Case Study

More fraud caught, fewer good players slowed down

Account takeover attacks on gaming accounts surged 78% in early 2026. This digital gaming platform had to stop more fraud without adding friction for the players buying in good faith. Pipl helped them do both. They auto-approved nearly a quarter of transactions with confidence, added $3.37M in low-risk approvals, and cut $340K in fraud from their riskiest segment.

The challenge

As purchase volume grew, fraud attempts grew with it, and so did the manual review queue. The platform’s rule set was under strain. Adding stricter rules was an option, but it came with a catch: every rule that caught more fraud also risked declining a real player at checkout.

Pipl Trust gave the team a calibrated score built on their own transaction history. That meant clear accept and reject signals in place of a growing pile of gray-area orders sent to review.

Their goals were clear

The goal was to hit all three at once, without trading one for another:

Catch more account-takeover fraud

Increase approvals for legitimate players

Cut manual review volume

Where the rule set broke down

The existing rules held up until volume and attack pressure scaled together. Then three problems stacked up.

Volume outpaced the rules

Fraud attempts and manual reviews climbed as purchase volume grew. More orders meant more cases the rules couldn’t settle on their own.

Account takeover was accelerating

Attackers were targeting gaming accounts at a sharply higher rate, putting pressure on a rule set that wasn’t built for that pace.

Stopping fraud couldn't risk approvals

Any new rule had to raise approvals too. Catching more fraud at the expense of good customers wasn’t an acceptable trade.

How the system changed with Pipl

Pipl calibrated a custom score on the platform’s own transaction history and validated it against known outcomes. The strongest signals came from two places: whether emails matched across the transaction, and how far apart the billing and shipping addresses were. With those signals scored, the team could set an accept line and a reject line, so the safest orders cleared quickly and the riskiest were stopped before they turned into losses.

The score was calibrated to the platform’s own transactions, not a generic model.

Email and address-distance signals separated genuine players from fraud.

74% of fraud was captured in a segment with a 12.9% fraud rate, 27.8x the average.

80% of good orders scored above the accept line, at 6.6x less risk than average.

Pipl's proof of signal intelligence

4.6x

more likely to be genuine when addresses and emails matched

6.6x

less risk among the 80% of good orders above the accept line

$340K

in fraud reduction from focusing on the riskiest segment

The real win? Clear decisions instead of manual review.

The team didn’t need more rules. They needed rules that could tell good players from bad actors. Pipl’s calibrated score sorted transactions precisely enough to support clear accept and reject decisions. That meant fewer orders sent to review, more legitimate purchases approved, and the riskiest activity stopped earlier.

Get these results for your organization

24.9% of transactions auto-approved with confidence

The highest-trust orders cleared without manual review.

$3.37M in new low-risk approvals

The top accept rule cleared legitimate purchases that would otherwise have been held or declined.

86.4% more fraud in the top reject rule’s segment

The reject rule concentrated fraud where the team could act on it.