Notifications
Clear all
Topic starter
11/07/2026 1:58 pm
TL;DR: Rules-based fraud tools can overload merchants with friction, while AI and machine learning can reduce false declines, speed decisions and better distinguish legitimate customers from fraudsters, according to Signifyd. The practical shift is toward risk decisions that preserve trust and conversion instead of treating security and customer experience as opposing goals.
NHIMG editorial — based on content published by Signifyd: CX Fraud Prevention: Balancing Risk and Customer Experience
By the numbers:
- 76% of consumers worry about data security when shopping online.
- seven out of 10 customers abandon their carts without placing an order.
Questions worth separating out
Q: How should security teams reduce false declines without weakening fraud controls?
A: Security teams should use layered risk scoring, then apply step-up authentication only when the event or behaviour is unusual.
Q: Why do rules-based fraud tools fail when transaction volume grows?
A: Rules-based tools fail because they rely on static thresholds that cannot keep pace with changing fraud tactics and customer behaviour.
Q: What signals should fraud teams use beyond basic login checks?
A: Fraud teams should use device, session, payment and behavioural signals together, because no single indicator is reliable on its own.
Practitioner guidance
- Separate low-risk and high-risk decision paths Use different approval logic for routine logins, checkout, refunds and payout events so the same control does not over-block low-risk customers.
- Replace brittle knowledge-based checks Retire knowledge-based authentication where fraudsters can guess or source the answers and use stronger factors only when the risk score justifies the added friction.
- Tune ML models against false-decline metrics Track approval accuracy, false declines, manual-review overturn rates and chargeback outcomes together so the fraud model is judged on both risk reduction and customer impact.
What's in the full article
Signifyd's full article covers the operational detail this post intentionally leaves for the source:
- A fuller breakdown of rules-based fraud tooling and why ruleset bloat leads to friction at checkout.
- Specific examples of how multi-factor and biometric-style checks are applied in customer journeys.
- The article's detailed view of how machine learning uses transaction and behavioural signals to reduce false declines.
- Further explanation of how encryption and secure payment gateways fit into ecommerce fraud prevention.
👉 Read Signifyd's article on balancing CX fraud prevention and checkout conversion →
CX fraud prevention and false declines: what teams need to know?
Explore further
11/07/2026 2:07 pm
CX fraud prevention is now an identity governance problem, not just a fraud problem. The article is right to frame customer experience and protection as linked, because every approval, decline and step-up decision is also a trust decision. For identity teams, that means fraud scoring, account assurance and access decisions need shared governance rather than separate optimisation targets. The practitioner conclusion is that customer identity assurance must be tuned as part of the broader identity programme.
A question worth separating out:
Q: Who is accountable when fraud controls create too much friction?
A: Accountability usually sits across fraud operations, IAM, product and customer experience leadership because friction is a governance outcome, not just a tuning issue. If controls are causing avoidable abandonment, the organisation needs ownership for the decision logic, the supporting data and the customer impact. That is why fraud governance must be shared.
👉 Read our full editorial: CX fraud prevention shows why security and conversion are linked
