The Social Media Analytics Expectations Gap

Social Media Analytics Expert Interview Series: Part 2 of 2 Last week ahead of the Social Media Analytics Summit, Useful Social Media published the second part of their Expert Interview Series. Incidentally, this week in the current issue of Inside Research ("The Bible of the Market Research Industry"), IR also published the results of a client roundtable on ‘Social Media - How is MR Coping?' in which the editor picked the client snippet of the month as "The problem is that there is more bad data being sold today than good data".

Reading these two I see some overlap in sentiment on both the good and bad issues related to social analytics. Too often when a client calls me to ask about our text analytics platform OdinText the question is "can you tell us what people are saying on the Internet". I have to explain that NO ONE can do this. Even if you could, why on earth would you put all your data into one tool and expect something useful?

There are no such expectations in regard to structured data. We are never asked to take all the surveys a client has fielded over the past couple of years and throw them into SPSS or SAS and analyze the results. While clients seem to understand that this doesn't make any sense with structured data, when it comes to text analytics, and especially with social media data, common sense and basic methodological considerations seem to be non existent for many.

Partially I think the problem stems from the marketing around social media that text analytics firms have perpetuated during the past few years. Many of these firms, perhaps because technologists and sales staff far outnumbered methodologists, have encouraged the black box/magical component to their products. As the field of text analytics continues to mature I'm hopeful that common sense will win out.

All text analytics, and therefore social media analytics as well, is not so much about "good" or "bad" data, it's more about the "right" data and the right tool/methodology to best solve your business need whether it's used on social media or survey data.