One of the challenges of running a social media platform like Pinterest is the fact that people generate a lot of content. Deciding which pins to show, out of the thousands that could be shown, has a huge effect on the effectiveness of Pinterest marketing campaigns. The pins that get shown in the home feed have a greater chance of being repinned or clicked. To help business owners improve their marketing, Pinterest has explained how “Pinnability” is used to determine what gets shown.
Back in August of 2014, Pinterest switched from a simple, chronologically based display of pins (where the most recent pins from boards the user subscribe to are shown), to a non-chronological system that sought to show the pins with most potential for engagement.
“The home feed, a collection of Pins from the people, boards and interests followed, as well as recommendations including Picked for You, is the most heavily user-engaged part of the service, and contributes a large fraction of total repins,” wrote Yunsong Guo, a software engineer on the Pinterest Recommendations team. “The more people Pin, the better Pinterest can get for each person, which puts us in a unique position to serve up inspiration as a discovery engine on an ongoing basis.”
Like Facebook, Pinterest is trying to improve the user experience on the platform by showing the best possible content. Pinterest has done this by running candidate posts through a Pinnability filter. Pinnability is the collective name of the machine learning models the company developed to help Pinners find the best content in their home feed. The system estimates the relevance score of how likely a Pinner will interact with a Pin. With accurate predictions, Pinterest prioritizes those Pins with high relevance scores and show them at the top of home feed.
There’s a lot of advanced statistical math involved with this system, but it boils down to three major components to the system. First, there’s Training Instance Generation, which is when Pinterest’s system learns which pins the users is likely to click on and interact with. After generating this historical data, the system creates a Pinnability model that attempts to predict which kinds of pins would generate the most interaction in the future. Finally, when the user logs into their Pinterest account and request the home feed, the system creates a pin order based on the quality of the pins.
Thus far, the program has proven successful at improving the engagement and click through rate on pins shown on Pinterest. According to Guo, as the Recommendations team continued to refine Pinnability, they observed significant boosts in Pinner engagement, with each new iteration of the system. These boosts include an increase in the the home feed repinner count by more than 20 percent along with significant gains in other metrics, like total repins and clickthroughs.
Though marketers are often annoyed when social networks employ systems like this, Pinnability is a good thing because it rewards marketers for creating quality content and ensures that Pinterest remains a viable platform into the future. Additionally knowing how Pinnability works helps marketers craft better campaigns. Check out this article for some tips marketers can use to improve their Pinterest campaigns
For more Pinterest news, read this article about the newly open-to-all Promoted Pins.