Configuring Your Risk Settings
When you tackle fraud, there’s an important trade-off: you can add more fraud prevention measures, but this might put off legitimate customers from signing up to your service. It’s important to consider how extra steps like bank account verification could impact your conversion rates. GoCardless uses its insights on payer risk to help you evaluate the trade-off between fraud prevention and conversion.
GoCardless assesses the riskiness of payers with a set of custom rules and our machine learning model. For our custom rules, we use information such as:
-
Is the bank account closed
-
Is the email linked to high failures, chargebacks or mandate cancellations
The more complex rules are captured by our machine learning model, which has 100+ different features. That model is trained based on a variety of inputs:
-
GoCardless' internal data from our more than 70k merchants
-
Data provided by our customers
-
Data from third party providers
An important part of the value GoCardless provides is aggregating data across our customer base - if we see a fraudster with another customer, we include that in the payer’s risk score and can help our other customers avoid that payer. And because we have the largest bank pay network in the world, these insights are super powerful!
Every business has a different risk appetite, so GoCardless allows you to toggle how many payers should go through extra verification steps (Verified Mandates) as they sign up to your service. If you choose less fraud prevention, you’ll probably have higher conversion rates, but you might also let in some fraudsters. If you choose more fraud prevention, the extra verification might decrease your conversion rates. GoCardless offers a helpful solution: use our recommended setting, and we’ll always choose the optimal balance between fraud prevention and conversion.
We try to estimate the impact this all makes on your bottom line on the right-hand side of the page. It’s impossible to know how much fraud you prevented if the fraudster was stopped from signing up in the first place, so this number is just our best estimate.
The % of payers asked to set up a verified mandate is a rough estimate based on the historical risk profile of your payers. If in the future the risk profile of your payers stayed the same as it was previously, then the estimated percentage of payers sent to Verified Mandates will be accurate. However, fraud is not always consistent over time, so this estimated percentage might differ from the actual percentage of payers we send to Verified Mandates, if there are more or less fraudulent payers than there have been previously. This is not a random sample; we prioritise payers who we suspect are more likely to be fraudulent towards your business to go through Verified Mandates.