How Dirty Is Social Data? An Analysis of Social Spam

By April 1, 2015 Analysis No Comments
DirtyData_Robot

Brands know that social media posts are a goldmine of consumer data that they can leverage for insight into the tastes, preferences and purchase intent of their target market. Even though savvy brands use social data to evaluate and inform their actions, they are being thwarted by non-consumer or social spam posts. Unless careful steps are put in place to remove social spam from their analysis, brands run the risk of missing or misinterpreting key consumer insights because of skewed data.

It’s a well known problem. Social media publishers have anti-spam measures in place, but they are fighting an uphill battle. But there are ways to identify and remove spammy posts from your data set and consumer analysis. Networked Insights’ models, for example, can identify social spam with accuracy greater than 80 percent making it possible to clean up the data set and achieve more accurate results.

The categories most weighed down by spam are those in which consumers make a lot of purchases, such as shopping, finance and tech. Significantly less spam occurs in categories such as religion, sports and science.

As indicated the chart below, some brands are particularly prone to spam, whether coupons, contests, giveaways and more. (For a complete itemization of what constitutes spam in this study, download the report below).

chart

The real issue for marketers is not spam count. It’s that all of that spam changes conversation themes that marketers discover using social data analytics., In a recent Networked Insights study, we used our discovery technology to analyze social data and organically uncover themes and topics that consumers care about. Then, we removed the spam to see how it may have affected the consumer insights that surfaced. The results were surprising – download the full report to see how drastically the conversation themes changed when we cut the spam. In effect, non-consumer posts actively warped the consumer insights that the social data revealed.

For more on how non-consumer content can harm brands (and how to avoid it), download the full report here.