The Arrival of Social Media Analytics: What Took So Long?

Social media has been one of the hottest trends in technology over the last decade. It has changed the way we talk to one another. We meet online. We converse online. We share online. And we complain together, online. Market researchers were quick to seize upon the opportunities provided by mass adoption of social media. Social media and “listening in” on the social chatter helps us better understand how brands are perceived, which points the way forward towards better campaigns and better targeting.

In a January 2019 article at, Michalis Michael reports that social media analytics for market research is trending upward and may be considered part of the mainstream, at last. This observation is echoed in a new market study featured on WiseGuyReports, which expects the Social Media Analytics Market, already a $2.5 billion market as of 2016, to grow more than 25% over the period 2017 to 2025. The question is, why has it taken so long?

First, let’s think about how social media analytics actually can help market researchers. The increase in the adoption of social analytics is a natural response to the relentless, undeniable rates of social media adoption among internet users worldwide. A social analytics process would implement a set of tools to perform customer segmentation & targeting, competitor benchmarking, multichannel campaign management, and customer behavioral analysis.

When you combine the monitoring of social media with the prowess of data analytics, you can achieve a virtually omniscient perspective, both wide-angle and telescopic views into the social space, where customers are candidly sharing their experiences. Social media analytics gathers intelligence about consumer choices and motivations in real time. It is like a master key that unlocks doors into how customers feel about brands and purchases. When people “talk” on social media, they do so in a freer, unsolicited fashion.  In a non-interfering way, social analytics watches the conversations and sharing behaviors, then assembles and analyzes them.

With this kind of intelligence, you are in better shape to assess brand positioning and marketing campaign strategy, generate stronger reports, gather more reliable competitive intelligence, and catch new trends before they become old news.  Think about how it applies to a retail context. Now, it is not farfetched to claim that social media listening reports will augment traditional retail measurement reports. Such reporting will be use machine learning tools to evaluate campaigns, brand reputation status, and identify influencers and leads.

It is a no-brainer to see why MR would take an interest in social media analytics, and years ago, analysts expected the techniques to take off rapidly, but the fact is, adoption rates have not gone as smoothly or as quickly. Michalis Michael’s greenbookblog piece explains why. It boils down to four key reasons.

First was the trust factor. Early social media monitoring tools were not as accurate as advertised, and market researchers were wary to buy in. Second was the superficial use of subscribed services, which contributed to less than useful results, and researchers assumed that social analytics didn’t really apply to their work. Third was the dissatisfying level of analysis. A researcher can rely on technology only up to a point, beyond which, human insights expertise is needed to make more sense out of the machine-driven analysis. In other words, the technology wasn’t mature enough to deliver the levels of quality promised to market researchers. Fourth was what we can call the inertia factor. Some market researchers were too stuck on the traditional methodologies they were most familiar with. Social analytics was too new, too untested, so they have been slow to adopt.

To be honest, the accuracy of social media monitoring technology HAS left a lot to be desired, with anemic reported accuracy rates at 60%, and lower. But accuracy rates can be improved when organizations invest properly in custom machine learning applications and effective curation of data (because if the data you are inputting into the machine learning model is irrelevant to the brand or campaign, you find yourself in the proverbial “garbage in / garbage out” jam). If social media analytics is really going to stay in the mainstream, vendors are going to have to prove that they can gather the right data and raise accuracy rates to the 80% range. Achieve that with reliability and consistency, and the researchers will come flocking en masse, with open arms.

Finally, it is wise to not set expectations too high. Social media analytics should not be conceived as a replacement technology. You still need good old human expertise, and the “old school” measurements of consumer sentiment and purchasing habits are not going to become passé anytime soon. What’s most needed is a hybrid strategy that merges the strengths of all analytical approaches. Find the role that social analytics can play best according to your needs, and let it add value to your existing reports.

Leave a Reply