How to Build a Signup Funnel in PostHog for a SaaS Site

A signup funnel in PostHog takes three things: the right events, a Funnel insight in step order, and a sensible conversion window. Then the steepest drop between two steps tells you exactly where to spend your next week.

Step 1: Capture one event per step

A minimal but honest SaaS signup funnel is four steps:

// 1. landing — autocaptured as $pageview, no code needed
// 2. user opens the signup form
posthog.capture('signup_started')
// 3. account created
posthog.capture('signup_completed', { plan: 'trial' })
// 4. first meaningful action (your activation moment)
posthog.capture('activated', { feature: 'connected_repo' })

Use stable, lowercase event names and keep them consistent — renaming an event later breaks historical funnels. Pick an activated event that reflects real value (a connected integration, a first project), not just "logged in again."

Step 2: Create the Funnel insight

In PostHog → Insights → Funnel, add the steps in order:

1. $pageview (optionally filtered to your landing path)
2. signup_started
3. signup_completed
4. activated

Set the conversion window to match your buying cycle — 7 days is a reasonable default for a trial product. PostHog shows the conversion rate between each step and the overall rate.

Step 3: Read the steepest drop

A funnel's value is the relative drop between steps, not the absolute numbers:

Step	Users	Step conversion
Landing	8,000	—
signup_started	1,200	15%
signup_completed	760	63%
activated	410	54%

Here the worst leak is landing → signup_started (only 15% even open the form). That points upstream to the page and CTA — not to the form itself. Always fix the steepest drop first; optimizing a later step with a small drop wastes effort.

The HogQL version (the unique part)

For ad-hoc or version-controlled analysis, reproduce the funnel with windowFunnel, which returns how many ordered steps each user completed within a window:

SELECT
  level,
  count() AS users
FROM (
  SELECT
    person_id,
    windowFunnel(7 * 86400)(
      timestamp,
      event = '$pageview',
      event = 'signup_started',
      event = 'signup_completed',
      event = 'activated'
    ) AS level
  FROM events
  WHERE timestamp > now() - INTERVAL 30 DAY
  GROUP BY person_id
)
GROUP BY level
ORDER BY level

windowFunnel(7 * 86400)(...) evaluates the conditions in order within a 7-day window and yields each user's furthest step (0–4). Grouping by level rebuilds the funnel counts — now in a query you can paste into a dashboard, diff in git, or schedule.

Keep it honest

Don't reorder events to flatter the numbers, and don't pick an activated definition so loose it always fires. A funnel is only useful if each step represents real progress toward value — then the steepest drop is a reliable map of where to work next.