Originally published on PYMNTS
Merchants are facing a sizable number of card declines, which are stubbornly averaging at around 15 percent. That challenge is not lost on FinTech startups. “Merchants are now getting swamped with declines,” Modo Chief Executive Bruce Parker told PYMNTS in an interview.
Yet around 12 or 13 percent of them fall into a general bucket of declines provided by issuers. As a result, merchants can’t always discern why exactly a decline occurred based on the transaction codes they receive. At the same time, merchants are learning a decline is not the end-all-be-all, and the merchant is left on the hook to figure out if it should reprocess a transaction.
Why? The merchant is concerned about how reprocessing a transaction might impact their own relationship to the card network and to their processing partner. That’s where FinTech startups like Modo come in: The company knows enough about the different reasons that declines happen, and so is able to help merchants find out “when is a decline really a decline and when is a decline an opportunity to ask in another way,” according to Parker.
As it stands, the fraud systems at major card issuers have been neural networks or old-school artificial intelligence (AI) by tradition, often referred to as black boxes. One cannot predict what these systems are going to do given a set of inputs, because each neural network has been training for the better part of 30 years to find problems using a completely different set of historical transactions.
These systems have, in essence, escaped their masters, and declines are occurring for reasons that aren’t always completely understood. But one can learn which paths work better based on factors like time of day or particular banks through particular processors. Modo sees this challenge as a form of interoperability, but not in the sense of taking something other than a card and interacting with it like a card.
In this case, Modo sees interoperability as how one takes a card and runs it down one of many different channels based on the likelihood that it will work. To make this happen, merchants need the ability to talk to many different processors. The trick is making multiple processors look like one entity, so merchants can maintain one connection and have one set of data, while improving the odds of getting an authorization.
On another note, Parker plans to speak at Money20/20 in Las Vegas next week to further efforts to help merchants solve for interoperability challenges.
