A: A fraud score (typically 0-100) is an antifraud tool's confidence that a click or lead is fake. Most teams set a blocking threshold without ever testing it, so here's a real before/after.
A lead-gen team blocked everything scoring above 80. Their advertiser still rejected 16% of accepted leads, meaning fraud was slipping under 80. They ran a two-week test dropping the block to 60.
Results:
— Volume of accepted leads fell 12%
— Advertiser rejection dropped from 16% to 5%
— Net approved leads (the ones that actually paid) rose 4%
So they blocked more, accepted less, and earned more — because rejected leads were costing them advertiser trust and a worse payout tier.
The catch: at threshold 50 they overblocked and lost good traffic. 60 was the sweet spot for that offer. The number is offer-specific; the method of testing it is universal.
Short version: test your fraud-score cutoff against real advertiser rejections — blocking more can earn more by protecting your approval rate.
Still stuck? Drop your case in the comments.
Clean Traffic Desk
@CleanTrafficDesk
Q: Lowering our fraud-score threshold from 80 to 60 — what did it actually do to revenue?
This post was published in the Telegram channel Clean Traffic Desk (@CleanTrafficDesk). All its posts are on the channel page.
Open on Telegram