Q: We added one filter and bot traffic dropped from 31% to 6% — which filter, and is that normal?

Q: We added one filter and bot traffic dropped from 31% to 6% — which filter, and is that normal?

A: Yes, that's a realistic result, and the filter was a headless-browser check (detecting automated browsers that render pages without a real screen).

Here's the actual setup from a media-buying team running pop traffic. Their tracker showed 31% of clicks producing zero scroll, zero mouse movement, and a navigator.webdriver flag set to true — the classic signature of automation tooling.

What they did:
— Added a JavaScript challenge that checks for webdriver flag + canvas fingerprint consistency
— Anything failing both got a 200 response but was tagged 'bot' and never counted toward payout

Result over 30 days: counted bot traffic fell from 31% to 6%, and their advertiser-side rejection dropped 14 points because the junk never reached the offer.

The 6% remainder was sophisticated residential-proxy traffic that no single check catches — that needs behavioral scoring over time.

Short version: a headless/webdriver check is the highest-ROI single filter; it won't catch everything but it kills the cheap automated noise.

Still stuck? Drop your case in the comments.