Feeling powerless when talking to banks is common for a payment manager. Listening to my peers, I come across this topic very frequently. Payment managers find themselves asking how to contact an issuing bank or discussing their frustration at not getting an actionable reply from their acquiring bank. They don’t have visibility into the payment chain – their payments are failing, and they do not know why.
However, the payment chain doesn’t need to be a total ‘black box’, it can be reverse-engineered. You don’t need cutting edge technology, but you do need a certain setup, a multi-acquirer setup. The principle is simple: with multiple routes for transactions, you can test one route against another and infer why they perform differently. Ideally, you would replicate every link in the payment chain and could mix and match all elements to create tests with limited variables. In reality, this is not always practical or cost-effective. For example, according to my experience, different gateway services connecting to the same acquiring bank have little effect on performance – however, multiple acquiring banks have performed very differently.
Running tests
A great way to benchmark your different payment routes is to split your volume equally between two or more different acquiring banks – often (but not always) a country can perform differently with different acquiring banks, and sometimes the cause can be traced back to a bug or an erroneous setting in the acquiring bank. Deep diving into the data identifies one segment of the transactions that is the cause of the low performance. In the past, I have identified issues the acquiring bank didn’t know about, such as card sub-types not being activated on a MID (Merchant ID), price limits wrongly applied to a local currency instead of the settlement currency, or unusual foreign card and country combinations being blocked.
Sometimes an acquirer will just perform better in a market – we don’t find any root cause, but we see the same trend across more than one merchant, and the acquiring bank performs consistently better than several competitors. Even without understanding the mechanics, this is a win for us, and we reward this acquiring bank with more volume.
A great way to look at the merchant’s reputation with an issuing bank is to start new traffic on a MID. From my experience, new traffic is given a general risk rating from the issuing bank. Over time, the acceptance rates will either go up or down. I like to imagine machine learning models taking on critical volumes of fraud and decline data – but when authorisation rates for a bank changed on a schedule every month, the reality may be a little less technical.
If the authorisation rate goes down, maybe your MID has a reputation issue. Different setups can be tested to understand what the issuing bank is concerned about. Removing failovers from one MID can show if high declines are an issue. Splitting traffic by fraud risk can give an overall uplift by the better performing good traffic and indicate you have a fraud or chargeback sensitive issuer. This can be useful information especially if you are running a dynamic 3DS setup.
Failovers
For this testing strategy to work, you need to recover the lost transactions to limit their impact on revenue. Testing between acquirers can mean you are running large volumes through a route known to be sub-optimal to achieve statistical significance. This is only possible if you have a robust failover strategy. Successful failover transactions give you two benefits: identification of payments that should have worked and a mechanism for limiting the cost of testing. It’s a hard conversation to have – explaining sales have been lost because you wanted to see how they failed. It’s a much easier conversation to explain why you have some increased processing costs from retrying transaction. Being able to safely fail gives us the freedom to test new setups and to ‘fail fast and learn faster’.
Communication
You are now communicating with the banks from a position of control; you have benchmarked your acquiring bank against its competitors and you can demonstrate transactions are failing when they shouldn’t. Corresponding with your acquiring bank is now different; you are no longer requesting support but informing them of their performance. You have multiple acquiring banks – so, the volume is easily rerouted until they have a solution in place. The internal pressure from reduced revenue is no longer there, the whole process is stress-free; you are back in control.
With each acquiring bank now getting a smaller slice of your business, the downside of this strategy is negotiating for prices or support resources. Some will be less inclined to go the extra mile for their client.
A multi-acquirer setup gives the payment manager good insights into what issues are affecting their authorisation rates, hypotheses can be tested, and data can be produced to back up the findings. Fixes can be implemented quickly and without relying on resources from the banks.
This editorial was first published in our Cross-Border Payments and Ecommerce Report 2020–2021, which assesses the change of pace that occurred in 2020 and provides a comprehensive overview of the major trends driving growth in this space, being the ultimate source of information for players interested in selling across borders.
About Liam Castagna
Liam Castagna is Head of Payment at Insparx GmbH. He is responsible for their Payment and Fraud vision and works tirelessly to optimise the payment flow. With 22 integrated Acquiring Banks benchmarked in the past year, he believes leveraging your own data is the most effective way to build business performance.
About Insparx
Based in Munich, Germany, Insparx GmbH is a leading global online dating merchant in over 40 different countries. With a fully subscription-based business model, they are experts in customer retention and managing an extended payment lifecycle.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now