GoHighLevel Funnel AB Test: Step-by-Step Setup

Split testing in GoHighLevel funnels runs on cookies, not a separate testing tool. Open the live URL once, and a cookie locks you into whichever variant you landed on.

Refreshing the page never shows you the other version. That detail trips up a lot of people running their first GoHighLevel funnel AB test, since checking both pages in the same browser just shows the same one twice.

The feature ships on every plan. It lives inside the funnel step rather than a standalone dashboard, and setting one up takes a few clicks. Knowing what the results can and can't tell you takes longer.

Setting Up a GoHighLevel Funnel AB Test

The setup lives inside a single funnel step, not a separate menu. Open the funnel, select the step you want to test, and click Create Variation.

GoHighLevel gives three ways to build that page: duplicate the control, start blank, or bring one in from elsewhere.

Once the variation exists, a slider on the funnel overview sets how much traffic goes to each version. Every visitor still lands on the same main URL.

According to GoHighLevel's own documentation on split testing in funnels, the system diverts them toward the control or the variation page based on slider position. Fifty-fifty is the obvious default, but nothing forces it.

Before any of this works, a domain needs to be attached to the funnel, usually the first thing worth checking after moving a funnel over through our GHL setup and configuration work.

That's also why marketers testing their own page get confused. Open an incognito window, or a different browser, to see the other variant, since your regular browser already picked one for you.

What the Split Test Stats Actually Measure

Once a test is running, the Stats tab breaks performance down by a fixed set of metrics rather than one conversion number. If you're testing a step that follows an opt-in funnel page or leads into a checkout page, these are the numbers worth reading.

Those definitions come straight from GoHighLevel's own funnel and website stats guide. Nothing in that tab calculates significance for you. You're reading raw counts for the control and the variation, deciding yourself whether the gap is worth acting on.

Where the Feature Runs Into Limits

GoHighLevel's split test compares exactly two versions of one step. There's no built-in way to run three or four variations in a single test.

One agency owner raised this on GoHighLevel's own feedback board, asking to test more than one element at a time and label what each variant was testing, whether that's the headline, the copy, or the creative. It's still an open request, not a shipped feature.

The other limit shows up in conversion tracking between steps. Commenters on that same feedback board report that click-through tracking to the next funnel step only works reliably when the tested page includes a form or an opt-in element.

Build a split test around a content page with no form on it, and visibility into whether either variant pushed people forward disappears, a gap GoHighLevel's own feedback thread confirms is unresolved.

None of that makes the feature useless. A GoHighLevel funnel AB test works best scoped to one variable, on a step that already captures a form or a sale, inside a funnel already built out through our GoHighLevel sales funnel work.

Pick the one thing you're changing before duplicating the page. Let the test run long enough to see real traffic on both sides, and read the stats tab yourself rather than waiting for the platform to call a winner. It won't.

Cover photo by Hexxen STL on Pexels