Everyone knows you can learn a lot by trawling data coming from social media services like Twitter, Facebook and Flickr. But sometimes the data will surprise you. For instance, you’d expect to be able to glean product feedback from Facebook’s public feed, but did you know that shoplifters tend to brag about it in social media?
Chris Moody, COO of Gnip talked about exploring social data and the real-world use cases for some of that data at the Strata Conference.
Gnip is a major provider of social media data. It has deals with many of the major social media services and provides normalized data to clients, which in turn process the data and serve (according to Moody) about 90% of the Fortune 500.
The Data Cocktail
Much of this is what you’d expect. Companies trawling Twitter data doing brand monitoring, for instance. But some of the uses are a little less obvious, and this is where it gets interesting.
First, Moody covered how folks should decide what data they need to follow. Twitter, for example, is great for getting real-time feedback to events, but (obviously) is very concise. Conversely, WordPress.com provides more depth on topics, but at a slower pace. IntenseDebate and Disqus provide more concise feedback (usually, anyway, comments are shorter than the posts) but are even slower than posts to WordPress because comments by necessity follow posts. YouTube is slowest of all, usually, due to the time it takes to create videos but often provides really rich insights.
So most of the use cases need a “data cocktail” that combines two or more sources to get what they want.
For instance, Moody talked about the partnership with geodata provider ESRI to provide data to local retail companies. This combines Twitter, Flickr and YouTube to provide a “lens of what people say in store.” This goes beyond supply chain information, because they can see what people say about things that are not in stock and whether they care.
This is also where Moody mentioned the tidbit about shoplifters. He says that retailers can use social data to see when people are stealing from their stores. He declined to specify what’s done with that information, but you can infer that retailers are using it to tighten up security and pursue the shoplifters.
Another case study involved J.D. Power and Tropicana. Moody says that they were able to glean insight from blogs and comments about Tropicana that show many children of Baby Boomers view Tropicana orange juice as “a reward.” Using that data, they decided to start placing vending machines close to the exits at gyms.
Many businesses want to know where the next big box store is going up, because they depend on the construction dollars or providing other services to the Walmarts of the world. Moody says that, typically, businesses would keep an eye out for paperwork to be filed.
Now? A lot of times you can see in social media when Walmart has come up in meetings with town councils and such long before paperwork is actually filed.
All of this depends on having the appropriate mix of information, and knowing what to look for. Unfortunately, the session was short on technical detail, but if you’re looking for ideas how to use social data it was well worth attending.