For a company that has a theoretical market value of about $1.5 billion, BuzzFeed doesn’t get a lot of respect from the traditional media industry. Many seem to think of it as just a viral content factory, one whose output involves primarily cat photos and/or animated GIFs. In other words, not something to be taken very seriously.

This seems more than a little short-sighted. Anything that has 200 million monthly visitors and about five billion video views is probably worth taking seriously as a media outlet. Not to mention the fact that BuzzFeed appears to have built a fairly substantial—and serious—journalism operation inside that viral hit factory.

But even apart from its news chops, no one can deny that BuzzFeed founder Jonah Peretti understands content and how it functions on the Internet better than just about anyone in the media. And the key to that understanding—as a recent Fast Company feature on BuzzFeed points out—is data.

When BuzzFeed raised $50 million from Andreessen Horowitz in 2014 (it also got $200 million from NBCUniversal last year), there was a lot of skepticism about the argument that the site was more like a technology company than a media company. What kind of technology could possibly be involved in posting cat photos?

Peretti: We want to make media for the way the world is today

But BuzzFeed’s innovation doesn’t have anything to do with how content is posted, it’s about how that content is tracked across the multiple sites, platforms, and services through which it is distributed. More than anything else, BuzzFeed is the poster child for the multi-platform approach to media—according to the Fast Company piece, as much as 75% of the site’s output never appears on its website at all.

This seems to be paying off. According to internal documents released last year, BuzzFeed was already generating revenue of about $100 million in 2014 and was profitable.

The site’s commitment to understanding how, when, and where its content is shared, and by whom, became obvious when Peretti made Dao Nguyen—the former head of the site’s data and growth team—BuzzFeed’s publisher in 2014. If BuzzFeed started as Peretti’s viral-content hobby, it has grown up using some hard science.

One of the site’s main tools is called Pound. It measures how much a specific post, photo, or video is clicked on and shared. The latter of those two things is the most important, not the first. As Matt Stopera explains in the Fast Company piece, sharing is the biggest determining factor in whether something is successful or not.

If something has a 1.5 viral lift and 100,000 views and above, that was worth doing. It’s a failure if you have 400,000 views and a 1.1 or 1.2 lift. That’s a flop.

Peretti and others have talked before about the criteria that they believe causes someone to share something—things that trigger strong emotions such as love, joy, etc. And these criteria are applied to both the news content and to the sponsored content or advertising that generates BuzzFeed’s revenue. In effect, understanding how content is shared makes BuzzFeed a better advertising tool as well as a better media site.

Sign up for Data Sheet, Fortune’s technology newsletter.

In a nutshell, BuzzFeed has been building a giant engine for understanding content flows and sharing behavior online, thanks to tools like Pound and The Hive—a kind of huge, real-time database that can ingest and track BuzzFeed content wherever it is shared across multiple platforms.

Many media entities are good at producing content of various kinds, whether it’s news or entertainment or both. Some use tools like Chartbeat, Parsely, or Omniture to track what happens to it. But there are very few companies that understand the science of online sharing as deeply as BuzzFeed does. And at a time when every media company is becoming a distributed-content business whether they like it or not, that is a powerful weapon.