Viralheat: Our Sentiment API Is Getting 300M Calls Per Week

Viralheat is probably best-known for its social media analytics and publishing tools, but it also offers some free APIs. And at least one of those APIs, the one that analyzes sentiment, is taking off, according to CEO Raj Kadam.

He tells me that Viralheat originally intended to use another company’s sentiment API, except he wasn’t satisfied with what he found — everything focused on keywords, rather than using natural language processing to glean the true meaning of a tweet or comment. So the company decided to build sentiment analysis tools of its own, and to make those tools available to other businesses through an API. One example: “sick” means different things in different contexts, and Viralheat can supposedly understand that someone talking about “sick jeans” has a different meaning than someone talking about being “sick at the hospital”.

Now Viralheat says that the API is being used 1,500 organizations and developers in industries like finance and academia, and that it’s receiving 300 million calls per week. (Each call represents a single piece of text, such as a tweet, that gets analyzed.) All of that usage is making Viralheat’s sentiment analysis smarter, because companies can correct the analysis, say if the API declares that a tweet is neutral in sentiment but it turns out to be positive.

One company making extensive use of the API is Zuberance, a startup that helps businesses turn their fans into brand advocates. Kadam notes that for Zuberance, it’s not enough to find people who are tweeting a lot about your company. You specifically want to find the ones who are saying good things, and will be amenable to spreading positive messages about the company.

And yes, Viralheat is providing a lot of this analysis for free. Kadam says he isn’t interested in turning the APIs into a big part of the company’s revenue, but some of the heavier users (like Zuberance) are paying.