What Are False Positives?

In anti-fraud circles, false positives are legitimate user actions that are blocked on suspicions of criminal activity. This could be a declined transaction, an aborted onboarding step, or a failed authentication.

Some companies measure false positives as a “customer insult rate” because they frustrate good customers who cannot proceed with their intended actions, and can even make them turn to competitors.

False positives are closely related to false declines, although the latter are more specific, as we’ll see below.

What Is the False Positive Rate?

This mathematical term describes the ratio of false positives as a proportion of all positives. For example, in fraud prevention, if five out of every 100 positives are false positives, this means a 5% false positives rate.

Considering their negative impact on the experience of legitimate customers, companies have very good reason to try to keep the rate of false positives as low as possible – ideally without compromising their defenses.

What Causes False Positives?

If your fraud prevention system is automated, a false positive will be caused by badly calibrated fraud scores or an overprotective risk strategy. 

Examples include:

• a purchase declined as fraudulent while it is actually from the legitimate cardholder (false decline)
• a user locked out of their account because the system thought they logged in from somewhere unusual
• a buyer blocked from the automated refund request because the way they attempted to go about it was flagged as suspicious

False positives also occur due to human error, for instance during a manual review process. It’s possible that the person looking at the data classified a certain action as fraudulent when it was in fact a high-risk but legitimate request.

Why Are False Positives a Problem?

The simple answer is because they frustrate legitimate shoppers, thus costing you in revenue and customer loyalty. 

False positives point to a poorly-adjusted risk management strategy. While some fraud vendors will err on the side of caution, a high rate of false positives can have negative consequences for your business. 

You may lose business as potential customers become frustrated with your company. In fact, 1 in 3 of US shoppers who experienced a false decline say they will not return to the same merchant – rising to 2 in 5 in Europe, per Forbes.

To make matters worse, false positives can create a negative cycle if they aren’t detected immediately. If you have a machine learning fraud prevention engine and the false positive is not labeled as such, for instance, it will not take these wrong results into account and may create even more false positives in the long run.

Last but not least, trying to solve false positives can put a strain on your resources, especially if the customer service team is dealing with angry customers.

Are False Positives the Same as False Declines?

A false decline is a type of false positive, specifically having to do with debit and credit card payments. The system looks at a legitimate attempt at a card payment but thinks the payment is fraudulent and declines it.

Shoppers dislike both false positives and false declines, which is what the term “customer insult rate” alludes to. It may sound harsh but shoppers can indeed feel offended if their payment is rejected. Moreover, in practical terms, they will need to invest extra time to either try again or locate identical products at your competitors and purchase them there. 

5 Ways to Avoid False Positives

Avoiding false positives takes time and effort, but it is worth it for businesses. 

1. Check the fraud prevention system: Have a look at all the risk rules in place and consider if they might be too broad or strict.
2. Test risk rules in a sandbox setting: Test risk rules in a sandbox environment and see if the false positive rates are too high. You may need to adjust your risk rules to prevent them from happening. 
3. Review your manual checks: Ideally, you will only review actions with a medium risk, as the low-risk ones should be automatically approved, and high-risk actions should be blocked. But it’s worth testing manual reviews with data on high and low-risk actions to see if you would also flag them in the same way.
4. Leverage machine learning & fraud analytics: If you manage to flag false positives manually, your fraud prevention with machine learning could suggest adequate risk rules that improve your accuracy rate.
5. Consider switching from chargeback guarantee tools: Chargeback guarantee tools are often calibrated to be overly protective. This is because the vendor has an incentive to avoid paying chargebacks for you. However, this is not necessarily aligned with your individual needs.

Overall, avoiding false positives is part of a concerted effort to efficiently and accurately catch fraudsters while reducing friction for legitimate customers.

Sources

Forbes: Three Digital Commerce Growth Opportunities