Social network analysis may well be emerging as one of the most important instruments in the alliance manager’s toolbox; if they can figure out how to use the tool.
It has become apparent for some time that collaborations are not only understood at the organisation level, but also at the level of networks of individuals working across organisational boundaries in any number of formal and informal ways. As analysis of the performance of alliances moves from the analysis of data aggregated at the organisational level towards the analysis of the relationships between the individuals involved in the collaboration, social network analysis tools may provide important insights into the weak points within the collaboration and more effective alliance management strategies. For example, any tool that promises to help identify where trust is weak in a collaboration is likely to be met with interest by alliance managers.
So what is social network analysis? At its most simple social network analysis proceeds by asking each individual involved in a network questions such as:
* How often do you talk with the following people outside formal reporting and management mechanism?
* How much do you typically communicate with each person relative to others in the group?
* How much do you trust each person relative to others in the group to follow through on commitments?
* Who do you typically seek work-related information from?
* Who do you typically give work-related information to?
* Who do you typically turn to for help in thinking through a new or challenging problem within the collaboration?
The data is then analysed and a network diagram is produced using various ratios. For example, the more people listing one of your team as a frequent contact divided by the total number of survey participants is a measure of that person’s centrality within the network.
It will become quickly apparent to alliance managers that these techniques are directly applicable to their role. For example, poor follow-through on commitments by one or a handful of a team could directly impact the trust between the teams if the individuals in question are central to the success of the collaboration. Likewise a research driven alliance should probably see a relatively high degree of contact between scientists on each team compared to a distribution alliances. If, on the other hand, the scientists aren’t talking when they should be talking that doesn’t seem to bode well for the future of the collaboration.
So if social network analysis is potentially so useful why has it not become a standard tool for alliance managers? We believe that there are a number of reasons for the flat uptake, so far, of social network analysis.
Bringing assessment down to the personal creates risks Asking your partners, for example, which of your colleagues they trust or do not trust brings a personal element into the analysis and the potential for conflict which many, if not most, alliance managers would rather avoid, even at the expense of useful data. Who wants to go to his colleague or even boss and tell him that nobody trusts him? To some extent this risk can be reduced by careful wording of the questions. But no matter how carefully worded they are asking partners questions pitched at the personal level that will often come into direct conflict with the team or corporate unity that executives want to present to external partners.
What should the network look like? Ask your partners to rate you on a scale of 1-7 and it is obvious that a higher rating is preferable to a lower rating. But what is the preferable network? Should there be lots of informal connections across the organisational boundaries or should the connections be restricted to key relationship management staff? Is a scientist with lots of connections with other scientists in other organisations a more valuable member of the collaboration team or simply a gossip or somebody looking for another job?
Lack of practical actionable output The output of most social analysis packages is geared towards academics and consists of dense networks of nodes, representing people, and connections. This type of output may look attractive in an academic paper but it has little or no practical application for alliance managers. Nor does it give alliance managers any practical tools for identifying areas where the alliance can be improved.
So do we believe that social network analysis is a dead-end for alliance managers? Not necessarily so. Some aspects of the insight that social network analysis brings can profitably be incorporated into partner feedback tools. However, social network analysis may have the greatest potential value where everybody agrees that a collaboration is running into serious problems and everybody agrees that something needs to be done to get the collaboration back on track. In that situation it may well be worth taking the risk of bringing the analysis down to the personal level in order to identify what is going wrong.
So we think that social network analysis may have a role as a Level 2 diagnostic tool. The partner healthcheck (the Level 1 diagnostic tool) is a useful tool for establishing that something is wrong with an alliance. However the healthcheck tool has some limitations when determining why the alliance has gone off the rails. This may be where social network analysis could into its own.