Content uploaded by Stephen P Borgatti
Author content
All content in this area was uploaded by Stephen P Borgatti on May 05, 2018
Content may be subject to copyright.
Content uploaded by Stephen P Borgatti
Author content
All content in this area was uploaded by Stephen P Borgatti on May 05, 2018
Content may be subject to copyright.
WHITE PAPER
Knowing What We Know:
Supporting Knowledge Creation and Sharing
in Social Networks
by
Rob Cross, Andrew Parker & Laurence Prusak
IBM Institute for Knowledge Management
August 2000
1
Institute for Knowledge Management
Sponsored by IBM and Lotus, the Institute for Knowledge Management (IKM) is a global consortium of
member organizations committed to understanding and developing tangible business value from knowledge
management.
This report was prepared for IKM members as part of a research project on “Social Capital: Networks.”
IKM Executive Director: Laurence Prusak
Editor: Don Cohen (doncohen@rcn.com); (617) 693-4606
Design and Production: AARTPACK, Inc. (www.aartpack.com)
Institute for Knowledge Management
55 Cambridge Parkway, LDB 3E
Cambridge, MA 02142
http://ikm.ihost.com
Copyright © 2000, IBM. All rights reserved.
2
Abstract
Many early knowledge management initiatives focused heavily on information
technology and codified knowledge and so missed performance improvement
opportunities from interventions targeting knowledge embedded within networks of
employees. Despite advanced technical solutions employed to manage organizational
knowledge, we continue to find that people are often more reliant on other people than
they are on databases when seeking answers to unstructured questions. As a result,
organizations creating more cohesive networks on knowledge related dimensions are
better able to collectively solve problems, create new knowledge and transfer explicit and
tacit knowledge embodied within employees. The following article reports on a cross-
industry research program assessing ways to promote knowledge creation and transfer in
networks of employees. Specifically, we have found four characteristics of relationships
important for knowledge creation in networks: 1) knowing what others know; 2) having
access to other people’s thinking; 3) having people be willing to actively engage in
problem solving; and 4) having a safe relationship to promote learning and creativity.
Mapping these dimensions in social networks yields targeted social and technical
interventions managers can employ to improve a network’s ability to create and share
knowledge.
3
Crafting an Answer
“So the call came in late on Thursday afternoon and right away I wished I hadn’t
answered the phone. We had received a last-second opportunity to bid on a sizable
piece of work that the Partner on the other end of the line really wanted to pursue.
Unfortunately I had little experience in [the subject matter] but happened to be the
one with availability at the time. I had no clue how to even begin looking for
relevant methodologies or case examples, so my first move was to tap into my
network to find some relevant info and leads to other people or databases. And in
fact I relied pretty heavily on this group of people over the next couple of days. For
example, Seth was great for pointing me to other people and relevant information,
Paul provided ideas on the technical content of the project while Jeff really helped in
showing me how to frame the client’s issues in ways that we could sell. He also
helped navigate and get buy-in from the client given his knowledge of their
operations and politics…I mean the whole game is just being the person that can get
the client what they need with [the firm’s] resources behind you. This almost always
seems to mean knowing who knows what and figuring out a way to bring them to
bear on your client’s issue.” --- Anonymous Interviewee.
The way in which this manager relied on his network to obtain information and
knowledge critical to the success of an important project is common and likely resonates
with your own experience. Usually when we think of where people turn for information or
knowledge we think of databases, the Web, intranets and portals or other, more traditional,
repositories such as file cabinets or policy and procedure manuals. However, a significant
component of a person’s information environment consists of the relationships they can tap
for various informational needs. For example, Allen (1984) found that engineers and
scientists were roughly five times more likely to turn to a person for information than to an
inanimate source such as a database or a file cabinet.1 In other settings, research has
consistently shown that who you know has a significant impact on what you come to know,
as relationships are critical for obtaining information (e.g., Simmel, 1950; Granovetter,
1973; Allen, 1984; Burt, 1992; Rogers, 1995; Szulanski, 1996; Shah, 1998) and learning
1 This finding has been recently replicated despite the explosion of distributed technologies. Cross (2000)
conducted in-depth interviews with forty managers in a consulting firm and found that, despite leading-
edge knowledge management technologies and practices, 85% of the respondents reported obtaining
information critical to the success of an important project through their network of contacts.
4
how to do your work (e.g., Lave & Wenger, 1991; Brown & Duguid, 1991 & 2000; Orr,
1996; Wenger, 1998).
Particularly in knowledge intensive work, creating an informational environment
that helps employees solve increasingly complex and often ambiguous problems holds
significant performance implications. Frequently such efforts entail knowledge
management initiatives focusing on the capture and sharing of codified experiences and
re-usable work products (Stewart, 1997; Davenport & Prusak, 1998; Davenport, De Long
& Beers, 1998; O’Dell & Grayson, 1998; Ruggles, 1998). To be sure these so-called
knowledge bases hold pragmatic benefits because they bridge boundaries of time and
space, allow for potential reuse of tools or work products employed successfully in other
areas of an organization and provide a means of reducing organizational “forgetting” as a
function of employee turnover. However, such initiatives often undervalue a crucial
“form” of knowledge in organizational settings: the employees and their network of
relationships that provide an ability to dynamically solve problems and create new
knowledge (Kogut & Zander, 1992; Nonaka & Takeuchi, 1995; Prusak & Fahey, 1998;
Nonaka & Konno, 1998).
As we move further into an economy where collaboration and innovation are
increasingly central to organizational effectiveness, we must balance our focus on the
technical and social aspects of knowledge management. Certainly we can expect
emerging collaborative technologies to facilitate virtual work and skill profiling systems
to help with the location of relevant expertise. However, as was so poignantly
demonstrated by reengineering, technology alone can only accomplish so much in the
pursuit of improving business performance (Hammer & Champy, 1993; Hammer &
Stanton, 1995). Improving efficiency and effectiveness in knowledge-intensive work
demands more than sophisticated technologies — it requires attending to the often
5
idiosyncratic ways that people seek out knowledge and solve problems in organizations.
Emerging collaborative technologies will be of limited use if not placed upon a
foundation of social relationships effective at knowledge creation and transfer.
With this in mind, we initiated a research program to determine social and
technical means of improving a human network’s potential to create and share
knowledge. The first step in our research was to understand the characteristics of
relationships that yield effective knowledge creation and sharing in human networks.
The second step was to map these dimensions of relationships among critical networks of
people in various organizations. Working with a consortium of Fortune 500 companies
and government organizations formed to study social aspects of knowledge management,
we developed empirical support for relational characteristics that facilitate knowledge
creation and transfer as well as insight into effective social and technical interventions to
improve knowledge creation and sharing in networks.
Relational Dimensions Facilitating Knowledge Creation and Transfer
Put an organizational chart (the formal structure) in front of almost any employee,
from line-worker to executive, and they will tell you that the boxes and lines do not
reflect the way that work gets done in their organization. These people intuitively know
that coordination and work often occur as a product of informal relations rather than
through traditional hierarchical channels or established policies and procedures. The
significance of these networks of relationships (or the informal structure) seems to only
be increasing today with trends toward ‘boundaryless’ organizations and telecommuting
(Hirschhorn & Gilmore, 1992; Kerr & Ulrich, 1995). Thus as we seek to promote an
organization’s ability to share and create knowledge we certainly can look to formal
organizational design features; however, we must also attend to the ways in which people
6
in networks become able to leverage each others’ knowledge. Social Network Analysis
(SNA)—a tool for mapping and analyzing relationships among people (or departments)
within an organization—offers a structural means to understand how knowledge creation
and sharing occurs within networks (Scott, 1990; Wasserman & Faust, 1994).
As a tool, SNA provides a means to represent networks where effective
collaboration can yield strategic advantage. However, we found limitations in the way in
which social network researchers have assessed “advice” or “information” networks to
date (see Monge & Contractor, 2000 for a review). First, aside from the concept of tie
strength, the network literature provides little insight into the characteristics of
relationships that promote effective knowledge transfer. While we have evidence that
ineffective relationships can impede knowledge transfer (Szulanski, 1996), we have little
understanding of the characteristics of relationships that make a network effective at
sharing and creating knowledge. Thus a first point of departure in our research was to
determine what makes a relationship effective in the creation of knowledge, because by
understanding this we could suggest interventions for networks.
Second, the traditional means of mapping an information or advice network has
been based on who helped whom with current or past tasks rather than who might be
helpful in tomorrow’s opportunities. However, the group of people that one knows and
has the potential to draw on for knowledge is not the same group of people uncovered by
network questions identifying who actually obtained information from whom in the
recent past. For example, it is often the case that people know of and are able to tap into
a much broader network of relationships for information or knowledge than whom they
have interacted with in the recent past for a given task. Thus, our second point of
departure was to better understand social and technical means of improving a network’s
capacity to collectively recognize and act on new opportunities.
7
With these goals, we asked forty managers to reflect on a project completed
within the past six months and identify all of the people who provided them with
knowledge critical to the successful completion of that recent project. We then asked
them to select the three most important relationships in terms of knowledge creation and
sharing and had them carefully describe both the relationships and the way they were
useful in helping to solve problems.2 3 Four features emerged that distinguished effective
from ineffective relationships: 1) valuing what they know4; 2) having access to them; 3)
having them actively engage in problem solving; and 4) having a sufficiently safe
relationship to ask important questions.5 Each dimension is described below.
Knowledge. The managers we interviewed often reported turning to contacts for
information because they considered those people knowledgeable in relation to some
aspect of the problem facing the manager. Thus, these contacts provided a critical
extension to the manager’s own knowledge when the manager had at least a semi-
accurate understanding of what her/his contacts knew. Relationships were valued for
2 Interviews generally lasted between two and three hours and followed a two-step process common in ego network
studies (Scott, 1990; Wasserman & Faust, 1994). First, the composition of each respondent’s advice network was
determined using a name generator technique (Burt, 1984; Marsden, 1990). Then the characteristics of each
relationship were further explored using name interpreter questions. In terms of theory development, we employed a
case-based logic in data collection by doing semi-structured interviews guided by a pre-existing theoretical model (Yin,
1994) that we held ‘loosely’ to allow for inductive theory development (Glaser & Strauss, 1967; Lincoln & Gubba,
1985). Our initial framework was informed by streams of research in social exchange theory and social capital, social
network analysis, transactive memory and distributed cognition, cognitive and social theories of learning and
communication studies. Interviews were transcribed, coded and assessed for inter-rater reliability using typical content
analysis procedures (Diesing, 1971; Lincoln & Gubba, 1985; Strauss, 1987; Miles and Huberman, 1994). A third party
independently coded interview transcripts with 93% agreement. To further empirically establish the importance of
dimensions uncovered in the qualitative work we conducted a follow on quantitative assessment. The relational
dimensions uncovered in our qualitative work accounted for 41% of the variance in a model of knowledge seeking even
after controlling for time in organization and formal hierarchical position.
3 So the sample was composed of 120 relationships.
4 This of course is subject to bias (Fiske & Taylor, 1984). However, one’s perception of another’s knowledge and
skills, even if inaccurate, informs who is turned to for what.
5 Readers interested in social capital may question why trust was not an explicit dimension. In the interviews, trust in
another’s intentions was NOT a defining characteristic of these relationships in terms of their effectiveness for
acquiring information, knowledge or advice. However, trust did come into play in two different senses. First, under
the definition of trust as predictability, we did find that being able to predict someone’s relevant knowledge in relation
to the problem at hand was important. Second, under the definition of trust as benevolence, the construct of safety was
also important in the learning potential of a relationship.
8
knowledge in two qualitatively different ways. First, some people were often sought out
for the specific knowledge they could contribute to some problem. Such people were
often skilled in technical domains and were sought out simply for the information that
they were likely to be able to provide. Second, other people were often not sought out for
specific knowledge in relation to a problem but rather for their ability to help think
through a tough issue. These people were tapped for advice in either defining or refining
complex problems and were considered good at identifying and making salient important
dimensions of such problems. For example, one manager indicated:
“At [Company X] we had access to background information and you know lots of case
studies and approaches that were really well written up. We had no experience in the
practice though of actually applying it on an engagement. So what was specifically
useful to me was to talk with Terry who helped me work some of this accessible content
into a workable approach…What I needed to know was: How might we apply this given
that we have not done it before. That was my key question. Rather than what do I need
to know about this subject matter.”
Access. Of course someone else’s knowledge is only helpful if they are willing
and able to make themselves accessible to you in a sufficiently timely fashion.
Interestingly though, knowing how to gain access to someone else’s thinking was not
always straightforward, but a learned feature of the relationship that was critical to its
effectiveness. Learning how to gain access to other people required one to develop an
understanding of a person’s response style and what medium was most effective for
establishing contact. Many relationships that evolved into highly productive ones were
initially very frustrating for the managers we interviewed as people often did not respond
in the depth, fashion or timeframe that the manager had anticipated. However, once these
managers learned a person’s approach to responding, and so had more accurate
expectations, they were better able to effectively tap into that person in ways that helped
them solve problems. This simple shift of expectations often dictated whether people
9
were effective in realizing the benefit of other people’s expertise or not. For example,
one manager indicated:
“I have gotten less frustrated the more I have worked with him because I have realized
that it is hard to get him to stick to a schedule. So now when I meet with him I have a list
of like ten things I need to get through. And we set up a meeting for 1:00 and I know that
it is not going to happen until 2:30 or 3:00 or maybe not until the next day, but when I do
see him I have my list and I am ready and we can run down it. It was important to learn
to accept that rather than be frustrated by it...”
Engagement. Yet access alone does not always result in the kind of dialogue
necessary to create actionable knowledge. More effective contacts tapped for knowledge
tended to willingly and actively engage in problem solving rather than overwhelm
someone with excessive information or offer solutions with little thought. It is important
to note that engaging in problem solving often did not mean a four-hour problem solving
session. Rather, it was a simple two step process whereby people would first ensure that
they understood the other person’s problem and then actively shape what they knew to
the problem at hand. In short, they were effective teachers. Most importantly, this stands
in stark contrast to other people our respondents indicated that often used their facileness
to get out of trying to be helpful. For example, one manager indicated:
“Some people will give you their opinion without trying to either understand what your
objectives are or understand where you are coming from or be very closed in their answer
to you. [She] is the sort of person who first makes sure she understands what the issue is.
I have been around people who give you a quick spiel because they think they are smart
and that by throwing some framework or angle up they can quickly wow you and get out
of the hard work of solving a problem. [She], for all her other responsibilities and stature
within the firm, is not like that.”
We found this aspect of engagement particularly interesting because it highlighted
an important dimension of knowledge transfer that is often overlooked. Often when we
talk about knowledge transfer we focus on the acquirer of the knowledge. For example,
Szulanski’s (1996) study, while acknowledging the difficulty created by an arduous
relationship, generally pointed to impediments to knowledge transfer from the
10
perspective of the acquirer of knowledge rather than the sender. This notion of
engagement raises an alternative dimension—that our energies might be better spent
attending to the sender of knowledge. Our interviewees indicated that a critical
behavioral difference between effective and ineffective people lay with first
understanding the problem and then shaping their knowledge to it. This is a behavior that
if developed among a network of people could significantly improve knowledge creation
and its effective diffusion.
Safety. Finally, asking for information can require the requestor to have some
degree of trust in the other person (Lee, 1997). Such trust often shapes the extent to
which individuals will be forthcoming about their lack of knowledge and concerns, as
defensive behaviors can knowingly and unknowingly block learning in critical
interactions (Argyris, 1982; Argyris & Schon, 1996; Edmondson, 1996). Virtually all
managers interviewed indicated that they felt safe asking other people for help, often
claiming that they were the kinds of people that were not afraid to admit their own
ignorance. However, the managers did say that the relationships characterized as highly
safe offered certain advantages in problem solving. First, they provided more learning
value as respondents were not overly concerned about admitting a lack of knowledge or
expertise. Second, several of the respondents also indicated that in the more safe
relationships they could be more creative. An important feature of these relationships to
them was that they were more willing to take risks with their ideas, with this often
resulting in more creative solutions. One manager indicated:
“[he] is always looking for the positive spin on something. I mean even if he thinks that
is crap and if he really thought that, he would always always find something positive or
he would say “Well I think we might be a little off track on that and heres why” and then
say why and of course there is learning that comes from that.”
11
Assessing a Network’s Potential to Share and Create Knowledge
The managers we interviewed indicated that knowledge, access, engagement and
safety were key characteristics of relationships that made them effective in knowledge
transfer. In contrast, they also recounted numerous times when knowledge transfer did
not happen due to one of the above dimensions not existing in the relationship (e.g.,
someone knew what they needed to know but did not make themselves accessible). With
the importance of these four dimensions established, the second step of our research was
to map each dimension within a critical network, such as research and development
initiatives or new product development efforts, to get insight into ways to promote an
entire network’s potential to share and create knowledge. Mapping the social networks
of groups that provide strategic advantage to an organization via their ability to share and
create knowledge helps us understand and propose specific social and technical
interventions to improve these networks.
This is an important advance over traditional network practices which, in terms of
knowledge sharing, have largely focused on past patterns of communication or advice. If
we have mapped a communication network and find that certain people are not as
connected as they should be it is difficult to tell what to do. Simply proposing more or
better communication is the oldest consulting recommendation in the book—and no one
today really needs more meetings. By taking a look at the aspects of relationships
underlying effective knowledge flow, we can offer more precise ways to improve a
network’s ability to create and share knowledge without overloading employees with yet
more meetings or e-mail.
First, one can analyze the knowledge, access, engagement and safety networks
separately to determine where a given group might be experiencing problems. For
12
example, if it is discovered that the knowledge network is sparse, it might make sense to
consider a skill profiling system or action learning sets—technical and social
interventions designed to help a network know what it knows. In contrast, if the access
network is sparse, then it might make sense to consider peer feedback or technical means
of connecting distributed workers (e.g., video conferencing, etc.) to make sure that people
within the network have access to each other in a timely fashion. The important point is
that each of the dimensions requires very different interventions to improve performance
(we expand on these below). As a result, analyzing the networks individually provides
more precise means of improving a network’s ability to share and create knowledge than
implementing a broad cultural intervention or distributed technology.
In addition, it is also instructive to assess the dimensions cumulatively to get a
better understanding of a network’s underlying knowledge creation and transfer potential.
In doing this, we can analyze networks where pairs of relationships exist (e.g., both know
and access) or networks where all of the relationships exist (e.g., know, access,
engagement and safety). For example, we conducted a social network analysis of 38
members within a practice of a consulting organization. The first network question asked
each person to indicate the extent to which they understood and valued the knowledge
and skills of their colleagues. This question was answered on a 5 point scale ranging
from strongly agree to strongly disagree for each of the other 37 people within the
consulting practice. A network diagram of the results is in the top half of Exhibit 1.
|Editor’s Note: Please Insert Exhibit 1 About Here|
The know picture demonstrates who in this network of people indicated that they
know and value other people’s knowledge and skills. Though relatively sparse compared
13
to similar networks we have assessed, this network does show a healthy core/periphery
pattern without distinct subgroups (which might represent a variety of political or
information flow problems). An interesting point to note in this diagram is the central
people — B, C, D, M and T — as these are the individuals most likely to be tapped for
informational purposes by the group.6 Just as importantly, it is interesting to note the
people around the edges of the network who are less connected (with one person being
completely isolated). Ultimately, these people are relatively less utilized by the group
and pose questions to management regarding why this might be.
The picture takes on added life when we also consider the access network where
each person rated their colleagues on the extent to which they were accessible in a
timeframe sufficient to help solve problems. Ultimately, both knowledge and access
relations must be present for information sharing behavior in a group to be effective. It
obviously does little good to value another person’s knowledge if you are not able to
access their thinking in a timely fashion. By combining the networks from these two
questions, we get a view of relationships that contain both dimensions — ones where
people both values what someone else knows and are able to access them. This network
diagram is shown as the Know × Access picture in Exhibit 1.
Several things are interesting in this network. First, we notice a fairly marked
decline in the number of connections among the group in comparison to the knowledge
network. While many central people remained central, it is important to note that several
people higher in the hierarchy shifted out to the periphery of the network. For example,
we now see that D, J and G (the three partners in this group) are all out on the periphery
of the network. Intuitively this makes sense (though it is often a surprise to those people
6 This assertion was validated by interviews in this setting and quantitative models of knowledge seeking in
several organizations.
14
in higher hierarchical positions). As people move higher within an organization their work
begins to entail more administrative tasks which makes them both less accessible and less
knowledgeable about the day to day more operational work of their subordinates. What
network analysis affords in this picture is an opportunity to assess whether those in positions
of formal authority are sufficiently central to the flow of knowledge, as well as to identify
those people that truly are influential knowledge brokers in the group.
The third question asked of the 38 consultants was who in the group they could
count on to actively engage in problem solving. The results of this question have been
incorporated into the Know × Access × Engage picture in Exhibit 1. Again, when the
engage network is added there is a significant decrease in the number of connections and
this is not trivial in terms of a network’s ability to solve problems. As outlined in the
initial interviews, it is often those people who are willing to engage in problem solving
that help both create actionable knowledge (rather than information overload) and ensure
that we are solving the right problem. In this view, we clearly begin to see the
importance of four people — B, C, M and P — for the network to leverage what it
knows. Just as importantly, there has now been a marked increase in the number of
people on the periphery of the network and five people completely disconnected.
Of particular interest at this juncture is the sub-group that has formed at the
bottom right hand corner of the network. The mere existence of a sub-group is not
necessarily a bad thing. On the one hand, a group that has splintered off from the main
network can represent untapped knowledge and occasionally political problems that must
be addressed. However, it might also be the case that to develop new products or
services management will make room and time for people to innovate as General Motors
did with their Saturn division or IBM with development of the PC at Boca Raton. A
common practice among many organizations seeking creativity is to allow such a group
15
to form and be creative outside of the requirements of day to day work and the pressure
that often exists to conform to current ways of doing things.
This was the case in this scenario. Roughly one year prior to this analysis, L had
been asked to develop a new service line in a technical network application. To do so he
hired several uniquely skilled people and spent a good bit of time pursuing development
of the service offering and sales opportunities — activities that had little to do with the
work of this particular group of consultants. As a result, this group had become isolated
from the main group over time, which in and of itself was not necessarily a bad thing due
to the group’s unique charter. What was a problem was that this group had become
linked back to the main network entirely through L. This put L in a particularly
influential position and also made the consulting practice susceptible to his leaving —
should he decide to go elsewhere this group would be largely detached from the workings
of the practice. This is not to say that new relationships could not be formed over time,
just that productivity would be greatly impacted as these relationships were renegotiated.
The final question we asked of this consulting practice determined whom each
person felt safe discussing work related issues with. The results of this question when
added to the three previous ones can be seen in the Know × Access × Engage × Safe
picture in Exhibit 1. With the incorporation of the safety network there is very little
change. This is because the safety network in this group was the most dense of all of the
networks. Ultimately, this is a sound indicator of the culture of this group for knowledge
creation and is obviously not a place we would look to intervene. It is also important to
note that based on our experiences a dense safety network is not always typical in these
scenarios.
Analyzing the combined network (i.e., Knowledge × Access × Safety ×
Engagement) provides a great deal of insight into who is critical as well as who is
16
currently less utilized within a group in terms of knowledge creation and sharing.
Understanding who is central to a group indicates people that might either be bottlenecks
or highly valued knowledge resources upon which the group is reliant. Only interviews
providing an in-depth understanding of a network can tell, but these people do pose
interesting questions to management. Has the group become too reliant on these people
should they decide to leave? Are these people hoarding information and so are
bottlenecks in terms of the group’s knowledge creation and sharing activities? In
contrast, should these people be rewarded for the somewhat invisible role they play in
supporting a group from a knowledge perspective?
If we discover that people are central in these networks for legitimate reasons,
management has an opportunity to begin acknowledging the work that these people do
for the group. In the words of one of the people central in the network in Exhibit 1, “I
spend about an hour and a half every day responding to calls and other informational
requests…[and] …none of that time gets seen in my performance metrics.” Network
analysis makes such interactions that are critical to a group visible, thus providing an
opportunity for management to acknowledge these people and the critical role they play.
For example, management might choose to better support knowledge creation and
sharing by offering central people such things as:
• Monies for efforts that might stimulate knowledge flow in a group via face-to-
face meetings, or to purchase technologies such as groupware.
• Cognitive and social space to allow room for both individual and collective
creativity and bonding to occur.
17
• Executive focus such as rewarding or promoting network enabling people to
both acknowledge their efforts and signal the importance of this kind of work
to others within the organization.
In addition to central or core individuals, we also find it important to better
understand why some people are peripheral in these networks. It might be that people in
these positions do not know what we thought they knew when hired. In these cases they
are peripheral for a legitimate reason and so reflect development or re-staffing
opportunities. Alternatively, it might be that these people are peripheral because they are
relatively new and the organization’s assimilation processes do little to help them
integrate into a network of colleagues. Given the increased turnover many companies
experience today, it is important to find ways to move people into the central part of the
network more and more quickly. This is often a process that can be improved by
focusing on the way that new people are brought into a group. At best, what most
organizations do when hiring a new person is to hold orientation courses that teach the
person about the computer system, their benefits and, perhaps, some homilies about the
culture and history of the company. It is rare to find practices where people are taught or
provided with opportunities to know what other people know in the organization.
Perhaps even more fatally, it is almost unheard of to find practices that teach the group
what individuals or other practices know. This is a critical shortcoming since, with work
increasingly being project-based, people will be brought into the center of the network
primarily as a result of what other people understand about their knowledge and skills.
On a more conceptual level, the combined network view offers unique purchase
on the elusive concept of organizational learning. Researchers in the field of
organizational learning have clearly captured the importance of declarative (know-what)
18
and procedural (know-how) knowledge in organizational learning (e.g., Cohen & Sproull,
1996; Moingeon & Edmondson, 1996; Walsh & Ungson, 1991; Walsh, 1995; Sanchez &
Heene, 1997; Walsh & Huff, 1997). In marked contrast, there has been far less attention
paid to organizational learning as a product of relational understanding of others or
know-who. Huber (1991: p. 89) claimed that an organization has learned when “through
its processing of information its range of potential behaviors has changed.” Thus if we
are interested in promoting an organization’s ability to react to new opportunities, we
need to account for the ways in which people in networks become able to leverage each
others’ knowledge. Changes in the knowledge, access, engagement and safety
relationships underlying a network’s future information processing behavior provides one
means of both descriptive and prescriptive traction on organizational learning.
Given the extent to which people rely on their contacts for information, this is not
a trivial question and at one level raises new ways of thinking about groups and their
performance in organizational settings. For example, rather than viewing senior
management as a cross-functional team, it may be more appropriate to consider it a rather
closed network that responds to opportunities presented by the environment. In this
view, more effective top management teams would be those that could configure more
effectively and efficiently around opportunities. Similarly, social network analysis can
help in understanding the impact that certain people may have on the trajectory of a
group over time. Organizations have often been claimed to be path dependent or
constrained by what they know. Such notions as absorptive capacity, core rigidities or
architectural knowledge have been claimed to lead to this path dependence over time
(Cohen & Levinthal, 1990; Henderson, 1992; Leonard-Barton, 1995; Arthur, 1996).
While critically important, this work has often been done at a level of abstraction that
makes interventions questionable. In contrast, the combined view of these networks
19
provides some idea as to precisely whose knowledge is primarily responsible for what a
group is likely to learn over time.
Promoting Knowledge Creation and Transfer
In applying these ideas in various organizations we have found two steps
particularly important. First, it is important to identify points of knowledge creation and
sharing within an organization that hold strategic relevance. The network of relationships
that one chooses to map is often a product of an organization’s strategy and how it
presents itself and its products to the market. Typical domains yielding benefit include
senior management networks, communities of practice and collaborative iniatives such as
new product development, R&D units or joint ventures and alliances.
It is particularly fruitful to map collaborative relationships that cross boundaries
of some form. Such boundaries might be hierarchical, functional, geographical or even
organizational as in joint venture or merger and acquisition scenarios. Understanding
how knowledge flows (or more frequently does not flow) across these various boundaries
within an organization can yield critical insight into where management should target
efforts to promote collaboration that has a pay off strategically for the organization. For
example, we mapped the relationships of one Fortune 500 organization’s top 126
executives to assess collaborative activity across divisions. This was an organization that
had grown by acquisition over the past several years with the primary intent that acquired
companies would go to market together. Thus the extent to which this group of leaders
was collaborating to integrate offerings held strategic significance for the organization.
We conducted a social network analysis of this executive network to assess
collaboration within and across divisions. While various network diagrams were
generated, the most insightful view came from a simple table demonstrating collaborative
20
activity amongst this network of executives. Exhibit 2 outlines a table of the percentage
of relationships that existed within and between each specific division (out of 100%
possible in each cell). Looking at the diagonal of the table we determined the percentage
of collaborative relationships that existed within the various divisions. This yielded an
opportunity to learn from practices within one division and apply them in others where
the work of each division required similar levels of collaboration. Similarly, we were
also able to determine which of the merged organizations (termed divisions in Exhibit 2)
had integrated well with other divisions. For example, a quick review of Exhibit 2 shows
that divisions 3 & 4 had a relatively high degree of collaboration; whereas divisions 1 &
7 had minimal contact.
|Editors Note: Insert Exhibit 2 About Here|
Though not a traditional network diagram, this view of collaborative activity
provided a great deal of insight into the inner-workings of the organization. The
company had acquired various organizations with the intent that they collaborate in
bringing their offerings to market. However, the social network analysis showed that
there was only limited collaborative activity in pockets of the organization. Various
reasons existed for this. In some settings members of the executive team were not sure of
what a given division did and so did not know how to even think about involving them in
their projects. In others, cultural barriers restricted people from seeking information
outside of their own division. And in some the complementarity of product offerings that
was presumed when an acquisition was made did not exist.
This kind of cross-boundary view identifies points where collaborative activity is
not occurring due to organizational boundaries and provides a more targeted view as to
21
where interventions might be employed. It is often not the case that you want high
collaborative activity amongst all departments within an organization. People have a
finite amount of time to put into developing and maintaining relationships. What
network analysis allows us to accomplish is to begin taking a portfolio approach to
considering what constellation of relationships are worth investing time, energy and in
some cases collaborative technologies in developing and maintaining. For example, in
the disguised scenario outlined above, it was not critical that Division 1 be tightly
connected to all other divisions to help the organization meet strategic objectives. To
provide strategic value to the organization, Division 1 really only needed to be well
connected to Divisions 3, 5 and 6. Thus, rather than engage in a company-wide initiative
to improve collaboration, more targeted and ultimately more successful interventions
were employed to facilitate collaboration at specific junctures.
Once a critical collective has been targeted, we have found that applying network
analysis based on the four dimensions of knowledge, access, engagement and safety can
improve a network’s knowledge creation and sharing potential. We further summarize
some of these ideas below.
Knowledge Dimension: How do we know what we know? Other people can only
be useful to us in solving problems if to some degree we know what they know. Even if
we are wrong, our initial perception will determine whether and how we tap into people
going forward. Problematically, most of our team interventions focus largely on shared
vision and process skills that help to create a harmonious environment, but do little to
educate team members of each others’ unique capabilities. This often does not help
people performing tasks to know how or when to tap into colleagues’ unique knowledge
and skills in future initiatives. This is a larger problem than one may think on first blush.
Studies consistently show that when work groups form to engage in a task they
22
experience what is called the unshared knowledge problem (e.g., Stasser, 1992 & 1995).
Rather than engaging in discussions that help them to learn the unique backgrounds of
individual members, they tend to focus on some domain that people have in common. Of
course this is a natural human process. By finding common areas of interest,
acquaintances or past experience, we are able to start a conversation with people we do
not know. However, the result is that we often never get to an understanding of other
peoples’ capabilities until extremely late in a given effort (if at all).
In more staid times, working relationships developed as a product of interaction
over longer time periods. This is surely not so in today’s business environment. As we
are working in an environment where we are increasingly forming and disbanding teams
on a project basis to accomplish work, we must pay more attention to the way that we
develop knowledge of others both within these initiatives and outside of them. If a
network is sparse on the knowledge dimension, then both social and technical
interventions designed to improve “knowing what we know” are warranted. For
example, on a technical front an organization might implement a skill profiling system or
a corporate yellow pages. On a social front, organizations such as the World Bank have
organized their employees into thematic groups that have Help Desks whom anyone
connected with the organization can contact. The individuals manning the Help Desks
are able to route people to others within the thematic group who have expertise on a
particular subject. Other companies and government organizations have regular
Knowledge Fairs where teams, communities or departments can set up a booth and
distribute information about the expertise that they have. Although this has limited scope
it has proven effective in increasing awareness of the projects and knowledge activities
taking place within the different departments and communities of the organization.
23
Access Dimension: How do we improve access to our collective knowledge?
Access undergirds the success of a network, especially in today’s environment where we
have little time to wait for written answers. Critical issues on which we may turn to
others for help or advice require turnaround of some sort within increasingly tight time
frames. For example, in the interviews of the forty managers in the professional services
firm, those relationships that people estimated were accessible in twenty-four hours were
by and large considered valuable relationships and ones that the person interviewed felt
was worth maintaining. This fell off sharply for those that were estimated to be
accessible in forty-eight hours, and those estimated to be accessible in seventy-two hours
were relationships that the manager no longer considered valuable and put no effort into
maintaining.
As organizations have become dispersed with people increasingly working from
diverse locations, improving access has largely centered on technical interventions such
as e-mail, cell phones, pagers, etc. Alternative ways of increasing access have included
focusing on the benefits of knowledge sharing within the company’s code of ethics. At
Buckman Laboratories all associates are empowered to speak with any associate at any
level; this is supported by a communication technology that gives each employee access
to all other employees. Access is also heavily conditioned by physical proximity and the
layout of workspace. Interestingly, Chrysler has gone full circle (from dispersion back to
co-location) and has brought all the people involved in new car development into one
building so that they can all have face to face access.
Engagement Dimension: How do we improve engagement in problem solving?
One of the most profound findings from our study of the forty managers was the
importance of a person engaging with another in problem solving. Ultimately, such
interactions led to actionable knowledge rather than information overload. Engagement
24
is a two-stage process of first ensuring one understands another’s problem and then
actively shaping what one knows to generate a solution—or teaching. British Petroleum
is one organization that has begun to systematically recognize the importance of teaching
by implementing peer reviews as an effective way to tap into others knowledge. Before
engaging in any significant task the individual or group invites peers to provide input.
Because the focus is performance, those with the most relevant knowledge and recent
experiences are tapped to participate. Through this peer review process not only is
performance on the task at hand improved, but people become much more aware of the
unique skills and abilities others can bring to projects. This creates a natural reason for
becoming familiar with the expertise of others. It also develops the needed norms of
reciprocity and trust that make sharing of expertise a natural process.
Technological interventions to improve opportunities for engagement have
included synchronous technologies such as VP Buddy at IBM, Same Time at Lotus or
white boarding applications that allow for dispersed engagement in a common problem.
Other successful interventions include the use of videoconferencing for visual interaction
between people in different locations. This has been particularly important at British
Petroleum where experts have been able to assist technicians who are working on oil rigs
thousands of miles away. Experts at BP’s offices have been able to engage in face to face
contact with the on-site technicians to analyze problems.
Safety Dimension: How do we ensure that learning and creativity occurs? In
terms of both creativity and the ability to ask questions that expose one’s lack of
knowledge, it is important to ensure that relationships are safe (at least from an
informational perspective). The first means of doing this is by increasing the visibility of
those relationships that are not perceived as entirely safe. Network analysis provides a
means of making these non-safe relationships highly visible and so discussible by the
25
group (under appropriate conditions). Using network analysis as a form of 360-degree
feedback (whether you disclose the names of people feeling unsafe or not) provides a
means of individually beginning to correct the problem.
Importantly, these interventions often take two markedly different forms. First, if
you see a network diagram where various people indicate that a specific person is not an
entirely safe relationship there are interventions that can be taken with that person. Often
such interventions just require creating an awareness, but may go in more depth as
necessary. Just as importantly though, it is important to look for the people who may be
claiming that a lot of other people are unsafe. In this scenario, someone claiming that a
variety of other relationships are unsafe in many cases may be reflecting their own
confidence levels.
Creating a greater degree of safety within relationships which allows for an
increase in learning and creativity is dependent upon allowing relationships to develop
over time. Although communications technologies such as e-mail are helpful in
maintaining relationships, when creating relationships it is more important to increase the
opportunity for face to face interactions between people. The World Bank has instigated
a program of brown bag lunches to encourage the development of relationships between
people. Another organization has encouraged face to face contact by monthly meetings
between different groups of researchers. These meetings consist of a discussion session
in the morning and a working session in the laboratory in the afternoon. This allows for a
free flow of ideas within the context of a real working environment. It also allows for
ideas to be experimented with which obviously leads to an increased opportunity for
innovation. Buckman Laboratories has used a more top down approach by creating a
code of ethics aimed at encouraging an atmosphere of confidence and cooperation
amongst all of its employees.
26
Conclusion
Where is the critical knowledge in the firm? Nelson and Winter (1982) proposed
that it lay primarily in organizational routines. We propose that it also lies in the
dynamic web of relationships existent in all organizations. A critical resource embedded
within organizations is the knowledge that workers bring to work on a day to day basis.
However, aside from human resource policies targeted to the attraction, development and
retention of identified valuable workers, there has been little effort put into systematic
ways of working with the knowledge that is embedded in people and their relationships
(Sveiby, 1997). Given the extent to which people rely on their own knowledge and the
knowledge of their contacts to solve problems this is a significant shortcoming. By
introducing social network analysis to understand how a given network of people create
and share knowledge, we are able to make these interactions visible and so actionable.
While social network analysis is not an entirely new concept, we have made its
application more tractable by offering specific dimensions of importance to better
understand how social networks create and share knowledge. A prominent sociologist
recently indicated that the emerging issue of social capital was to be the “killer app” of
network analysis. If this is the case, then rich and dense networks of relationships may
not only work as knowledge enablers within organizations, but also hold promise for
future assessment of the creation and value of social capital.
27
Exhibit 1
Know Network
Know × Access Network
28
Know × Access × Engage Network
Know × Access × Engage × Safety Network
29
Exhibit 2
Cross-Division Collaboration
Division
1Division
2Division
3Division
4Division
5Division
6Division
7Division
8
Division 1
33%
Division 2
5% 76%
Division 3
11% 18% 45%
Division 4
2% 11% 21% 38%
Division 5
6% 7% 12% 6% 75%
Division 6
7% 2% 13% 7% 2% 76%
Division 7
1% 3% 16% 6% 8% 2% 36%
Division 8
10% 2% 9% 6% 3% 10% 0% 90%
Avg.
6.0%
6.8%
14.3%
8.4%
6.3%
6.1%
5.1%
5.7%
30
References
Allen, T. (1984). Managing the Flow of Technology. Cambridge, MA: MIT Press.
Argyris, C. (1982). Reasoning, Learning and Action. San Francisco, CA: Jossey-Bass.
Argyris, C. & Schon, D. (1996). Organizational Learning II: Theory, Method and
Practice. Reading, MA: Addison-Wesley.
Arthur, B. (1996). “Increasing Returns and the New World of Business.” Harvard
Business Review, July-August, pp. 100-109.
Brown, J.S. & Duguid, P. (1991). “Organizational Learning and Communities-of-
Practice; Toward a Unified View of Working, Learning and Innovation.” Organization
Science, 2(1) pp. 40-57.
____. (2000). The Social Life of Information. Cambridge, MA: Harvard Business School
Press.
Burt, R. (1984). “Network Items and the General Social Survey.” Social Networks, 6,
pp. 293-339.
____. (1987). “Social Contagion and Innovation: Cohesion versus Structural
Equivalence.” American Journal of Sociology, 92, pp. 1287-1335.
____. (1992). Structural Holes. Cambridge, MA: Harvard University Press.
Cohen, W. & Levinthal, D. (1990). “Absorptive Capacity: A New Perspective on
Learning and Innovation.” Administrative Science Quarterly, 35, pp. 128-152.
Cohen, M.D. & Sproull, L.S. (Eds.) (1996). Organizational Learning. Thousand Oaks:
CA: Sage Publications.
Cross, R. (2000). “More than an Answer: How Seeking Information through Other
People Facilitates Knowledge Creation and Use.” Boston University Working Paper.
Davenport, T., De Long, D. & Beers, M. (1998). “Successful Knowledge Management
Projects.” Sloan Management Review, Winter, pp. 43-57.
Davenport, T. & Prusak, L. (1998). Working Knowledge. Boston, MA: Harvard
Business School Press.
Diesing, P. (1971). Patterns of Discovery in the Social Sciences. Chicago, IL: Aldine-
Atherton.
Edmondson, A. (1996). “Learning from Mistakes is Easier Said than Done: Group and
Organizational Influences on the Detection and Correction of Human Error.” Journal of
Applied Behavioral Science, 32(1), pp. 5-28.
Fiske, S. T. & Taylor, S. E. (1984). Social Cognition. New York, NY: McGraw-Hill, Inc.
31
Glaser, B. & Strauss, A. (1967). The Discovery of Grounded Theory: Strategies for
Qualitative Research. Hawthorne, NY: Aldine de Gruyter.
Granovetter, M. (1973). “The Strength of Weak Ties.” American Journal of Sociology,
78, pp. 1360-1380.
Hammer, M. & Champy, J. (1993). Reengineering the Corporation: A Manifesto for
Business Revolution. New York, NY: HarperBusiness.
Hammer, M. & Stanton, S. (1995). The Reengineering Revolution: A Handbook. New
York, NY: Harperbusiness.
Henderson, R.M. (1992). “Technological Change and the Management of Architectural
Knowledge.” In Kochan, T.A. & Useem, M. (Eds.) Transforming Education. Boston,
MA: Sloan School of Management.
Hirschhorn, L., & Gilmore, T. (1992). “The New Boundaries of the "Boundaryless"
Company.” Harvard Business Review, May-June: 104-115.
Huber, G.P. (1991). “Organizational Learning: The Contributing Processes and
Literatures.” Organization Science, 2(1): 88-115.
Kerr, S. and Ulrich, D. (1995). “Creating the Boundaryless Organization: The Radical
Reconstruction of Organization Capabilities.” Planning Review, 23(5), p. 41-45.
Kogut, B. & Zander, U. (1992). “Knowledge of the Firm, Combinative Capabilities and
the Replication of Technology.” Organization Science, 3(3), pp. 383-397.
Lave, J. & Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation.
Cambridge, UK: Cambridge University Press.
Lee, F. (1997). “When the Going Gets Tough, Do the Tough Ask for Help? Help
Seeking and Power Motivation in Organizations.” Organizational Behavior and Human
Decision Processes, 72(3), pp. 336-363.
Leonard-Barton, D. (1995). Wellsprings of Knowledge: Building and Sustaining the
Sources of Innovation. Boston, MA: Harvard Business School Press.
Lincoln, Y., & Guba, E. (1985). Naturalistic Inquiry. Beverly Hills, CA: Sage.
Marsden, P. (1990). “Network Data and Measurement.” Annual Review of Sociology,
16, pp. 435-463.
Miles, M. & Huberman, A. (1994). Qualitative Data Analysis (2nd Edition). Thousand
Oaks, CA: Sage.
Moingeon, B. & Edmondson, A. (Eds.) (1996). Organizational Learning and Competitive
Advantage. Thousand Oaks, CA: Sage.
32
Monge, P. & Contractor, N. (2000). “Emergence of Communication Networks.”
Forthcoming in F. Jablin and L. Putnam (Eds.), Handbook of Organizational
Communication (2nd Edition). Thousand Oaks, CA: Sage.
Nelson, R. & Winter, S. (1982). An Evolutionary Theory of Economic Change.
Cambridge, MA: Harvard University Press.
Nonaka, I. & Konno, N. (1998). “The Concept of “Ba”: Building a Foundation for
Knowledge Creation.” California Management Review 40(3), pp. 40-54.
Nonaka, I. & Takeuchi, H. (1995), The Knowledge-Creating Company. London, UK:
Oxford University Press.
O’Dell, C. & Grayson, C.J. (1998). If Only We Knew What We Know. New York, NY:
Free Press.
Orr, J.E. (1990). “Sharing Knowledge, Celebrating Identity.” In D.S. Middleton & D.
Edwards (Eds.), Collective Remembering (pp. 169-189). Newbury Park, CA: Sage.
____. (1996). Talking About Machines. Ithaca, NY: Cornell University Press.
Prusak, L. & Fahey, L. (1998). “The Eleven Deadliest Sins of Knowledge Management.”
California Management Review 40(3), pp. 265-276.
Rogers, E. (1995). Diffusion of Innovations (4th Edition). New York, NY: Free Press.
Ruggles, R. (1998). “The State of the Notion: Knowledge Management in Practice.”
California Management Review 40(3), pp. 80-89.
Sanchez, R. & Heene, A. (Eds.) (1997). Strategic Learning and Knowledge
Management. New York, NY: John Wiley.
Scott, J. (1990). Social Network Analysis. Thousand Oaks, CA: Sage Publications.
Shah, P. (1998). “Who Are Employee’s Social Referents? Using a Network Perspective
to Determine Referent Others.” Academy of Management Journal, 41(3), pp. 249-268.
Simmel, G. (1950). The Sociology of Georg Simmel. New York, NY: Free Press.
Stasser, (1992). “Discovery of Hidden Profiles by Decision-Making Groups: Solving a
Problem Versus Making a Judgement.” Journal of Personality and Social Psychology,
63(3), pp. 426-434.
____. (1995). “Expert Roles and Information Exchange During Discussion: The
Importance of Knowing Who Knows What.” Journal of Experimental Social
Psychology, 31, pp. 244-265.
Stewart, T. (1997). Intellectual Capital: The New Wealth of Organizations. New York, NY:
Doubleday.
33
Strauss, A. (1987). Qualitative Analysis for Social Scientists. New York: Cambridge
University Press.
Sveiby, K. (1997). The New Organizational Wealth: Managing and Measuring Knowledge-
Based Assets. San Francisco, CA: Berret-Koehler.
Szulanski, G. (1996). “Exploring Internal Stickiness: Impediments to the Transfer of
Best Practices Within the Firm.” Strategic Management Journal, 17 (Winter Special
Issue), pp. 27-43.
Walsh, J.P. (1995). “Managerial and Organizational Cognition: Notes From a Trip Down
Memory Lane.” Organization Science, 6(3): 280-321.
Walsh, J. & Huff, A. (Eds.) (1997). Organizational Learning and Strategic Management.
Greenwich, CT: JAI Press.
Walsh, J. P. & Ungson, G.R. (1991). “Organizational Memory.” Academy of
Management Review 16(1), pp. 57-91.
Wasserman, S. & Faust, K. (1994). Social Network Analysis: Methods and Applications.
Cambridge University Press.
Wenger, E. (1998). Communities of Practice. Oxford, UK: Oxford University Press.
Yin, R.K. (1994). Case Study Research: Design and Methods (Rev. Edition). Newbury
Park, CA: Sage.
