Metrics, social networks and simplistic interpretations
A recent news item about Google’s attempts to use the social graphs of its employees to find new staff led me to ponder the likely impact of the growing availability of social media analytics.
This is a subject I have written about on a few occasions in my blog in Spanish, particularly in the context of social media analysis sites such as Klout, Peerindex, Kred, and others and their growing influence on the labor markets of some countries. We could say much about these types of metrics, but in light of the way that they have evolved, they cannot be dismissed as irrelevant or absurd.
The temptation to use somebody’s relationships to evaluate their ability or competence in certain areas is without doubt a procedure that is ahead of its time, as shown by some of the horror stories published by magazines like Wired in which somebody is excluded form the selection process despite their 15 years in the industry because they had never heard of Klout.
On the one hand, the progressively widespread use of the social networks does not mean that every professional uses them, and much less that any use would reveal much about their knowledge or experience of something, instead perhaps providing a very superficial outline of what we might call their social life.
On the other, there is a big difference between subjects: for some areas, for example those related to technology, the use of the social networks, or certain aspects of a company’s strategy are undoubtedly efficient when it comes to capitalizing on influence, other areas are seen as more “opaque”, and despite their importance are held in low esteem by most people.
If everything functioned on the basis of social mapping, the world would be made up of hundreds of thousands of social media ninjas, masters, gurus, and mavens and there would an alarming shortage of professional profiles. If we fell ill, we would find thousands of people prepared to tell the world about our sickness, but very few capable of helping us get better.
At the same time, it would be unreasonable to think that all these social reflections of what we get up to is simply smoke and mirrors. I have never felt particularly special because I have a high Klout score, or this or that number of followers on Google+, or a great many LinkedIn endorsements, but apart from being very grateful, I suppose that these collective perceptions at least mean that I have something to say about the subjects highlighted in that multidimensional image. That said, I am the first to understand that certain precautions need to be taken in interpreting those numbers, and above all, to avoid drawing simplistic conclusions.
Few of us are unaware, and I would be the first to say so, that the so-called pioneer effect can be profitable: having opened a Twitter account relatively early meant that mine was one of the recommended accounts when the application launched in Spanish, which meant that I went from nine thousands followers to around 80,000, as part of a process that obviously said nothing about my abilities, a concern I wrote about at the time.
I have experienced similar situations with other applications, and that tend to self-regulate over time — it would be absurd to think that a lecturer who few people would recognize in the street could, in the wake of the spread of the application and its use by people who really are famous, would continue to have the most followers. But I do think that you can’t fool all of the people all of the time, and that if a lot of people say that somebody knows a lot about a particular subject, or respect their opinion, then they probably do have something to say on the matter. A case of no smoke without fire, transferred to the social networks.
Conclusion: evaluating the importance of social metrics requires a large dose of common sense. Google seems to believe that when looking for people with a different attitude to life, which they describe as “googliness” and define as: “a mashup of passion and drive that’s hard to define but easy to spot,” a person’s social profile or inclusion among employee circles might be helpful. If you have a collective with certain attitudes, it’s not unreasonable to think that its relationship to other people will be influenced by these attitudes, and that this will create what researchers call “self-selection bias” toward candidates with similar profiles.
In the case of other companies, social metrics can help to indicate other things, such as influence or “resonance” in relation to certain subjects, how other people see the candidate, etc., all of which are worthy of careful study. A high Klout score, for example, means nothing in itself unless we subject it to close scrutiny to find out which aspects of it are relevant, a process that Klout itself provides the tools for. A low Klout score may mean that somebody is not particularly inclined to share matters related to their professional life on the social networks, but that doesn’t necessarily reflect their aptitudes, at least for the moment.
Are we headed toward a world in which everything we know, our professional abilities, our aptitudes, etc., will have to be reflected via society, that we will have to be validated by the group in order to be considered of value? The growing popularity of the social networks and the fast-growing trend of including all kinds of profiles within them, along with the development of specialized tools in each knowledge area would seem to indicate that we are. In countries with labor markets that really work, it is difficult to deny the influence of these types of social media analytics in selection processes, although, as we have already commented, their inclusion in such processes seems, for the moment to be based on relatively simplistic criteria.
It’s probably fair to say that this simplicity is a reflection of the newness of the process, and that it will correct itself over time: for some professional profiles, many of them belonging to people in senior management positions, it would seem that social media analytics are important. It is probably just a matter of time before they extend to other areas: these analytics are not entirely dissimilar to the relevance indices used among the scientific community. Perhaps now is the moment to begin thinking about it.
