$ man posthog-gtm-signal

PostHog as GTM Signal

Product analytics turned into an outbound signal source. Watching what bounces, what dwells, and which pages prospects hit, then feeding that behavior back into plays.

by Shawn Tenam

analytics tools get treated as reporting: a dashboard someone checks monthly to feel good about traffic. that wastes the data. a prospect who spent four minutes on your pricing page is a warmer signal than any enrichment field you can buy, and it's sitting in your own event stream for free. the shift is treating PostHog as a signal source that plays can query. dwell time on a landing page tells you which account is actually reading. a bounce off the top of a page tells you the hook failed for that segment. pages hit in sequence tell you what a prospect is evaluating. all of that is targeting data, not vanity metrics.

PostHog events flow into my plays two ways. first, prospect behavior: personalized landing pages carry company context in their events, so when an account visits, dwells, and scrolls, a script picks that up and moves the account up the priority queue. deep dwell routes to a founder-touch sequence instead of the standard one. second, content feedback: I query dwell and bounce across properties to see which pages hold attention and which lose it, then the next batch of pages gets built from what held. the queries run through the PostHog API on a schedule, so the signal is a column in the dataset by the time the next play runs.