Case Study: How Y Combinator lifted engagement 40% with PostHog experiments
Key results
The challenge
Y Combinator wanted to improve the matching algorithm behind its member networking product and understand user behavior more completely than general web analytics allowed. It needed a way to test changes and measure their impact on engagement.
The solution
Y Combinator used PostHog's experimentation suite to run an A/B test hiding member profiles that had been stale for 3, 6, 9, or 12 weeks, and used PostHog analytics to capture behavioral data that other platforms, including Google Analytics, were dropping to adblockers and third-party cookies. Results were measured against a control group.
“PostHog's experimentation suite is really great. We recently used it to improve our matching algorithm by running an experiment which hides profiles that have been stale for 3, 6, 9 or 12 weeks.”
CLCat LiProduct & Engineering Lead, Y Combinator
The results, in context
The six-week stale-profile experiment group sent 40% more messages and had 35% higher match acceptance than the control group. The story is published on PostHog's customer hub and dated October 5, 2022.