The other day, I posted about choosing the right social media listening tool for the right job. I usually don’t post again so quickly on the same subject, but I got so many questions that I thought I needed to. In my last post, I made the point that simple (and free) engagement tools (such as Hootsuite) work just fine if all you want to do is to monitor for crises (and you’re OK with watching many irrelevant results fly past as long as you eyeball the correct stuff). Any use of social media that is good enough for a person to look at individual social conversations can get by with just these free tools. But if you need to look at aggregate data to do market research or it is too expensive or error prone to have a person look at every tweet to find the relevant ones, then you need a better tool–one that incorporates text mining and probably that uses machine leaning technologies based on human-analyzed training data.
I serve as Chief Strategist for Converseon, a social media listening leader, so I might be a bit biased as to what is required, but the enterprise clients we work with (IBM, 3M, and many others) are looking for higher quality tools that help them glean the insights that help them make better decisions.
Technology only takes you so far before you need to invest in people that know how to analyze the data. It’s not much different from what I observed in the late ’90s with Web Analytics. The technology can collect the data, but unless you have a smart analyst telling you what it means and helping you decide what to do about it, all you have are some reports of visitors to your site–boring and unhelpful in improving your company’s results.
But people really want it to be automated. They want it to be automatic. Even Converseon is responding to this need by offering an API that anyone can hook up to for adding insights to their data. Now, the truth is that what the API does is based on lots of human analysis, but just like Google’s link analysis algorithm, if you can respond quickly enough, no one cares how long it took to codify that human knowledge. They just want it to be fast and accurate.
So, the bottom line is that you need to focus on how you want to use social media to choose the right tool. If you can live with inaccurate data because you are scanning everything with human beings to find the gems, use the cheapest tool that you can. If you’d rather have someone else pay the human being to get it right before people on your payroll look at it, then you need to make a different choice.
But make no bones about it. Somewhere a human being looks at the data. You need to decide when that happens, whose employee it is, and how much they are getting paid versus what the value is.
