Case Study

Bot orders caught, fan orders approved

Scalping bots were buying tickets under throwaway identities, and the rules built to stop them were turning away real fans too. A Trust Score calibrated to this platform's own order history gave the fraud team a way to tell the two apart: reject a narrow band of high-risk orders, and approve low-risk fans with confidence.

The challenge

Few categories attract automated buyers the way ticketing does. On this platform, scalping bots accounted for most of the fraudulent order traffic. Disposable inboxes and addresses that couldn't be verified let them submit orders in volume, each one arriving under a new identity.

The fraud team's defenses caught some of it, at a price. Rules strict enough to slow the bots also declined fans who had done nothing wrong, and each of those declines meant lost revenue and a buyer who might not return. The team needed a cleaner way to separate one from the other.

Their goals were clear

Three outcomes mattered, and progress on one couldn't come at the cost of another:

Stop more scalper and throwaway-identity orders

Say yes to more low-risk fans

Trade blanket manual review for targeted accept and reject rules

Where static systems broke down

Rules written for last season's bots were struggling to keep pace. The same three gaps kept appearing.

New identities on demand

Each disposable email or unverifiable address gave a bot a clean slate, leaving rules built on past identities little to work with.

Real buyers paid for the caution

On average, 12.09% of orders were auto-rejected, and genuine fans were part of that number. This causes customers to abandon their purchase.

Too little fraud for blunt rules to target

Fraud made up 0.08% of orders, so any rule broad enough to stop it also stopped a large share of good orders.

How the system changed with Pipl

Pipl built a Trust Score around this platform's historical ticket transactions, linking each past chargeback and approval to the identity signals present at checkout. Email and IP signals did the most to distinguish bot purchases and disposable-email fraud from ordinary fan activity. The score was then translated into two decision thresholds the team could apply alongside its existing rules.

Each historical order was scored against its known outcome, chargeback or approval

The gap between billing and IP locations became one of the sharpest risk signals

Email age and breach history identified fans with long, clean digital histories

A reject threshold at 50 and an accept threshold at 850 turned the score into decisions

Pipl's proof of signal intelligence

69.1x

higher fraud on disposable emails and unverifiable addresses

95.7%

the average risk rate, in the segment holding 80% of fraud

21.6%

fraud rate under the 50 threshold, compared with 0.08% overall

The real win? A clear line between scalpers and fans

With two thresholds in place, manual review no longer has to be the default for every uncertain order. Orders under 50 can be rejected on the spot and orders over 850 approved, and inside that accept band the auto-reject rate fell to 1.5%, compared with a 12.09% average.

The fraud team gets fewer borderline calls and a score whose reasoning they can inspect. Fans get a checkout that treats them like fans.

Get these results for your organization

Est. $4.53M in fan orders moved to approval

Orders scoring above 850 carried little fraud, giving the team room to approve revenue that broad rules had held back.

1.11% lift in acceptance rate

Acceptance grew by focusing automatic approvals on the score's most trusted band.

52.7% of fraud value stopped by one reject rule

The sub-50 threshold alone accounted for more than half of all fraud dollars.

Est. $101.1K cut from fraud losses

Catching throwaway-identity orders at the threshold reduced the losses they would otherwise have caused.