BuzzFeed wants to find a better way to weigh its viral success, and it wants to convert clicks to Pounds.

At the NewFronts presentations in New York today the company unveiled its new proprietary system for analyzing how content races across the the web — and sharing that information with its advertisers. Specifically, the “Process for Optimizing and Understanding Network Diffusion” is meant to shed a little light into the hidden corners of the social web. BuzzFeed has said that 75 percent of its 200 million monthly users are visiting the site through social media, so it would only make sense the company would want to better understand the patterns and habits of social sharing.

What exactly does Pound do? BuzzFeed publisher Dao Nguyen:

It follows propagations from one sharer to another, through all the downstream visits, even across social networks and one-to-one sharing platforms like Gchat and email.

Pound is the Process for Optimizing and Understanding Network Diffusion.

Pound does not store usernames or any personally identifiable information (PII) with the share events. Each node in the sharing graph is anonymous. We are not able to figure out who a user is by looking at the graph data. Pound data is collected based on an oscillating, anonymous hash in a sharer’s URL as a UTM code.

Instead of roping off social sharing into platform-based categories (how many came from Twitter vs. Facebook vs. Pinterest), Pound is designed to trace the way stories or videos are shared. It builds a pathway that shows all the routes a piece of content can take from being shared by one person over chat to exploding on Twitter.

What does it look like in action? BuzzFeed decided to examine the case of the infamous blue/black/white/gold dress that brought the Internet to a standstill in February. The initial post from BuzzFeed has been viewed more than 38 million times. Using Pound, they broke that down:

Nguyen said they plan to use Pound to gather more granular insights on the type of content that resonates with people and to optimize their sharing strategies. And since Pound made its debut at the marketer-friendly NewFronts, the company will use the tool to give advertisers a better picture of their audience. They’ve already used it to evaluate the effectiveness of sponsored content:

Pound data proves that sponsored content is shared just like editorial content. @Target's fanny basket content is great example. #NewFronts

— BuzzFeed Partner (@buzzfeedpartner) April 27, 2015

It should not be a surprise BuzzFeed wants to build a better machine for understanding how content performs in the wider world outside its own website, as the company has been laying out its vision for distributed content over the last several months.

In March, it was reported that the company, along with The New York Times and National Geographic, was considering hosting content on Facebook. At SXSW, BuzzFeed CEO Jonah Peretti offered a glimpse of the “cascade” of data that Pound provides. (Indeed, compare Pound to The New York Times’ similar-in-spirit Cascade project.)

Peretti told the audience then: “For us, it increasingly doesn’t matter where our content lives,” he said. “That can actually be a huge advantage.”

One measure of that success has been BuzzFeed’s success with video, reaching 1 billion monthly video views, with only 5 percent of that coming from BuzzFeed.com.