Case Study

Higher precision, fewer false flags

This global social media platform had a CAPTCHA problem, but the issue wasn’t just about bots. Good users were getting flagged. Risky ones were slipping through. And their model’s precision was starting to plateau.

By strengthening identity signals, aligning scoring thresholds to business objectives, and validating user trust before registration, the platform improved performance at every level while reducing friction for the users they actually wanted to keep.

The challenge

With millions of new registrations per month, this platform relied on a sophisticated lifecycle model to score risk, but the system wasn’t calibrated to distinguish trust.

CAPTCHA was being over-triggered for good users. Some risky accounts weren’t being challenged at all. And signal strength between identity attributes wasn’t informing decision logic. They needed a way to shift from reactive risk scoring to proactive trust validation, without re-architecting their flow or raising false positive rates.

Their goals were clear

The hard part wasn't naming them, it was hitting all four at once without one undermining the others:

Reduce unnecessary CAPTCHA friction for good users

Improve fraud recall at high-precision thresholds

Strengthen signal resolution across email, phone, and IP

Unlock confident approvals without increasing risk

Where static systems broke down

A risk-only model can spot what looks suspicious, but it can't recognize who deserves to be trusted. By the time the team brought in Pipl, three problems were stacking up on both sides of the funnel.

Good users were flagged too often

Even legitimate signups were challenged by CAPTCHA due to low scores, not low trust.

Some fraud went unchallenged

40% of known-risk registrations were missed by the native model but flagged by Pipl Trust.

Signals lacked connectivity

Phone, email, and IP attributes were scored in isolation and not crossed to validate trust.

How the system changed with Pipl

Pipl introduced Trust Scores calibrated to the platform’s own registration outcomes, using both labeled and unlabeled events. Three scoring strategies were tested: Pipl Trust standalone, a blended average of both scores, and a joint model with the platform’s score used as an input feature.

Historical CAPTCHA, fraud, and approval decisions were used to calibrate thresholds

Identity signals were evaluated for strength and cross-signal consistency

Connectivity scores validated relationships such as phone to IP or email to name

Thresholds were aligned to business tolerances for precision, friction, and fraud

Pipl's proof of signal intelligence

3x

less fraud on matched email + phone

59%

less fraud on users in Pipl's identity graph

42%

less fraud on matched phone + address

80%

less fraud with 3+ strong identity signals

The real win? Trust at the point of entry

Once deployed, Pipl’s Trust Score sharpened decision-making across the funnel. CAPTCHA was reduced for good users. Risky signups were caught earlier. And model lift was immediate, with blended and joint models outperforming baselines across every metric.

Get these results for your organization

Pipl flagged 40% of risky registrations missed by the native model

Trust signals identified high-risk users that existing logic scored as safe, strengthening the top of the funnel.

Recall at 100% precision jumped to 10.94

Trust calibration enabled stronger detection without increasing false positives, nearly doubling recall in the safest zone.

CAPTCHA accuracy improved 4.2 points at 70% precision

Median scores for fraud vs. legitimate users separated cleanly, enabling smarter rule design.