English for PostHog Analytics

PostHog combines product analytics, feature flags, and session replay in one tool, which means its vocabulary spans several disciplines at once. Being precise about the difference between an “event,” a “funnel,” and a “cohort” keeps discussions with product managers grounded in what the data actually shows, rather than in vague impressions. This guide covers the core terms.

Key Vocabulary

Event — a single tracked action (like signup_completed or button_clicked) sent from the app to PostHog, the basic unit all analysis is built from. “We’re missing an event for the ‘export’ button, so we currently have no visibility into how often that feature is used.”

Funnel — a sequence of events analyzed together to measure the conversion rate from one step to the next, revealing where users drop off. “The funnel shows a forty percent drop-off between ‘added to cart’ and ‘started checkout’ — that’s the step we should investigate first.”

Cohort — a saved group of users defined by shared properties or behavior (like “signed up in the last 30 days” or “used feature X”), used to segment analysis. “Let’s compare retention between the cohort that used the new onboarding flow and the cohort that used the old one.”

Feature flag — a toggle that controls whether a feature is visible to a given user or group, used for gradual rollouts, A/B tests, and kill switches. “We shipped the redesign behind a feature flag so we can roll it out to five percent of users first and watch for errors.”

Session replay — a recorded, replayable reconstruction of a user’s actual session, used to see exactly what a user experienced without needing them to report a bug. “Instead of guessing what went wrong from the support ticket, let’s just pull up the session replay for that user’s session.”

Retention curve — a chart showing what percentage of users who performed an initial event return to perform a follow-up event over time. “The retention curve flattens out around week three, which suggests users who stick around that long tend to stay long-term.”

Common Phrases

• “Is this event actually firing, or are we missing instrumentation on that button?”
• “Where’s the biggest drop-off in this funnel, and do we have a hypothesis why?”
• “Can we build a cohort for users who hit this error, so we can see what else they have in common?”
• “Is this feature flagged, or did it ship to everyone at once?”
• “Let’s pull a few session replays before we assume this is a widespread issue.”

Example Sentences

Explaining a rollout plan in a design review: “We’re gating this behind a feature flag, starting at ten percent of users, and we’ll watch the funnel and error rate before increasing the rollout percentage.”

Reporting a data gap: “We can’t answer that retention question yet because the ‘feature used’ event isn’t instrumented — we’d need to add that tracking call before we can build the cohort.”

Discussing a UX problem with a teammate: “I pulled three session replays from users who dropped off at the payment step, and all three hit the same validation error on the card form — that’s likely the real cause, not general checkout friction.”

Professional Tips

• Say “event” specifically when discussing tracked data, and reserve “action” for describing what a user did in general conversation — mixing them makes it unclear whether something is actually instrumented.
• When reporting a funnel problem, name the exact step where drop-off happens rather than saying “checkout is broken” — it focuses the investigation immediately.
• Use “cohort” rather than “segment” when referring to a PostHog-defined group, matching the tool’s terminology in written reports.
• Mention explicitly whether a feature is behind a flag when discussing a new release — it changes how a bug report should be triaged.

Practice Exercise

1. Explain in two sentences why a funnel is more useful than a single conversion percentage for finding a UX problem.
2. Write a one-sentence report describing a missing event that’s blocking an analysis.
3. Describe, in your own words, how a feature flag and a cohort might be used together during a staged rollout.