Mopsos

Reviewing KM and the social network by Patti Anklam, I was wondering about the statement made that it creates new credibility for KM by providing executives with concrete data to illustrate the dynamics of informal networks. "SNA cracked the code of talking KM with the executive team. ", it says...

I am under the impression that there are major obstacle to the use of Social Network Analysis in corporate settings. It goes beyond the traditional legal issues regarding profiling of individuals, which incidentally are much more stringent in Europe than in the US.

The #1 problem with SNA surveys is that you actually ask people to give names. The #2 problem is that there can be tens of thousands of names to process. Employees will necessarily feel uneasy, to say the least, about answering questions about who is knowledgeable about this, and who is meeting with whom. –Why would the management want to analyze the social fabric of the company anyway ? Whose business is that ? Is that another trick for downsizing ? etc. I challenge the willingness of employees to participate in systems which coud be used to minimize the impact of their own eventual layoff!

I can only think of a few set of rules that would allow SNA to be performed in companies:

• Linking people together on the map should be an automated process, like analyzing the origin and destination of e-mails (a.k.a. Tacit Systems) or phone conversations.


• The interviewee owns his social network profile and it should not be used without his consent


• Only the names that the interviewee has actually declared could be seen on the map. Others would be hidden.


• Only aggregate results would be communicated to the management (no names)

Hi, Martin,

I appreciate your point of view, and I agree wholeheartedly that there are cultures in which SNA is inappropriate. Certainly, there are corporate and organizational environments when it would be unwise to use SNA. Downsizing is absolutely one of those environments. Responsible SNA practitioners would never accept to do consulting that was designed to identify individuals for downsizing.

On the subject of giving names, SNA does not necessarily ask people to "give the names of other people." In a basic survey, all the names in the organization are presented, and the people surveyed indicate the extent of their communications patterns. There are forms of SNA where people may be asked to provide the names of people they communicate with most frequently, and again it is the responsibility of the SNA practitioner to understand the cultural context and use the survey form that is most appropriate.

I agree with you on the use of electronic means for analyses of very large populations. This is where the emerging tools that mine the interactions of email, document access, or (increasing) social networking software become valuable. There is definitely a trend toward the gathering of data regarding connections in a number of these applications, which is key trend in SNA.

Context is everything. Whether determining the data gathering method, preparing the form for a survey, interpreting the results, or following up the results from the analysis of data gathered electronically, it is always important to understand the business goal, to communicate to the network these business reasons, and to treat individuals with the utmost sensitivity.

patti

Posted by: Patti Anklam at August 31, 2003 04:41 PM

Mr Roulleaux-Dugage,

Thank you for pointing out this paper, and also for making the work of Sebastien Paquet available. The socio-technical approach once disappointed me but this is the opportunity to learn more about it. The SMA paper and your comments ringed a bell and this led to a rather long contribution which I hope won't bore you over bounds.
I should make perfectly clear from the start that I do not know about US organisations social issues. I could not tell what can be tolerated or not by US employees as you could. I am concerned about French organisations.

I think this SNA is inherited from practices that go back a long time. In the late 1930's, Moreno established the rules of sociometrics, based on the use of the sociogram ("sociometrics" and "sociogram" are free translations, it may not be the exact spelling). Moreno first used this with children. He asked every child in a group (say, for instance, a class) who he would like to be friend with and who he thinks would want to be friend with him. Moreno then put the results in a table and draw a diagram of social relations in the group. These results could reveal different types of social roles placed on a continuum between 4 extremes (the ones "wanting" everyone and "wanted" by everyone, the ones wanting no one and wanted by everyone, the ones wanting everyone and wanted by no ones, and the ones wanting no one and wanted by no one). Sociometrics was then used as a base for group therapy (Moreno was one of it's founders) for which the sociogram was a starting point and the reference tool to assess the evolution of the group. Of course, there had been other questions than "who do you want to be friend with", but sociometrics is about the quality of relations inside a group, and mechanisms of acceptance in and rejection from a group. When dealing with adults, the typical questions are "who do you prefer", "who don't you like" and "who are you indifferent to". This may still be used in companies to put together small teams and to help manage conflicts, in a manner that a picture of an existing situation at a given moment would.

Interestingly enough, sociometrics appeared in the field of sociology of small groups (another free translation), a part of Psychosociology. This particular type of sociology is concerned with groups in which everyone knows (at least superficially) the other ones. This means that they know the other's names, their faces, they had the opportunity to chat with them and they remember those events. For each one, the other ones are persons, not just people.

Now apparently the ambitions of SNA is to use a well-known tool, the sociogram, widely outside it's theoretical bounds. This raises practical as well as theoretical difficulties. I think there certainly is a theoretical use for SNA, for research purposes. A practical use, well… that's what we are interested in, aren't we? And the constraints and limitations of this tool seem to rule it out. Note that the author, Patty Anklam, clearly expose some of them.

As a French cognitician (or cogniticist? could you help me on that?), the professional code of ethics lays for me a clear path trough the matter of confidentiality. Actually, my view is very close to yours. If I had to deliver a diagram like the ones that Patty Anklam gives as illustrations I could do it in only two situations, in France : 1) every person present in the diagram accepted to sign a paper describing the ins and outs of the study and giving the authorisation to display his/her name (as you said, the probability to obtain this is rather low) or 2) the diagram is drawn from the output of group meetings involving every person mentioned in the diagram. The thing is, group meetings about information exchange are themselves a very efficient action towards change, making the drawing somewhat useless and anyway outdated.

As it is presented, SNA is focused on individuals, with their personal history and personality. I believe than in this case, if a part of the team chooses not to participate, the validity of the whole description is forfeit. The usefulness of SNA lies in the taking into account of details. So I believe that the second and third rules you propose mean "not to use SNA".

As you say, what the consultant could deliver would be global results and/or conclusions about what to do to improve the situation, not the precise description of the situation. However, as the author Patty Anklam says, SNA is supposed to be but a step to other, deeper, analysis. And the only way she propose to cross this step is to show the detailed results, complete with names and all, to someone who knows, meaning "the managers". This apparently rules out your fourth possibility.

A system allowing to know who is writing to who and who is calling who would technically be a great solution. I can hardly imagine how perfect my job would be if I always had such complete and practical information within reach. And objective too. That would be a great help. But apart from the fact that it would certainly win the Big Brother Award, we could hardly do without subjective information, that is, how people see things. Nonetheless, that would be a great help, because we always have much difficulty in seeing what people do behind what people say they do.

Still, "what's the benefit after all"? Maybe a map leading to the discovery that some employee is new in the company, that some group of people worked in the same company before, or that a surprisingly strong proportion of employees regularly report to their manager, would be a bit frustrating. But I wonder if a computer aided scan of a large company, on a regular basis, would be a good start for some deeper analysis afterwards. I mean in setting a number of processing rules, it could points to possible "bottlenecks" or lack of communication between teams of any size, beyond the individual level (which means without personal information as the name). Provided there is no need to draw a map of course. Seeing clearly into a diagram with several thousands of nodes in a reasonable time may be beyond human cognition. What do you think of such a possible use?

Sincerely,

Yves Pierrot

PS : I juste noted Mrs Anklam is visiting the page. This could lead to interesting discussions. Actually, I hope so.

Posted by: PIERROT Yves at August 31, 2003 05:40 PM

Hello again,

I believe it is "rang a bell" not "ringed". Sorry.

Yves

Posted by: PIERROT Yves at August 31, 2003 05:42 PM

Whew! I could not imagine such a flow of very interesting comments. It gave me food for thought, and maybe I changed my mind a little bit.

In a big company like Schneider Electric, where I work, there are around 60 000 people in 130 countries dealing with similar problems. One of the most important challenge in the years to come (if not the single most important) is to make sure that employees that are doing similar things are indeed connected to one another, especially within communities of practice, where important knowledge is being shared and created.

After all, it may well be that SNA applied to a community of practice can be perceived by its members as far less threatening than if it were applied to a "managerial" organization, say, a country organization or a plant. A community of practice is based on trusted relations among its members, and a community leader has no right to hire and fire anyway.

In such a community setting, the output of SNA could be a "community networking index" that would certainly be a good indicator of the community's capability to respond to difficult customer issues and to innovate within its domain of knowledge. This would clearly be both valuable for the management and for community members to know this and to track the evolution of the index.

On a more personal and human note, I tried to make the point that whatever analysis is performed, the individual must always see the immediate benefit for him, stay in control of his/her destiny, and be respectful of others. No one individual should have access to a personal profile that is not his own, unless the owner of this other profile has granted him/her the right to see it. In a tightly bound community such as a family, a group of friends or a community of practice, it might be possible. In a bureaucracy, no way!

Posted by: Martin R. Dugage at September 1, 2003 01:11 PM

The original post has this: "Employees will necessarily feel uneasy, to say the least, about answering questions about who is knowledgeable about this, and who is meeting with whom."

About which I disagree. In our consulting work at Adaptive Path, we've done a lot of talking-to-employees research to understand how the company actually operates. And rarely are there obstacles to this. People love talking honestly about themselves and their situations, even when they're being recorded.

I say this because Martin's comment is not unique. A lot of people assume that folks won't want to give names and specifics. But you'll be surprised what you can learn by simply asking and listening.

Posted by: peterme at September 4, 2003 09:57 PM

I agree with you Peter. A new consultant hired by a company is usually perceived as unbiased, thus trustworthy. Employees give consultants they don't know the benefit of the doubt. They are usually pleased to talk, especially about their frustrations.
But I personally observed that consultants tend to wear out rather quickly. Whether they want it or not, they become sooner or later entangled in corporate politics, and trust fades away. I know people in my company who will never speak to McKinsey again!
Very often, hiring consultants sends a very clear message to employees that managers don't trust them to solve the problem themselves(a.k.a. Dilbert cartoons). Hiring consultants can be the right thing to do as a quick fix, but it can indeed damage trust and deplete social capital. Coaching, on the other hand does not.
There might be an interesting paradox in the very notion of "knowledge management consultancy" (and SNA consultancy in particular). Is it a viable business model?

Posted by: Martin R. Dugage at September 4, 2003 11:23 PM

A sister group, at BT Exact, used SNA as one of the techniques used to understand Internet usage in the home. Through a combination of large scale surveys, data gathered by instrumented PCs and a far smaller number of interviews, a picture of who families used PC to communicate (in combination of old-style telephony use). SNA was used to capture a person's web of contacts, and that then was used as a talking point, during the structured interviews.

Jump to http://www.btexact.com/ideas/papers?doc=73010 for more information.

Posted by: Nick Kings at September 8, 2003 11:43 AM

Thanks for signing in, . Now you can comment. (sign out)