June 28, 2022
When talking about payment transactions, there is a common misunderstanding that a transaction is something that cannot be influenced and should remain untouched. But is that really the case?
Fraudio's CEO João Moura will give a talk at MPE '22, in Berlin, where he will go into detail about the payment industry's current state of affairs, how transactions are processed in the present, and how we can optimise that process in the near future. This all started as a vision that Fraudio presented 3 years ago, at MPE '19, and which Fraudio is now the current front-runner.
Here's a small teaser of João Moura's MPE '22 talk:
Everyone wants to optimise payment transactions in order to cover 3 aspects: improve authorization rate, reduce transaction costs, and keep risk under control.
Currently, you can only optimise for two of the three at any given time, but not all of them.
What happens if you're not optimising for low fraud ratios? What happens if you're not optimising for low costs? What happens if you're not optimising for higher authorization rates?
What happens if a transaction is optimised for high authorisation rate and low cost? An example could be mail order or telephone order transactions where the authorisation rates are super high and they are very cheap. The sacrifice here is the risk incurred, as these types of transactions are vulnerable to being exposed to fraudsters.
Similarly, when we optimise for low risk, the authorisation rate goes down as users drop out from the paying experience and the cost naturally increases as tools such as 3DS are enabled.
That's the world that exists right now.
João’s vision is that, in the future, AI can take a bigger role in the ecosystem whereby centralising data and using machine learning to learn from each transaction - this central entity can supervise and modify transactions to be optimal for a high authorisation rate, low cost and low risk at the same time without compromising either.
Traditionally, once a transaction starts its journey through the payments ecosystem - it has a certain form. It contains the MCC code of the merchant, a channel is defined as a form of payment, certain risk controls are put in place as it goes through the system, etc. From the start of the journey to the end, the transaction never changes its parameters…
You can actually look at payment transactions and if you know that they are not fraudulent, then you can route them in a specific way and in a particular format that minimises costs, and maximises the likelihood of authorization. But this is something that has to be done in a holistic way in which you cannot just have a little piece of software that tries to optimise costs, because if you're optimising for cost, then you're going to very likely have higher fraud ratios and lower authorisation .
Now, how can you optimise for all three at the same time? When you know that something is not fraudulent, because you have a good tool for it, then you can switch, and optimise other parts of the transaction. You can tweak the parameters in the transaction in real time before the transaction is processed. It is not something you can do after processing when the transaction is settled or rejected, for instance.
For example, you can adjust security features or parameters that are going to increase the cost. There's really a lot of optimisation you can do once you have a high degree of certainty that the transaction is not fraudulent. But you can only do this when you have a good centralised oracle that can understand all three facets of a transaction holistically at the same time.
Would you like to go more in-depth in this topic? Then we have the pleasure to invite you to this year's MPE '22 Berlin event. We'll be attending, sharing this talk live, and have booth presence during MPE conference. Be sure to visit us there, to see this talk in full, and schedule our talk.
See you soon!
How about trying our solution and experiencing the next generation for yourself?