An A/B test goal defines what counts as a "win". Pick the wrong goal and you'll either declare a winner that doesn't matter, or run the test forever waiting for a signal that never comes. Here's how to pick well.
The three goal types
1. Pageview goal
Counts a conversion when a visitor lands on a specific URL. Use this for straightforward funnels — the classic example is /checkout/complete or /thank-you as the success page.
Good for: order completion, signup confirmation, content engagement (/blog/article).
2. Click goal
Counts a conversion when a visitor clicks a specific element. Define it by CSS selector — e.g. .add-to-cart or #newsletter-submit.
Good for: button clicks, link tracking, anything that doesn't trigger a pageview. Watch out for selectors that change with your theme — use a stable class or attribute.
3. Custom event goal
Fires when your code calls netaj.track("event_name"). Most flexible but requires a small JS change. See custom events.
Good for: video plays, scroll-depth milestones, in-app actions, anything that needs context (e.g. order value).
How to pick
Ask: what's the smallest measurable thing that, if it improves, means my change worked?
- If you're testing a hero image, "scrolled to product list" or "clicked add-to-cart" beats "purchased". Closer to the change = faster signal.
- If you're testing checkout, only "purchase complete" matters. A button-click win that doesn't translate to revenue is fool's gold.
Don't stack goals
One experiment, one goal. If you track three goals "to be safe", you'll find one of them looks significant by random chance — and you'll convince yourself it's real. Pick the goal up front, commit to it.
Next steps
Once your goal is set up, read about when to call a winner so you don't stop the test prematurely.