Calculating the cost of chargebacks might not sound especially complicated on the surface. Dig a little deeper though and there are several approaches which produce wildly different outcomes. The result of that is that you can have a different understanding of what fraud is costing you, and consequently take very different stances on the right risk thresholds for your business.

Choose the wrong thresholds and you could wind up costing your company thousands or even millions of dollars. Here we run through three of the more common methods of counting the cost of chargebacks, and how they match up with the actual financial impact of a fraudulent transaction.

The order and goods method

Let’s get this one out the way first. You may come across fraud vendors scaremongering with talk of chargebacks costing you the order value, the cost of replacing goods, and the fee. Everything but the kitchen sink. But this is overstating the case somewhat.

In the normal course of a sale you wouldn’t gain the payment and keep the goods too. You’d gain the payment in exchange for the goods. For a given genuine order, you always expect to “lose” the goods and their associated value. So to say that a chargeback has cost you both the lost payment and the goods starts to look on paper like double counting.

At a stretch, there is a cannibalisation angle which justifies counting both the order and goods costs. It’s conceivable that this fraudulent order has taken the place of a genuine order that would otherwise have happened. For example, the fraudster resells the goods to one of your regular customers off platform. Or the fraudulent order was from a normally genuine customer, committing friendly fraud on this occasion.

In most cases, it’s unlikely that cannibalisation applies to all chargebacks. So, if you’re not going to count both the order value and the goods value, which one do you focus on?

The lost order method

When I speak to fraud teams they often use the lost order method to calculate the costs. This focuses on the value of the lost order. The point of focus is that transient period in which the payment was banked before it was reversed. This could be 10, 30 or even 90 days, depending on your business. Essentially, this approach compares where you are now with where you were for a while after the order looked successful.

This approach reflects some loss aversion bias. The order was in your grasp, and now it’s been taken away. At one point you were up, now you’re down. Let’s face it, no one likes having their sweets taken away from them! Still, this method is better than double counting.

The lost order approach has the benefit of simplicity and lines up with general practice across departments. By focusing on topline order revenue, it can help align your fraud team’s numbers with those of your colleagues in other teams.

What’s more, leaning into loss aversion can be beneficial in making the case for investment. This is partly because it produces higher numbers than some other approaches, such as the net position approach (below).

Overall, this method looks at perceived losses more than actual losses. But in doing so, there is one disadvantage. You find yourself ignoring the margin on the order – the cost of the goods sold.

The net position method

This brings us to the net financial position method. This compares where you are now relative to where you were before the order even happened. You set to one side the payment that got reversed. Instead, you ask yourself – what is my net financial position now, compared to before the order happened?

This approach highlights your margin.The focus isn’t the lost payment, but the cost of replacing the lost goods which aren’t going to be returned by the fraudster. Or, in the case of a marketplace, the lost payout to suppliers who fulfilled the order.

You can think of this in terms of the lifecycle of that order. And how the balances and the net position change over time compared to the starting point:

This is a more sophisticated approach to counting chargebacks. That said, it can be more difficult to explain and gets complicated if you have large variations in margin across your product line. One great strength of this method, however, is that unlike the others it gives you an accurate indication of how many successful orders you need to cover the net loss, on a gross profit basis.

Summing up

Depending on how you choose to calculate the costs of chargebacks, you could end up with divergent results. Following the single $100 order used in each of the examples above, and a 50% margin, we end up with this range:

So which should you choose? Well, it’s important to think about how you’re making this calculation today and how well it lines up with your business objectives.

In the tradeoff between blocking genuine customers and accepting a given volume of chargebacks, a business focused on profitability will prioritize margin. So here net position calculation is likely the most appropriate method for you.

If the margin is low, the number of good orders needed to recover the costs is high. So you might choose a conservative risk level to minimize the chargebacks. If the margin is high the reverse is true. So you might want to optimize for minimizing the block rate.

Equally, a business focused on growth and top line order volume might be more concerned with minimizing blocked orders regardless of margin. In that case, the lost order method might be the more suitable choice. There’s also a case to be made for using two different calculations. One for internal audiences - to speak their language - and one for optimizing the risk thresholds.

Whatever direction you choose to go in, you should work with your fraud vendor to figure out what the right strategy might be. Once the decision is made you can analyze your risk settings in a new light, using tools such as Ravelin’s threshold analysis tool (below).

This provides insight and helps you set the appropriate levels to match your strategy.

Learn more about how we can help your business reduce fraud and chargebacks.