For years, customer relationship management (CRM) data has been the bedrock of direct marketing. With the recency of social data, we’ve witnessed the birth of retargeting and predicting consumer behavior. But how exactly does big data relate to social analytics? Can social data ever stand on its own? And what can it truly tell us about consumers?

PluggedIn BD brought together industry experts in an attempt to answer these questions. Here’s what we learned.

Social can be a world of its own.

If you’re a brand, you have a choice: You can either use social analytics as a part of a family of data or you can hammer on that data individually. Generally, it’s more powerful when it’s integrated. We’re seeing brands retarget on Facebook with great success, especially when it comes to videos. And when we compile offline data, we’re able to use Facebook custom audiences, loading up an email and doing a direct match to target that way.

But we’re suffering from a lack of metrics.

The biggest question of social media since its inception has been: What’s the return on investment (ROI) of this? Ultimately, we want to know how many units have actually sold. We want to track sales. That’s what it always comes back to. However, because the media world is so fragmented and there’s very little tracking across platforms, we simply don’t have enough data on social.

Take Instagram, for example. We have very basic analytics to work with. Any measurement is going to be indirect. On Instagram, you’re only able to use links in the profile, and for any sort of metrics, you must put tracking on that link or use a special landing page. And if you’re using Snapchat for a campaign, forget it. In a few years, we predict that the platform will offer data, but right now we’re simply out of luck.

So how do we see a more complete picture of the consumer?

As hard as gathering data can be on social, it’s necessary — especially if you’re looking at younger demographics. Social has fragmented media well beyond what any of us thought possible just five years ago. The kids aren’t watching commercials; the kids are on Snapchat. There has been a huge shift in how people are consuming media and how we are able to reach them. We need to figure it out. And that’s where social comes into play.

  • Create a social fingerprint. While Facebook is a more personal space, on Twitter, you don’t always know your followers (and you certainly don’t know their birthdays). You can’t pinpoint identities. What you can do, however, is gain context. Just by looking at followers, you can likely guess their political affiliation, location, brand affinity, and (based on those brands) income level. Think of this as a social fingerprint.
  • Leverage influencers. There’s also a powerful influencer aspect to Twitter and other platforms where users are likely to follow celebrities. While it may be hard to determine exact ROI with social data, we do know that influencers drive sales.
  • Direct to a place where metrics exist. When you’re using a platform that has no data, you sometimes need to drive consumers toward a place where data lives. (The one caveat here is that you’ll likely lose some people along the way.) If done properly in an isolated campaign with the right targeting, social media can drive actions that can then be measured.
  • Use social data as an information gathering tool. This will be especially important as social media and the way people consume content evolves over the next few years. Brands need to steer away from being commercial or creepy with data, and use it to help consumers receive better products and entertainment. It’s a value exchange. Take Toyota, for example: Imagine that the company is coming out with a new electric car. Product development and testing can be done on social. The brand can start seeding design and gauge reaction on different platforms. From here, it can learn where it should be marketing the car and who is the target market. When using data, brands must aim to help rather than sell.

So yes, social data can be really challenging. We’re scraping it and it’s not adding up to much. But there’s still a place for it in the realm of big data.