"Social Listening" data is getting better and better. What should we do about it?
Historically, marketing analytics has always been ahead of the data. Certainly regression techniques have been around forever, but it couldn’t be applied to marketing until the data became available (and cheaper) to warehouse over time.
However, today there are plenty of providers who can provide “social listening” data: basically text scrubbed from thousands and thousands of blogs, tweets, and facebook updates. This text is then categorized into positive, neutral, and negative comments and quantified over time. It’s an unprecedented mechanism to get wide swaths of direct voice-of-customer, with much smaller risk of survey bias. Further, it’s importance is magnified because these do not just reflect the ideas of the individual, but one has to assume that blogs are read by at least one other person whose opinion is affected as well.
Net Promoter Score, or the metric obtained by asking solely whether a person would recommend a product or service to someone else, made a bit of a splash a few years back, though interest in still slowly growing. It’s based on the idea that personal recommendations hold much more sway upon someone than any advertising or marketing. This blog data offers exactly that metric, as these are actually people recommending (or not) your product to their friends, family, or readership. It is also better than survey-based metrics, as it is less prone to selection and self-reporting biases.
So can this be incorporated into a robust marketing accountability program? Yes, but it’s richer than that.
When used amidst other datasets in an analysis of sales drivers, this data should be treated as an “interim indicator” of demand. It isn’t sales, so it’s immediate dollar value is zero. However, it may indicate future sales- which can make it valuable to a marketing measurement plan.
But it’s real value is as an assessment of the other non-marketing factors that drive your business. How?
1) Real time product quality tracking. Imagine knowing at 10AM that there has been an upswing in defects in the southeast- how much bad PR could be averted? A few firms are already doing this real-time in airlines, and sports drinks, though with mixed results.
2) Assessing the impact of a competitor’s product launch– even before it happens. If there’s one thing about the internet, it’s that people don’t hold back with their opinions, both good and bad.
3) Finding the customer and how they use your product– many a creative agency brief has missed the mark due to a lack of understanding of who their consumer is. Even a quick scan through words nearby your product mention yields a brand association understanding that would cost millions via other traditional methods.
So while it’s not quite as convenient and easy as turning all that text into a single number and flipping it over to the analytics teams, the data can be invaluable when used in the right way.
PS – if you ever have the chance to have a beer with the people in this business, you won’t regret doing so. Ask them about the challenges of coding words like “ass” in text streams- whether a product “kicks,” “sucks,” “smells like” is critical to coding it as a positive, neutral, or negative impression. Further, whether a product is “shit” or “the shit” is certainly an important distinction.