Skip to main content
View .mdOpen in ClaudeOpen in ChatGPT

Creating an A/B funnel for your app

~15 minutes · Intermediate

Most A/B tools only measure click-through rate on a single page. FunnelFizz's A/B split condition measures the whole funnel, including conversion to paid customer, which is what you actually care about.

Who this is for​

  • You want to know which hero copy actually converts, not just which gets more clicks.
  • You're comfortable with URL parameters and have a way to vary content based on them (server-side or client-side).

The scenario (outline)​

You have two hero copies:

  • Variant A: "See Your Entire Funnel. Fix What's Broken."
  • Variant B: "Read Your Customer's Mind. Know Their Every Move."

You want to know which one drives more trials — not just which one gets more clicks.

Steps (outline — full version coming)​

1. Serve each variant based on a URL param​

  • On your landing page, read ?ab= from the URL.
  • If ab=fox, show variant A. If ab=owl, show variant B. If neither, random-assign.
  • In your Twitter post: https://example.com/?ab=fox
  • In your LinkedIn post: https://example.com/?ab=owl

3. FunnelFizz auto-captures the variant​

  • The tracking script records ?ab= values as abVariant on every event from that visitor.
  • No extra config needed.

4. Split the funnel by variant​

  • Between AWARENESS and CONSIDERATION, add two splits:
    • ab = fox → branch "Fox variant"
    • ab = owl → branch "Owl variant"
  • Each branch now shows full-funnel conversion rate.

5. Wait for statistical significance​

  • You need meaningful volume per branch. Rough rule: ~100 conversions per variant minimum.
  • Watch the conversion rates. If one is materially higher after enough volume, that's your winner.

Full write-up with statistical-significance calculator coming soon.

Related: