A seller gets banned from a peer-to-peer marketplace for shipping counterfeit goods. Within 48 hours, they’ve registered a new account using a different email address, a prepaid debit card and a slight variation on their business name. The reviews are gone, but the operation is intact. They list the same products, resume shipping and start accumulating five-star ratings all over again.
The platform’s fraud team flagged the original account. Trust and safety implemented the ban, but no one measured what happened next.
Why Treating Ban Evasion as a Fraud Problem Makes It Worse
Most marketplace operators categorize multi-accounting as a fraud issue. On paper, that makes sense — a banned user creating a new identity to circumvent enforcement is deceptive behavior, but classifying it as fraud confines the response to fraud tooling and fraud headcount, which is where the problem starts compounding.
Multi-accounting does damage beyond fraudulent transactions. It taints the data layer that every other team depends on. Product teams use seller ratings and buyer behavior signals to make platform design decisions. Growth teams use marketplace activity metrics to forecast demand and allocate resources. When a banned actor re-enters the ecosystem under a clean identity, those signals degrade. The review system rewards a seller with no legitimate track record, search rankings reflect artificial activity and customer support absorbs complaints that trace back to an enforcement gap nobody is tracking. Research shows that 11-14% of reviews on major eCommerce platforms are potentially fake.
The cost of multi-accounting isn’t measurable as a single line item on a fraud report. It spreads across seller quality scores, buyer retention rates, operational overhead and platform reputation — none of which the fraud team’s dashboard captures.
A Metric Nobody Owns
Ask a marketplace fraud team how many banned users successfully re-register and you’ll get silence. Not evasion, an actual absence of data. Most platforms don’t track ban-evasion repeat-offense rates because no team has been tasked with measuring them.
That measurement gap is itself evidence of a structural problem. Fraud teams own account-level enforcement. Trust and safety owns policy. Product owns the registration flow. Growth owns acquisition volume. Multi-accounting sits at the intersection of all four, which, in practice, means it falls under nobody’s quarterly objectives.
The result: each team addresses the fragment of multi-accounting that falls within its scope. Fraud flags suspicious accounts—trust and safety bans policy violators. Product tightens registration fields. Growth resists friction that might reduce sign-up conversion. No one stitches the full picture together, and no one measures the outcome of the combined effort — or its absence.
The organizational design isn’t accidental. Marketplace teams are typically structured around the customer lifecycle: acquisition, activation, engagement, monetization. Multi-accounting cuts across that entire lifecycle laterally. A problem that doesn’t fit neatly into one team’s workflow tends to get partitioned rather than owned, and partitioned problems don’t get measured.
What Measurement Would Look Like
Treating multi-accounting as a platform integrity metric rather than a fraud incident changes the operational model in two ways.
First, it forces a shared definition. Fraud, trust and safety, product and growth need to agree on what constitutes a re-registration event, what severity thresholds trigger escalation and what data each team contributes to the detection layer. That alignment doesn’t exist today on most platforms because the teams have never been asked to build it.
Second, it introduces a recidivism rate — the percentage of banned users who successfully create and operate a new account within a defined window. That number, once visible, becomes actionable. If 30% of banned sellers re-register within 90 days, that’s a platform integrity failure with a measurable cost. If 5% do, the existing controls are working. Without the metric, neither scenario is distinguishable from the other.
The signals required to catch multi-accounting already exist: device fingerprinting identifies hardware and browser configurations even when names, emails and payment methods change; network analysis maps relationships between accounts sharing IP addresses or referral chains; behavioral signals, like listing cadence and pricing patterns create profiles that are difficult to fabricate from scratch. The detection capability isn’t the gap. The measurement mandate is.
The Real Cost of Not Measuring
Platforms that treat multi-accounting as a fraud subcategory will keep catching individual accounts and missing the systemic problem. Every re-registered seller who builds a clean reputation is a data integrity failure that compounds over time. Buyer trust erodes. Seller quality scores lose meaning. Enforcement becomes performative — visible action without measurable impact.
Multi-accounting deserves its own metric, its own owner and its own cross-functional mandate. Platforms that don’t build that structure will keep banning accounts and wondering why the same products keep showing up.
Frequently Asked Questions
Ban evasion occurs when a suspended or banned user creates a new account to resume activity on a platform. Common methods include new email addresses, prepaid payment instruments and device or browser changes designed to avoid matching the banned account’s fingerprint. For a full breakdown of how this plays out at the seller level, see seller onboarding fraud.
Because no single team owns the metric, fraud, trust and safety, product and growth each manage a portion of the problem, but none has a cross-functional mandate to measure re-registration rates after enforcement actions. The result is an absence of data rather than a deliberate gap.
Device fingerprinting, network analysis and behavioral signals are the primary detection layers. These identify returning users even when surface-level identifiers — name, email, payment method — have changed, because hardware configurations, IP relationships and behavioral patterns are harder to alter fully.
Undetected re-registrations degrade the data every other team relies on. Seller quality scores reflect activity from accounts with no legitimate history. Search rankings incorporate artificial signals. Growth and product decisions built on corrupted marketplace data produce outcomes that can’t be explained by surface-level analysis.
The core measure is the repeat-offense rate: the percentage of banned users who successfully open a new account within a defined window, typically 30, 60 or 90 days post-ban. That figure, tracked over time and broken down by ban reason, gives fraud, product and growth teams a shared number to act on.
