Analyzing Social Media Youre Doing it Wrong

Brands often count numbers without understanding context. They focus too much on one channel while ignoring all others, or fail to gauge the collective opinions of the crowd. Instead they focus on a few overly positive or negative mentions, and stay too caught up in the moment to observe trends over time.

Each of these faux pas is easy to commit, and every one of them can have profound consequences. But it doesn't have to be that way.

Get to the Why

So your brand had 100,000 mentions in the last thirty days and 70 percent came from Twitter. Great. But what does this tell you? People are talking you, right? Taken a step further, 82 percent of tweets that expressed an opinion were positive. Better. Now at least you know the majority of things that were said were good.

Even still, this tells you very little about your brand’s positioning in the minds of consumers. If you compared your brand against another using the information above, how would these measures really help decision making? For the most part, they wouldn’t. While it is valuable to know where a brand stands in comparison to competitors, it’s much more powerful to know why.

This is where context comes in. Understanding important attributes within conversations, such as perceptions of cost or pricing, levels of customer satisfaction, expressed purchase intent or acquisition, and opinions of quality, actually help brands understand how they are positioned in the minds of consumers.

Context Inspires Action

So how do you actually get to context? Many of the more advanced social intelligence platforms operate on a flexible query language, some allow users to create custom filters, and a few even have themes and emotions built-in. Customizable filters and theme detection are of course the easiest, however, with some ingenuity, queries can also be adapted to uncovering deeper context (but that’s a whole different post).

After incorporating any three of the methods above, what do you look for? The first step is to establish your baseline with overall topic mentions, net sentiment, and any other available measures such as passion, gender distribution and/or demographic data. With your baseline established, consider the attributes that are key to the product, service, category and industry you’re researching.

Take luxury cars as an example. On the product side you might want to know the breakdown of conversations about specific components such as the suspension, motor, brakes, electronics. You might also want to know about perceptions of the dealership, where customers get maintenance performed and price.

With this information in hand you can ask questions: What percentage of the baseline does each attribute make up? What is the sentiment for each attribute? The passion? Gender distribution? When an attribute is present in mentions, what happens to purchase intent -- does it increase or decrease? Finally, how do the attributes of one brand compare against competitors? To really go the extra mile, ask what do millennials think compared to baby boomers, car enthusiasts compared to golfers, or moms compared to dads?

These conversations are happening on social media, and they don’t always align with the message the brand wants consumers to digest. By moving beyond the high level number of brand mentions and instead diving into the contextual elements, brands can move from filler to actionable intelligence.

Title image by Angie Harms (Flickr) via a CC BY-ND 2.0 license