Business intelligence and analytics processes are meant to expand the understanding of certain topics or situations, to then improve decisions and energize actions that take an organization to better and smarter places. Obviously bad decisions can be made when analytics are not rigorous, and fail to dish up insight and meaningful information. Bad analytics start with a poorly thought-out purpose, and are made worse by choosing unreliable, inappropriate and incomplete sources, particularly if multiple sources of information are needed to make intelligent decisions.
Too many analytics vendors of all sorts push “fast and easy” for analytics processes – when often quite the opposite must happen when first setting up such processes. Much thought, discussion and even research are often needed to fully understand objectives and the means to potentially get there. Are enough questions being asked? Are all of the stakeholders and those ‘in the know’ being consulted? Do we understand the shortcomings of certain data sources?
Even when big data analytics are focused on interpreting patterns of data – exploration without specific questions in mind – it’s essential to have parameters, comparison intelligence, metrics, and oversight by knowledgeable people, to determine if the analytics results are promising — or junk. This sort of data exploration is highly iterative which means wrong turns can be taken if only the data is considered.
Analytics targets that aren’t “black and white” concepts or topics can be the most problematic. When the problem or topic includes strong qualitative aspects, data analytics may only illuminate part of the story, and potentially can point in the wrong direction.
With the growth of social media, there have been increasing efforts to measure the notion of Influence – and then to identify the “best” influencers for the business focus de jour. The biggest obstacle for measuring influence is that influence is ephemeral, highly subjective and difficult to identify — let alone quantify. The nature of influence is qualitative – and continuously fluid. Influence isn’t just about the “influencer” but also about those who are influenced, the context for such influence, and whether influence even took place. Influence processes are multi-faceted, not one-way streets.
Many professionals are active on sites like Twitter and Facebook. Professionals on such sites frequently bring influence and credibility that mainly emanate from a large body of work that occurs outside these social sites. So there are problems with analytics that measure social site participation attributes without considering influence earned elsewhere. Yes, a Twitter presence and worthwhile tweets can be part of the “influence” story – but many people may follow these professionals on Twitter because of who they are and what they write in blogs or accomplish in client projects, not because of what they do each day on Twitter.
Professionals may also have a presence on Klout or Peerindex, sites that contend that they measure “influence” by tracking activity on social sites. Supposedly the more activity on multiple social sites, the higher the score. But this is “one-way street” analysis, not multi-faceted the way influence processes really occur.
Using social media sources for something like influence is messy and often introduces more questions than answers. The credibility of how the influence metrics are derived from sites like Twitter and Klout comes into question. It’s already been proven repeatedly that a Klout score can be gamed and often doesn’t reflect actual power or real influence. When considering individuals who may be deemed “influencers”: reliability, credibility, expertise, integrity, top-notch work are important parameters and all probably can be measured to a certain extent – but not necessarily as characteristics of a Twitter or Klout presence.
The heart of the matter of such attempts at analyzing the ephemeral is always: what value does this exercise bring? Often such attempts have an equally ephemeral shelf life of significance and relevance, even when credibility isn’t an issue.