All /
Why is a variation recording conversions for an element it doesn’t contain?
If you're using an Element Click conversion goal and notice a variation recording a small number of conversions for an element that doesn't exist in that variation, here's why, and why it's not a problem.
How On-Page Tests Work
In an on-page elements test, all variations are pre-rendered in the page HTML at the same time. AB Split Test then uses JavaScript to show the correct variation to each visitor while keeping the others hidden. This means that even when Variation A is showing, Variation B's elements still exist in the page's HTML and they're just visually hidden from the user.
Why Does This Cause Extra Conversions?
In rare cases, bot traffic using headless browsers (tools like Puppeteer or headless Chrome, commonly used by AI scrapers, monitoring tools and crawlers) can programmatically trigger a click on a hidden element even though real human visitors can't see or interact with it.
This results in a small number of conversions being recorded on a variation that doesn't visually contain the tracked element.
Example
- Variation A — original hero section, does NOT contain
#signupButton - Variation B — new hero section, contains
#signupButtonas the conversion goal
Even though #signupButton is hidden when Variation A is showing, a bot may still trigger a click on it, recording a conversion against Variation A.
Should I Be Worried?
No. This is statistical noise and has no meaningful impact on your results.
In real-world tests, bot-triggered conversions are tiny compared to real human conversions. For example:
- Variation A: 4 conversions / 95,000 visits = 0.004%
- Variation B: 1,100 conversions / 94,800 visits = 1.16%
The gap between variations is what matters and bot noise won't change which variation wins.
A few unexpected conversions on a variation without the tracked element = bots, not a bug. Your test results are valid.