Primary Research in the World of Unstructured Data Analytics

I get a bit uncomfortable when people oversell unstructured data analytics as a panacea for all customer insight ills. Don't get me wrong - I love the discipline and have spent great part of life advocating it. I still advocate it but what irks me is the thought being propagated that it is an alternative (probably cheaper) to some know your customer initiatives.

As a company in operation you should already have a fair "idea" about who your customers are and what they want, especially employees in your front offices dealing with the daily customer. If you do not have that, no amount of analytics can help because that analytics is as good as the base hypothesis and underlying model. (Though there is some merit in automated "discovery") The best way of knowing your customer is to serve her yourself; discuss openly across the table. First rule of customer centricity should also be to put yourself into customer's shoes. A well designed primary research program is a goldmine of specific actionable insights that you would not get elsewhere. Additionally, you can pick up subtle body language cues that are difficult to get otherwise. Sometimes customers say something in a survey while end up doing another thing. (even primary research has hypothesis, analysis models and mental models just like automated analytics)

Where does unstructured analytics fit in? (I include all kinds of social media as well as texts from internal call center etc. here). It was a happy world before social media kicked in giving voice, clout and control to end customers. Now there is too much data that cannot be humanly analyzed. You need automation, models to make sense out of it and probably make it actionable. This output should be used to either augment or correct an already developed understanding of customer and to quickly fix issues. It can also be used to validate the hypothesis or shifts in consumer behavior but in no uncertain terms can this be used as a sole source of customer insight and truth. It helps, supports, enriches, disproves, dissolves but never should create or initiate an insight. That task is better left to primary research.

Can primary research be done online (yes e.g. surveymonkey) and is the above true for all situations and all industries. I am not too sure if above thoughts completely apply to pure digital only businesses but they do apply to traditional ones.