Data Analytics Startup Topsy Aims for the Local

With Twitter users sending more than 400 million tweets per day, there’s no dearth of data flowing through the social pipes. That’s rich, potentially lucrative information from people all over the world.

That’s also a huge problem to any organization who wants to make sense of it — be it a brand, a government institution, or a hedge fund in Manhattan. For one, it’s a lot of data to sort through. And second, it’s not immediately clear just how to best use that data.

Enter outfits like Topsy, one of the companies chosen as a “Certified Product Partner” to Twitter and therefore given direct access to the “firehose,” the full stream of every tweet ever sent. Topsy, along with other similar services, offers a set of professional analytics tools for sale, so that customers can find key data points like influential Twitter accounts, relevant content for specific time periods, even sentiment analysis around tweeted terms.

The latest offering: Topsy is beefing up its geolocation efforts to help surface exactly where tweets are originating from in the world, based on factors like regular references to physical landmarks or events specific to a particular country, state or city.

So, to a brand aiming to find better, faster ways to reach audiences in the moments they need to be reached the most, that location data could prove valuable.

Say there’s a bunch of tweeting in Dover, Del., about donuts, for whatever reason. A brand like Dunkin Donuts can find that center of activity and figure out what’s causing that flurry of donut-related activity. From there, the company could buy a promoted tweet advertisement from Twitter, and target the ad specifically to users tweeting in that area.

“Instead of just being there when the lights go out at the Super Bowl, we’re trying to offer real-time engagement at a local level at any time,” said Jamie de Guerre, SVP of Product and Marketing, in an interview.

While helpful, the updated geo-inference model isn’t bulletproof. Since the tech relies upon repeat references to locations and landmarks, it’s susceptible to error (after all, I don’t have to live in London to make reference to Trafalgar Square). And as of right now, the best measurement is at the country level; Topsy can ID the country of origin in about 95 percent of tweets, but that number drops to 50, 30 and 25 percent when moving to the state, county and city levels, respectively.

But Topsy is betting that this location data will be far more useful to customers who don’t otherwise have the ability to zero in on tweet origins, perhaps giving brands a localized edge in appealing to new customers.

It’s worth noting that Topsy isn’t alone in the space. A number of companies in Twitter’s Certified Products Program offer analytics products, and Topsy competes with data resellers like Datasift and Gnip to provide that access to the Twitter firehose. But Topsy is positioning itself as an analytics dashboard provider paired with full data-reseller privileges, unique in its focus and the level of access it has to Twitter data. That could sway potential customers.

The updated tool is available to Topsy Pro customers beginning Monday.