When last we saw our heroes, they were very busy eating at a rather chic spot in New York while managing to carry on a conversation, hold a baby, talk to friends from the Bronx, and gesticulate a lot.
I'm not sure we were the neatest table in the place.
However, it was at this point in the discussion that the question was raised: why do the results of network analysis sometimes go horribly wrong?
It's A Pain to Fill Out Forms
Many network analysts begin by asking employees to fill out surveys to list their skill sets. This can cause problems.
First of all, the surveys are often long and involved. As Nancy White pointed out, all the information must be present for the data to be truly useful to the social network analysis.
As Dean Landsman added, this problem can be mitigated either through financial rewards or by asking managers to begin filling out what they know about their teams' skills and interests. Once the data is plotted, more can always be added.
In fact, Dean suggested that the analysis will be more accurate by beginning with the managers' view of their team-members' skill sets rather than by having team-members fill out their own surveys. Any individuals' biases and blindness about his or her own talents can get in the way of gathering useful information.
But Do Managers Necessarily See More Clearly Than Their Teams?
In the same way that every employee is not a natural connector, every manager might not be a natural (or even trainable) expert at identifying areas of prowess in his or her own team. Just like their team-members, managers' views can be colored by their needs, personal issues, an their own set of biases. It's best, then, to invite a facilitator to walk the managers through such an analysis. The results will help ensure accuracy and identify the managers who might have enough natural talent to serve as facilitators as the network information is updated.
A Bigger Problem
As discussed at great length in other posts, often business leaders want to determine and measure the outcome of processes in advance. It's a shortsighted and dangerous strategy with social networking. Just as a creative person can't be cajoled into creating art on demand, the best connectors create networks as they appear to them.
As Nancy White says, this is exactly where the strategic challenge lies -- how much should people outside the network (say, those in charge) activate and push the networkers to do their job and how much should they stay away to let the connectors be free to do what they do best?
You can do a lot of damage with network analysis if you use the data to legislate from outside. Social networks don't work when they're treated as though they can be activated at will, as though by a switch. Connectors will not be able to do their best work under these circumstances. Yet, when is it a good idea to give some direction? And what is the best way to give it?
Networks as Organic: Members as Nodes
Dean added to this discussion that there are two important points to keep in mind:
1. Networks are made up of living people who excel at times and at other times will not. Furthermore, these people have responsibilities beyond the network -- to projects within their departments, to their lives outside of work, and so on.
2. Networks are made of nodes that do not work well with a hub and spoke model. The network must be flexible enough that if one node goes down, there are ways to go elsewhere or around it in order to get the information or skills you need. If someone in the network changes jobs or leaves, it shouldn't bring down the network.
So How To Make It Work?
Companies must change their values and compensation structures for social networks to foster innovation and optimum team-building. Reward people for the ways in which they disseminate knowledge rather than for the fact that they originated it or hold it.
Nancy gave an example in the Michael Smith Genome Center in British Columbia. Every piece of data has an RSS feed and is distributed. Raw data, as well as collections of analyses, are available to everyone in Canada (in the case of government funded projects) and to a smaller network for those that are funded by private organizations.
Scientists are rewarded not on the basis of published papers but on the quality of the data streams. Managers are rated on their ability act as filters -- to route the data streams to those who need them. Higher-level managers are rewarded on their ability to identify strategic opportunities.
The system reduces the power of seniority and raises the value of connecting ideas with those who are interested. This fosters innovation.
Social network analysis is an excellent tool to optimize a company's talent. It can help you graph where the skills lie, improve processes, and solve problems in new ways that lead to innovation.
However, networks are organic and require guardians whose talent allows them to grow in appropriate ways and to leverage resources for maximum value. However, leadership must rethink their reward structure and trust the network guardians to experiment in an informed way. This requires long-term thinking, experimentation, and the willingness to let go of control.