We’re excited to officially launch our ‘Beat’em’ feature today. The feature, which analyzes online shoppers’ heart rate to identify fraud, was piloted for the past year with select beta merchants with outstanding results, and is now available to all eCommerce merchants.
‘Beat’em’ utilizes the same knowledge on which polygraphs are based, to identify excited or nervous behavior. A baseline heart rate is established within the shopper’s first few clicks or swipes on the site. This baseline is then compared to the customer’s heart rate at the checkout stage – when fraudsters use stolen credit card information to complete the transaction. Thanks to ‘Beat’em’, the Riskified shopfront beacon, a snippet of code embedded in every page of a merchant’s website, can now track not only browsing behavior within the site, but also the visitor’s typing speed & heartbeat.
This feature was under development for a long time and was carefully tested before being implemented. In the first stage, after installing the special beacon with several merchants, the bio-data was analyzed and compared to the decisions reached by Riskified’s machine learning models. The underlying assumption was that in cases where the heart rate at the checkout stage is significantly higher than the customer’s baseline rate – this might indicate fraud.
Changes in online shopper heart rate at different web pages:
In the second stage, the data was incorporated to Riskified’s models in “shadow” mode – meaning the decision that would have been reached with the bio-data versus the decision reached without it was compared. In some cases, legitimate customers exhibited increase heart rate at the checkout stage, possibly due to positive excitement over the purchase, which could lead to a false positive decline if the decision was based on bio-data alone. However, when the heart rate data was weighted properly in relation to all other available data points, Riskified’s models achieved an extremely high level of decision accuracy.
Once the contribution of the heart rate data was statistically validated, Riskified began incorporating it into the pilot merchants’ risk models, with impressive results. ‘Beat’em’ allowed Riskified not only to avoid approving fraudulent transactions, but also to approve orders that merchants would typically consider risky. We’re excited to offer this advanced technology to all Riskified merchants at no additional cost!
Still under trial is ‘Beat’em’ for mobile commerce. Thanks to touchscreens, ‘Beat’em’ for mobile can measure not only heart rate, but also skin conductivity, another metric commonly utilized in lie detectors. When a user is nervous or anxious, this can increase the levels of sweat in their eccrine sweat glands (on the palms of the hands and feet). Sweaty hands are more conductive than dry skin, which causes an increase in bulk conductivity. As with the heart rate, a baseline skin conductivity measure is established before the user reaches the checkout and payment pages.
‘Beat’em’ marks a revolutionary advancement in online fraud prevention and provides a strong advantage to online retailers in fighting fraudsters.
Interested to learn more? Contact us!