At Pagos, we’re working on intelligent systems that monitor and optimize the performance of merchants’ payments processing technology. Since there are so many steps involved in successfully processing a payment, there are many opportunities for a transaction to fail or be declined along the way. Sometimes, transactions are declined for reasons the merchant can address and the payment attempts can be retried. Other times, transactions are declined for good reasons, like attempted fraud. In general, we can classify declines into two categories: hard or soft. A hard decline indicates the transaction will not be processed successfully as is, while a soft decline suggests the transaction might succeed again in the future as is, if it is retried. Pagos gives merchants visibility into the distribution of their soft and hard decline rates.
When it comes time to diagnose a problem in your payments stack, the best approach is to narrow the transactions you’re investigating to a specific currency, market, processor, or network—at Pagos we call this “navigating the metric tree.” Payments are very complex, so we use the metric tree to limit the scope of our work to a specific context.
For this blog post, we’re going to look at our platform data from the opposite perspective: with no specific customer or metric-tree context. This may seem counterintuitive, because the value proposition of Pagos’ “awareness” products (like Peacock, our data visualization product) is all about increasing specificity and context in a merchant’s view of their payments data. Why would we be interested in looking at payments data without drilling down?
The reason is that an aggregated view of payments data across customers enables Pagos to make inferences about market-level trends in merchants’ payments performance. Once merchants who ask “How do my payments perform?” have the ability to slice-and-dice their data with Peacock, their next questions are “Is this normal?” and “How do I compare to other merchants like me?” The ability to answer these kinds of questions is the foundation from which Pagos can give advice and recommend specific actions our customers can take to improve their payments performance.
So our goal in this post is to find some high-level, normative statements to make about the patterns and costs of merchant payments. A great place to start is to look into the patterns of declined transactions.
An NSF decline is a transaction the issuer declined because the customer either didn’t have the funds available to cover the transaction or is above the credit limit on their credit card.
NSF declines are considered soft declines; even though they represent lost revenue, these kinds of declined payment attempts may be recoverable for a merchant, but at a cost.
For a given set of transactions, the average order value (AOV) equals the total value of those transactions divided by the number of transactions. AOV gives an idea of the central tendency of those transactions. In the context of approved transactions, you could think of the AOV as a weighted approval rate measure; approval rate tells us the proportion of successful payment attempts and AOV gives a rough estimate of the revenue generated by each of those transactions.
AOV isn’t just useful for understanding the revenue impact of approved transactions. If we calculate the AOV of NSF declines, we have a measure of the amount of revenue per transaction lost specifically when a cardholder doesn’t have the funds available to complete their payment.
Recall our goal: what can we say, irrespective of specific customers or metric tree segments, about the average transaction declined due to NSF? How much was that transaction worth a year ago? How much is it worth today? How often do these declines happen?
Based on a stratified random sampling of anonymized customer transactions (denominated in USD, from issuers in the United States) over the past year, we know the following:
Inflation in the US from November 2021 to November 2022 was roughly 7%. As such, we’d expect to see an increase in AOV over this period, because consumers are paying higher prices for the same goods and services. However the value of the average transaction declined due to NSF increased 15% over that same period, suggesting the increase isn’t explained by inflation alone.
This data suggests that an increasing number of cardholders have less available funds or credit, and that is not good news for merchants. For merchants who transact with customers on a non-recurring basis, sales lost to declined payments today are more frequent and more valuable than one year ago. For subscription merchants, it portends future growth in churn.
Given the frequency and value of NSF declines is up across the board, these kinds of failed payments represent a larger share of merchants’ uncaptured revenue, relative to one year ago. Merchants who have implemented a strategy to retry transactions declined due to NSF will be able to recover some of that revenue. Attempting to recover revenue with a retry strategy comes with a tradeoff though: merchants still have to pay the usual fee structures assessed by their gateway/PSP/acquirer to process a transaction. Ultimately, this amounts to a more nuanced tradeoff between the revenue ‘recovered by retry’ vs. the cost of attempting the recovery. This is a particularly difficult situation for merchants who don’t have tools for insight into their declined transaction distributions.
Speaking of merchants who don’t have visibility into their declined transactions… is that you? If so, chances are you’re leaving a lot of money on the table in the form of declined transactions that could be completed successfully, if only you knew which transactions and which steps to take to remediation. Fortunately for you, we’re making just the tools you need here at Pagos. To learn more about everything Peacock can tell you about your decline data, sign up today!