalconver-
sations often occur between designers who are known to
face specific technical challenges (who are expected to ask
for help) and designers who are known to have pertinent
expertise (who are expected to give help). A designer said:
I thinkthat people here feel reallyfree about just throwingthings
out just in casual conversationsin the halls. [Youask] "Oh, Lee, I
got a problem,maybe you have an answer for me." You stop and
help if you've been into brainstormerson some of these things be-
fore, and so have some exposure to it, or hear about it at the Mon-
day morningmeeting and thinkabout it, or if it's obvious what
someone is doing. You stop and throw your ideas out.
Analogicthinkingallows designers to take in specific imple-
mentations of technologicalsolutions yet retrieve useful so-
lutions for new problems that could not be predictedor that
are, in form, distant from the technologicalsolutions in
IDEO'smemory. The intimate knowledge requiredto enable
such generative metaphors requiresproblem-solvingarenas
in which communicationof complex problems and solutions
are possible. IDEO'sbrainstorms,other scheduled meetings,
e-mails, and informalconversations create such richcommu-
nicationsand allow the retrievalof specific technologicalso-
lutions that often take far differentforms than those in
which they entered the organization'smemory.
DISCUSSION
We blended network and organizationalmemory perspec-
tives in a model of technology brokeringthat explains how

one organizationdevelops innovativeproducts.This firm ex-

ploits its network position to gain knowledge of existing
technologicalsolutions in some industriesthat may be po-
tentiallyvaluablein others, but are rareor unknown. It acts
as a technology brokerby introducingthese solutions to in-
dustries where they are not known, and, in the process, cre-
ates new productsthat are originalcombinationsof existing
knowledge from disparateindustries.The organization'slinks
to many industriesprovideits designers with access to a
broaderrange of technologicalsolutions than they would see
workingin a single industry.Designers acquireand store
such solutions in the organization'smemory. Then, by mak-
ing analogies between new design problems and old solu-
tions they have seen before, they retrievesuch knowledge
to generate new solutions to design problems in other indus-
tries.
Because the primaryaim of this paper was to blend network
and organizationalmemory perspectives in a model of tech-
nology brokering,we have devoted limitedattentionto ex-
plaininghow IDEOencourages and supports employees to
follow the internalroutinesthat make such brokeringpos-
sible. Ourfield study suggests that the organization'srole is
critical.The structureof the work, norms for collaboration,
formaland informalrewardsystems, and employee selection
processes may help explainwhy IDEO,and perhaps other
organizations,have employees with the skilland motivation
to carryout these routines.
Organizational Support for Technology Brokering
First,the structureof work at IDEOcauses individualdesign-
ers to face a continualflow of new problems requiringengi-
neering solutions. This flow of new problems provides incen-
tives for them to develop, and opportunitiesfor them to
exploit, a wide-rangingknowledge of potentialsolutions.
IDEO'sclient base ensures that the organizationwill encoun-
ter a range of new problems and new solutions. Withinthe
organization,however, the engineers also do not specialize
in any single industrybut, instead, often move to new indus-
tries after completing a single project in one industry.Engi-
neers also often transferon and off long-termprojects to
prevent "burn-out"and allow them to pursue interests in
other areas. In a given year, an engineer may design por-
table computers, vacuum cleaners, medical products, and
office furniture.One projectmanager described this lack of
specialization:"As a designer you love varietyand, not hav-
ing to do the same thing for years on end, it keeps you fresh
and it makes you more confident that you can use some-
thing you learnedin this area and move from there." Workis
also structuredso that teams are formed and disbanded
aroundindividualprojects and often pull in additionalmem-
bers for brainstormsor short bursts of effort. This move-
ment between teams and projects not only provides design-
ers with a wide range of experiences, it also providesthem
with exposure to the skills and backgroundsof their col-
leagues. So the constant flow of new problems, combined
with the movement of engineers from team to team and
industryto industry,creates opportunitiesfor engineers to
develop variedtechnologicalbackgroundsand to learnabout
others' distinct knowledge and skills.

These varied backgrounds,and awareness of what others
know, enables the engineers within IDEOto act as individual
technology brokersby allowingthem to draw on their own
diverse technical knowledge to help others. Much of this
benefit, however, depends on IDEO'sstrong norms for de-
signers to share their disparateknowledge and help one an-
other. We have proposed that these norms (andthe associ-
ated values) can be summarizedas an "attitudeof wisdom"
(Suttonand Hargadon,1996). Buildingon Meacham's (1990)
writing,people who have an attitude of wisdom are coopera-
tive because they are neithertoo arrogantnor too insecure
to ask others for help and because they treat what they
know with humilityand what others know with respect. Fur-
thermore,wise people realizethat they know things that
others do not, so they constantlytell others what they know
and offer others help and advice.
Newcomers at IDEOare taught the attitude of wisdom at
IDEOand old-timersare remindedof it by the everyday in-
teractions of the designers, interactionsthat are most visible
in organizationalroutines such as brainstormingand Monday
morningmeetings. Brainstormsfoster technology brokering
at IDEOby pullingtogether groups of designers to work on
an identifiedproblem. Designers call these meetings to seek
the help of other designers at IDEOwho are not alreadyin-
volved in that project. In doing so, they demonstrate that
they are neithertoo insecure nor too arrogantto ask for
help, that they are treatingwhat they know with humility
and what others know with respect. One designer described
brainstormsas "useful in getting detailed knowledge about
your projectout so it stimulates others to suggest solutions
or offer leads." Anothersaid, "The main reason I use brain-
storms is to generate ideas that I know I wouldn't have on
my own." The designers who attend brainstormingsessions
do so because they believe they can contributedistincttech-
nical solutions to the problemand because, if they don't
help with others' projects, the favor will not be returned.
This same sense of sharingwas visible in the Mondaymorn-
ing meetings, when designers would announce problems
they were workingon (e.g., "Who knows about adhesives
for sheet metal?")or would present potentiallyuseful or in-
teresting ideas they had recently seen (e.g., handingaround
an example of a co-molded plastic handle with rubbergrip)
during"show and tell." These interactionsmade visible the
norms for asking for help, sharing knowledge, and giving
help, which taught newcomers and remindedinsiders how
they were expected to behave at IDEO.
Third,IDEO'sformaland informalrewardsystems provide
substantialsupportfor such collaboration.Top managers de-
termine designers' pay and responsibility,and while they
place weight on the numberof hours billed,compensation
decisions are based largelyon informalreputationamong
fellow designers and formal peer reviews. A top manager
described IDEOas a "peer-orientedmeritocracy,"so pay
and status are closely related. A designer emphasized, "the
only way to enhance your status in the organizationis by
earningthe respect of your peers." Designers do earn re-
spect from their peers through individualefforts that pro-
duced good designs, but a designer's reputationis based at

least as much on using his or her skillto help others. A de-

signer explained:
People realizethat the way to be respected and to get ahead is to
be out there. It doesn't work just to be a grindon your project.I
mean, you can do a great job on your projectand meet all your
goals, but I thinkthe rewardstructureis set up in a way that if you
don't participatein other stuff, that it's probablya little bit of a de-
merit. People don't know you as well so your capabilitiesaren'tas
well understood,you're not as likelyto be invitedon projects,
you're not going to be in demand. So there's a benefit for spread-
ing your knowledge and your skills aroundbecause you get to be
seen by more people and so you become more desirable.
A designer's reputationamong his or her peers is also en-
hanced by asking for help. Followingthe attitude of wisdom,
people who don't ask for help are thought to be either too
insecure or too arrogant,to lack humilityabout what they
know and respect for what others at IDEOknow. One engi-
neer compared his experiences at IDEOto other engineering
organizations:"Where I worked before, you just didn'task
for help. It was a sign of weakness. . . .At IDEO]we don't
have time to screw around.At the first hint I don't know
something, I'llask 'Does anyone know about this?' The
whole thing here is you've got to leverage as much as pos-
sible. You ask for help. You are expected to ask for help
here." There is especially low tolerance at IDEOfor engi-
neers who don't ask for help and then produce poor de-
signs. One designer asserted that makingmistakes was
viewed as an expected and inevitablepartof the design pro-
cess. Failedor weak design efforts that represented the
combined best efforts of a numberof IDEOengineers were
viewed as understandable.In contrast, designers who made
mistakes but had not asked for help were not easily for-
given. Other designers sometimes reprimandedthem for
failingto follow IDEO'smethodology, spread negative gossip
about them, and, if they repeatedlyfailed to ask for help,
shunned them by, for example, not invitingthem to attend
brainstormsor to work on interestingprojects.
Fourth,as with performanceevaluations,employee selection
is done by (future)peers who look for new designers with
the righttechnical knowledge and skill (includingback-
grounds that bringnew design solutions into IDEO)and an
inclinationto follow IDEO'swork practices and norms. A per-
son is not hired unless at least ten designers express strong
supportfor offering him or her a job. Furthermore,approxi-
mately 70 percent of IDEO'sengineers are graduates of
StanfordUniversity'sDesign Divisionof the MechanicalEngi-
neering Department.IDEO'sCEO, DavidKelley,is a tenured
professor in the departmentand, along with at least ten
other IDEOdesigners, teaches productdesign in this pro-
gram. A senior designer and long-timelecturerdescribed
IDEO'srelationshipwith Stanford:"I have twenty-five of my
students here now. And Kelley,I'm sure, can count more
like forty or fifty of the whole seventy. So we use, or have, a
nice relationshipwith Stanford."DavidKelleyand many of
IDEO'sfirst employees graduatedfrom this program,so
there is much overlapbetween IDEO'sand Stanford'sde-
sign philosophy.As one designer and teacher put it, "We're
not only following the philosophyof the Stanfordproduct
design program,we're setting the philosophy.It's not clear

who's drivingwho now." As a result, IDEOengineers (and
others) who teach these design courses socialize students in
IDEO'sdesign process and core norms. These IDEOengi-
neers are also able to select newcomers from a pool of stu-
dents whom they have observed workingon design pro-
jects. Promisingstudents often performsummer internships
at IDEO.A senior designer said, "It's a great way to inter-
view people. We just see these superstars coming through
and they see us and they want to work here and we love to
have them so we just grab them every year." In addition,
the remaining30 percent or so of IDEOengineers who did
not attend the Stanfordprogramusuallyfirst worked with
IDEOengineers as clients or contractworkers, so they
learnedabout IDEO'swork practices and core norms and
were screened carefullyby insiders before being offered
jobs. The result is that IDEO'sselection process not only
screens potentialnew employees for pertinenttechnical
skills and willingness to seek and offer help in the design
process, it may also serve to instillsuch knowledge and be-
liefs in designers before they are hired.
These preliminarydata from IDEOprovidehints about
means that other organizationsmay use to supportand en-
courage employees to act in ways that support internaltech-
nology brokeringroutines, so as to establish and exploit indi-
vidualties to distinctive knowledge domains. By structuring
work so that employees are exposed to a wide range of in-
dustries, individualemployees become well versed in di-
verse, and perhaps otherwise disconnected, domains and
the technologies within each. By constantlyformingand dis-
bandingteams, employees are exposed to the diverse
knowledge held by their coworkers and learnwho has what
kinds of expertise within their organization.By developing
and reinforcingstrong norms for exchanging informationand
for asking for and giving help, employees will feel comfort-
able asking for help, will know the rightpeople to ask for
help, and those who are asked will feel compelled to help.
By providingrewardsfor sharinginformationand helpingoth-
ers that are at least as great as the rewards for individual
accomplishments, employees will cooperate ratherthan
compete, or perhaps compete with one another over who
shares the greatest amount of pertinentinformationand
who is most helpfulon others' projects. Finally,by screening
employees for culturalfit as well as technical knowledge and
skill,and by relyinglargelyon what potentialemployees
have shown they will do instead of what they say they can
do, an organizationis more likelyto be composed of people
who act, individually,as technology brokersand who help
their coworkers do such brokering.
Directions for Future Research
Oureffort to blend network and memory perspectives sug-
gests that networktheory might be developed furtherby
devoting more attentionto the transformationand combina-
tion of ideas and resources as they flow through network
actors. The transformationand combinationdescribed in this
paper occurs predominantlythrough individualactions within,
and not between, such actors. Networktheory describes the
fragmented structureof knowledge across differentdomains
and explains the value of people and organizationsposi-

tioned as brokersin structuralholes. But this perspective

treats network actors largelyas conduits that pass along un-
changed ideas and resources to others. Littleattention is
devoted to if, how, or why those ideas and resources are
transformedand combined into new solutions for other ac-
tors and subgroups.
IDEO,like Edison's laboratory,does more than just transfer
technologicalinformationfrom groups where it is plentifulto
groups where it is dear. IDEOacquires such information,
stores it, changes it, and retrieves it to create new combina-
tions of old things. Throughthis process, brokerscreate new
value (and new knowledge) by adaptingand recombiningex-
isting technologicalsolutions in creatingthe specific forms of
new productsand processes that meet the needs of differ-
ent markets. Networktheory may be enhanced by future
research that considers how brokerswho span structural
holes change the ideas and resources that they transferand,
additionally,how brokers'abilityto add value to ideas and
resources helps them maintainand furtherexploit their net-
work position.
The purpose of an inductivestudy like the one reportedhere
is to guide and inspirenew ideas, not to validateexisting
ideas. The extent to which the local explanationof innova-
tion summarizedin figure 1 develops into a more general
theory of technology brokeringdepends on how well it, or
its descendants, explains innovationin other settings. One of
the first questions for future work on technology brokeringis
whether or not this local model resembles innovationpro-
cesses in other settings or is idiosyncraticto the firm that
we studied. The extent to which our model generalizes to
other organizationscan only be determined by hypothesis-
testing research in large, representativesamples of other
organizationsinvolvedin creative problemsolving. A variety
of existing cases suggest, however, that the process we ob-
served at IDEOis much like that used in other organizations
doing creative work.
Management consultingfirms like McKinsey& Co. and
Andersen Consultingprofitby bringingto client organizations
management techniques that clients were often not previ-
ously aware of but that have potentialvalue to solve their
currentproblems. For McKinsey,the result is reportedto be
a set of clients whose "demandfor organizationalknowl-
edge and experience cuts across nearlyevery importantcli-
ent relationshipregardless of industry"(Katzenbachand
Smith, 1993: 98). Takingadvantage of this position between
clients and the knowledge these clients seek requiresrou-
tines to acquire,store, and retrieve such knowledge in forms
that their clients can use. To do so, McKinseyuses formal
and informalmechanisms to store and retrieve potentially
valuableknowledge. One example is their "RapidResponse
Team," established to "respond to all requests for best cur-
rent thinkingand practice by providingaccess to both docu-
ments and experienced consultants"from across the com-
pany (Katzenbachand Smith, 1993: 98). ArthurAndersen
similarlybenefits from its abilityto brokernew solutions
based on past experiences. The firm's promotionalmaterials
promise to "quicklyproduce innovativesolutions" by draw-
ing on a knowledge base that "abounds with breakthrough

quantitativetools along with qualitativebest practices com-
piled from client experiences and exhaustive research,"and
they promise to do so by using the formalizedroutines cap-
tured in their GlobalBest Practicesapproach(Business
Week, 1996). As a result, these management consultants
providetheir clients with new solutions that are combina-
tions of (what they believe to be) the best management
techniques they have seen elsewhere.
Technology brokeringmay not be limitedto consultants. The
3M Corporationis a large manufacturingfirm that often finds
new uses for existing technologies by adaptingand introduc-
ing those technologies in new markets. The surface prepara-
tion technology known as "microreplication" that 3M origi-
nallydeveloped for overhead projectorsin 1964 has diffused
and evolved to provide3M with innovativeproducts in elec-
tronics (magneto-optics),adhesives (smartadhesives), abra-
sives (structuredsurface abrasives), reflective materials
(street signs and lane markers),illumination(lightpoles), film
(liquidcrystal displayfilm),and lenses (low-profileoverhead
projectors)(Stewart, 1996). Our local model of technology
brokeringmay offer insights into how organizationslike Mc-
Kinsey,ArthurAndersen, and 3M positionthemselves within
a network of imperfectlyshared technological knowledge
and how they acquire,store, and retrieve past technologies
for implementationin new designs for other industries.
Futureresearch might also focus on specifying the environ-
ments in which technology brokeringis likelyto occur. Our
perspective suggests that the primaryfeature of such envi-
ronments will be a fragmentationin knowledge and commu-
nicationbetween technologicaldomains. When ideas exist in
one domainthat are potentiallyvaluablein others, individuals
and organizationscan create innovativenew concepts by
acquiring,storing, and retrievingthese ideas in new combi-
nations and by transferringthese combinationsto new audi-
ences. Technology brokeringappears especially likelyto oc-
cur when new technologies are developed that have
potentialvalue in a wide range of industriesbut such knowl-
edge is not yet widely diffused. Examples include electro-
magnetic power at the turn of the century and information
technologies in recent decades.
The general applicabilityof electric power meant that inven-
tors like Edison could profitby findingproblems in a wide
range of industriesthat electric power could solve. In a simi-
lar,though less grand process, IDEOadapts information
technologies for use in a range of industriesthat had previ-
ously lacked such knowledge. The historianHughes (1989),
in language reminiscentof Marchand Olsen's (1976) gar-
bage can model of decision making,described the creative
process used by independent inventorsat the turn of the
century as solutions in search of problems. The presence of
technologies with potentialvalue to a broadrange of indus-
tries allowed organizationswith knowledge of these existing
"solutions"to create new products routinelyby crossing in-
dustry boundariesin search of new problems. ElmerSperry,
a contemporaryof Edison,whose firm pioneered the use of
electric motors in gyrostabilizersand gyrocompasses, stated
his rationalebehind this strategy: "If I spend a lifetime on a
dynamo [i.e., electric motor] I can probablymake my little

contributiontoward increasingthe efficiency of that machine

six or seven percent. Now then, there are a whole lot of [in-
dustries] that need electricity,about four or five hundredper-
cent, let me tackle one of those" (Hughes, 1989: 54). Theory
about technology brokeringmight be advanced by consider-
ing how, given the natureof the technologies involvedand
the distributionof knowledge about them, environmentsfa-
cilitate or hamper innovationthroughbrokering.
Ourevidence from IDEOalso suggests that internalroutines
are essential to organizationsthat exploit attachments to dis-
parate industries.Organizationsthat face many different
problems benefit from routinesfor acquiring,storing, and
retrievinga broadrange of technological knowledge when
that knowledge will be useful in solving future problems.
IDEOhas clear ties to a wide range of industries,but other
forms of organizationsmay hold similarnetwork positions
without sharingthe same organizationalforms. There may
be alternativeways of organizingfor technology brokering
that reflect differentenvironmentsand differentstrategies
and result in differentsets of internalroutines. For instance,
IDEOand other consultingfirms continuallyexplore the envi-
ronmentfor new solutions while solving the specific prob-
lems of clients in a range of industries.Others, like Edison
and Sperry,may innovate by specializingin a single, emerg-
ing technology and exploringthe environmentfor possible
applicationsof that one solution. Stillother organizations
may gain access to a range of industriesthroughmultiple
divisions and share discoveries (andfailures)in one industry
that may have potentialvalue in another. Large,multidivi-
sional corporations,for example, have internalizedaccess to
the technologies and marketneeds of differentindustries.A
divisionoperatingin one industrymay brokerpotentiallyvalu-
able technologies to other industriesby sharing knowledge
between divisions. Mueller(1975: 326) described the discov-
ery by DuPontresearchers of Duco lacquer,which reduced
the dryingtime of automotive paintfrom days to hours.
Seeking an improvedphotographicfilm, these researchers
recognized in a failed experimentthe potentialfor a new
productin a wholly differentindustry.By adding pigments to
a congealed and useless solution for photographicfilm, they
created a vastly improvedautomotive lacquer.DuPont's ac-
cess to such a broadrange of industriesand its internalrou-
tines for sharingproblems and solutions turned a failed pho-
tographicfilm experiment into a highlysuccessful
automotive product.Organizationsthat have developed rou-
tines for technology brokeringmay be better able to take
advantage of such serendipitous discoveries. Theoryon tech-
nology brokeringmight be enhanced by consideringthe rou-
tines that such firms use to match potentiallyvaluabletech-
nologies found in some parts of the organizationwith needs
in other parts.
CONCLUSION
Our model of technology brokeringsuggests that innovation
can and should be studied by consideringboth the social
structureof technological knowledge and the internalrou-
tines of organizationsable to exploit that structure. Innova-
tion through brokeringmay generalize beyond technological

innovationswithin the productdevelopment process. Scien-
tists, artists, management consultants, and others involved
in creative problem-solvingefforts often buildinnovativenew
ideas by recombiningexisting ideas. It is an old notion that
innovationsare builtfrom existing works, but the image of-
ten remains of the lone genius inventingideas from scratch.
Technology brokeringoffers a perspective on innovationand
innovatorsthat recognizes the value not of inventionbut of
inventive combination.
REFERENCES
Allen, Thomas J. DiMaggio, Paul Law, John

1977 Managing the Flow of Tech- 1992 "Nadel's Paradox revisited: 1987 "Technology and heterog-
nology. Cambridge, MA: MIT Relational and cultural aspects enous engineering: The case
Press. of organizational structure." In of Portuguese expansion." In
N. Nohria and R. G. Eccles W. E. Bijker, T. P. Hughes and
Attewell, Paul T. Pinch (eds), The Social
(eds.), Networks and Organi-
1992 "Technology diffusion and or- Construction of Technological
zations: Structure, Form, and
ganizational learning: The case Systems: 111-134. Cam-
Action: 118-142. Boston: Har-
of business computing." Or- bridge, MA: MIT Press.
vard Business School Press.
ganization Science, 3:1-19.
Fernandez, Roberto M., and Roger March, James G.
Basalla, George 1972 "Model bias in social action."
V. Gould
1988 The Evolution of Technology. Review of Educational Re-
1994 "A dilemma of state power:
New York: Cambridge Univer- search, 44: 413-429.
Brokerage and influence in the
sity Press.
national health policy do- March, James G., and Johan P.
Burt, Ronald S. main." American Journal of Olsen
1983 "Range." In R. S. Burt and Sociology, 99: 1455-1491. 1976 Ambiguity and Choice in Orga-
M. J. Minor (eds.), Applied nizations. Bergen, Norway:
Glaser, Barney G., and Anselm L.
Network Analysis: A Method- Universitetsforlaget.
Strauss
ological Introduction: 176-194.
1967 The Discovery of Grounded March, James G., and Herbert A.
Beverly Hills, CA: Sage.
Theory: Strategies for Qualita- Simon
1992a "The social structure of com-
tive Research. New York: Al- 1958 Organizations. New York:
petition." In N. Nohria and
dine. Wiley.
R. G. Eccles (eds.), Networks
and Organizations: Structure, Gould, Roger V., and Roberto M. Marsden, Peter V.
Form, and Action: 57-91. Fernandez 1982 "Brokerage behavior in re-
Boston: Harvard Business 1989 "Structures of mediation: A stricted exchange networks."
School Press. formal approach to brokerage In P. V. Marsden and N. Lin
1992b Structural Holes: The Social in transaction networks." So- (eds.), Social Structure and
Structure of Competition. ciological Methodology, 19: Network Analysis: 201-218.
Cambridge, MA: Harvard Uni- 89-126. Beverly Hills, CA: Sage.
versity Press. Huber, George P. Meacham, John A.
Business Week 1991 "Organizational learning: The 1990 "The loss of wisdom." In R. J.
1996 Advertisement for Arthur contributing processes and Sternberg (ed.), The Nature of
Andersen and Co. February the literature." Organization Creativity: 181-211. New
26, p. 85. Science, 2: 88-115. York: Cambridge University
Callon, Michel Hughes, Thomas P. Press.
1980 "The state and technical inno- 1989 American Genesis: A Century Merton, Robert K.
vation: A case study of the of Invention and Technological 1973 The Sociology of Science:
electric vehicle in France." Enthusiasm. New York: Pen- Theoretical and Empirical In-
Research Policy, 9: 358-376. guin Books. vestigations. Chicago: Univer-
Cohen, Wesley M., and Daniel A. Kantrow, Alan M. sity of Chicago Press.
Levinthal 1987 The Constraints of Corporate Miles, Matthew B., and A. Michael
1994 "Fortune favors the prepared Tradition. New York: Harper & Huberman
firm." Management Science, Row. 1994 Qualitative Data Analysis.
40: 227-251. Katzenbach, Jon R., and Douglas Thousand Oaks, CA: Sage.
Cyert, Richard M., and James G. K. Smith Millard, Andre
March 1993 The Wisdom of Teams: Creat- 1990 Edison and the Business of
1963 A Behavioral Theory of the ing the High-Performance Or- Innovation. Baltimore: Johns
Firm. Englewood Cliffs, NJ: ganization. Boston: Harvard Hopkins University Press.
Prentice-Hall. Business School Press.
Latour, Bruno
1987 Science in Action. Cambridge,
MA: Harvard University Press.

Mueller, Willard F. Rogers, E. M. Torrance, E. P.

1975 "The origins of the basic in- 1983 The Diffusion of Innovation, 1988 "The nature of creativity as
ventions underlying Du Pont's 3rd ed. New York: Free Press. manifest in its testing." In
major product and process R. J. Sternberg (ed.), The Na-
Rosenberg, Nathan
innovations, 1920 to 1950." In ture of Creativity: 43-75. New
1982 Inside the Black Box. New
R. R. Nelson (ed.), The Rate York: Cambridge University
York: Cambridge University
and Direction of Inventive Ac- Press.
Press.
tivity: Economic and Social 1994 Exploring the Black Box: Tech- Usher, Abbot Payton
Factors: 323-358. Princeton: nology, Economics, and His- 1929 A History of Mechanical In-
Princeton University Press. tory. New York: Cambridge ventions. New York: McGraw-
Neustadt, Richard E., and Ernest University Press. Hill.
R. May Schon, Donald A. Walsh, James P., and Robert D.
1986 Thinking in Time: The Uses of 1993 "Generative metaphor: A per- Dewar
History for Decision Makers. spective on problem-setting in 1987 "Formalization and the organi-
New York: Free Press. social policy." In A. Ortony zational life cycle." Journal of
Ogburn, William F. (ed.), Metaphor and Thought: Management Studies, 24:
1922 Social Change. New York: 137-163. Cambridge: Cam- 21 6-231.
B. W. Huebsch. bridge University Press. Walsh, James P., and Gerardo R.
Osborn, Alex F. Schumpeter, Joseph Ungson
1957 Applied Imagination. New 1934 The Theory of Economic De- 1991 "Organizational memory."
York: Scribner. velopment. Cambridge, MA: Academy of Management Re-
Harvard University Press. view, 16: 57-91.
Padgett, John F., and Christopher
K. Ansell Sutton, Robert I., and Andrew B. Weick, Karl E.
1993 "Robust action and the rise of Hargadon 1979a The Social Psychology of Or-
the medici, 1400-1434." 1996 "Brainstorming groups in con- ganizing. Reading, MA: Addi-
American Journal of Sociol- text: Effectiveness in a prod- son-Wesley.
ogy, 98: 1259-1319. uct design firm." Administra- 1979b "Cognitive processes in or-
tive Science Quarterly, 41: ganizations." In B. M. Staw
Petrovski, Henry (ed.), Research in Organiza-
685-718.
1992 The Evolution of Useful tional Behavior, 1: 41-74.
Things. New York: Knopf. Stewart, Thomas A. Greenwich, CT: JAI Press.
1996 "3M fights back." Fortune
Ribot, T. 1992 "Agenda setting in organiza-
Magazine, February 5: 94-99.
1906 Essays on the Creative Imagi- tional behavior: A theory fo-
nation. London: Routledge & cused approach." Journal of
Kegan Paul. Management Inquiry, 1: 171-
182.


Hargadon-1997-Technology Brokering and Innovation in A Product Development Firm

Uploaded by

Copyright:

Available Formats

You might also like

Hargadon-1997-Technology Brokering and Innovation in A Product Development Firm

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hargadon-1997-Technology Brokering and Innovation in A Product Development Firm

Uploaded by

Copyright:

Available Formats

Technology Brokering and Innovation in a Product Development Firm

Author(s): Andrew Hargadon and Robert I. Sutton

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

hole" (Burt,1992a, 1992b), a gap in the flow of information

By highlightingthe structureof resource flows across group

The notion that brokerstransformand blend informationis

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

50 to start-upcompanies. IDEOusuallycharges clients for

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

solutions that were new to that industrybut were used

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

about the nuances of how such processes unfoldwithin or-

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

useful informationfor designing new ones (Marchand Si-

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

organizational memory in forms that fit the new combina-

others.NtokPsto DOs raiainlMmr

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

0)cnc !E00 U)2coC cu

U E 0= >/ )Co C a)C;. Co o U c->

Ot C o CQ5co0E ocn- / EcC coU) * C)

726/ASQ, December 1997

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

-) C _r_ -0 aC 0" CC Co0

c~~~~~ a, C'- ~~~~~~~~~,5 a,. o CD.~a a)

a EZO a, ~CL MC Co- a, CU0 a) a~.- a, :

~~ aC~~~~~I5 E~~~~~ECU>'- - C-CaoC

727/ASQ, December 1997

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

can cause significanteconomic and competitive changes

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

products, manufacturing,and disk drives. In addition,IDEO

Most social networkanalyses measure currentsocial rela-

At IDEO,designers view their communityas a valuable

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

damental processes presented by Walsh and Ungson (1991):

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

ing, and ultimatelydesigning such productswhile workingon

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

seem to lie aroundIDEO'soffices as remindersof interest-

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

problems of the water bottle design and a previous project

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

one organizationdevelops innovativeproducts.This firm ex-

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

This content downloaded from 138.100.4.44 on Wed, 14 Jan 2015 03:46:32 AM

c~ a, C'- ~~~~~,5 a,. o CD.~a a)

aC~I5 E~~~ECU>'- - C-CaoC