Michael Benisti is the co-author of a book in French on fraud management (Lutter efficacement contre la fraude: Enjeux stratégiques et opérationnels, available on Amazon) and teaches e-commerce payment and fraud in business schools in Paris. He is the current head of payments and revenue protection at Ledger.
We sat down with Michael to get his thoughts on the entrance of payment providers into the chargeback guarantee business.
1. We’re seeing a lot of payments providers — PayPal and Stripe to name just two — get into the chargeback guarantee business. As a long-time fraud manager, is this a game changer for you? Do you think this is a trend that the traditional fraud prevention industry needs to be worried about?
It is clearly an interesting trend that could have a significant impact on the industry.
However, this is not exactly new. Paypal has been offering their merchant protection program since 2012 to most merchants. As a reminder, Paypal have their own fraud algorithm and will bear the risk in case of fraud chargeback as long as the merchant is able to bring a shipping proof and ships the good to the address communicated to Paypal at the time of the order.
Stripe just got into the chargeback guarantee business with their latest product “Stripe Chargeback Protection” costing 0,4% of transaction value insured.
When you think about it, the amount of data you have will be a determining factor in your ability to fight fraud. With such an important network of merchants, Stripe states that there is an 89% chance that new card payment for a merchant will actually already been used in the Stripe network before. That will definitely help Stripe in their risk assessment.
It’s also one less thing to deal with, especially for smaller merchants (Stripe historical target) that do not have the resources to learn and implement a fraud management strategy. At 0,4% fees, the “Stripe Chargeback Protection” product might be worth considering especially when dedicated fraud providers often have minimum monthly fees that can be significant ($1000-$3000) and in some cases even setup fees.
PSPs main weakness in terms of fraud management is for a larger merchant using a multi-PSP approach with smart-routing, as in that case it might be easier to have one single tool for fraud rather than having one per PSP.
2. Would you say chargeback guarantee models are going to increase their market share at the expense of more traditional fraud solution providers?
Chargeback guarantee is clearly not a one size fits all model. And I speak from experience.
In my previous position (Vestiaire Collective, a second-hand luxury goods marketplace), I was a very strong supporter of Riskified, a chargeback guarantee machine-learning based fraud engine. It was the perfect fit at the time and I still absolutely love the product and technology today.
However, when I joined Ledger I made the call not to go with Riskified again. The B2C products we sell on ledger.com – Ledger Nano S and Nano X – are called hardware wallets. They allow you to secure your cryptoassets’ private key on a physical device. Due to their nature, those products are much more difficult for a fraudster to resell. We also strongly recommend our community only to buy from us directly or from our list of authorized resellers. Because of this, the inherent e-commerce payment fraud risk for Ledger is lower than what it was in my previous experience. On top of that, while my previous company was a marketplace, Ledger manufactures its own products and is able to have higher margins. The combination of those effects (lower inherent risk, higher margins) makes me less risk averse as a fraud manager.
The main issue with chargeback guarantee is the risk appetite of the fraud provider. For high value orders in particular, you could often end up accepting initially rejected orders by fraud provider. The maths are pretty straight forward. Let’s consider the chargeback-guarantee fraud provider takes a cut of 0.5% while you as a business have a gross margin of say 25%. If the fraud provider validates a fraudulent order and therefore has to refund for the chargeback, it will take the provider 200 orders to recoup the loss and breakeven. As the merchant, making the same mistake would not impact you as much : only 4 orders would allow you to breakeven. Therefore, it is expected for chargeback-guarantee fraud provider to be more risk-averse than the merchant if both players behave in a strictly economical way.
I was told by one executive of a major chargeback guarantee fraud provider that this is even included in their fraud models. In other words, negotiating your fees really hard will not necessarily have a net positive impact for your business as the provider will tend to decline more.
3. What advice can you give to traditional vendors trying to differentiate the value they bring merchants from the solutions provided by payment providers?
Well I am not a sales guy but here is a short list of competitive advantages for standalone fraud solution vendors compared to payment providers fraud module:
- Bring more value by tackling more topics than payment fraud. Sift for example is promoting their content integrity product while Riskified is also positioning itself on account takeover fraud which can easily become a huge topic in a marketplace with digital wallets for instance.
- Explain that they are a much more suited solution for multi-PSP merchants as you would only need to configure one single fraud solution
- If your are focusing on fraud and on fraud only, you should hopefully be more advanced in terms of technology. Demonstrating technological superiority is the tricky part of course.
4. On the tech side, where should traditional vendors be looking to innovate? What new features and services would you like to see them create to innovate where a payment service provider either can’t, or won’t?
To be honest, I am not sure at this stage that the large payment service provider will put barriers on what they can do. Right now, I would say that in terms of fraud solutions they are not as good as the latest generation of standalone fraud solutions based on machine learning but as they see more transaction volume the benefit from more data to crosslink orders & buyers with.
5. Are you seeing any adaptation by fraudsters in methodology or tactics as they look to avoid detection by payment providers’ solutions?
They always do, that’s for sure. What’s great for payment providers is that the amount of data they have can really allow them to judge if the order data points makes sense or not. They are not exactly there yet but I am sure we will get there eventually.
6. Your experience spans from high-end luxury fashion to cryptocurrencies. Do you think there are certain types of industries that should either avoid, or trend towards, using a payment provider solution?
Not really. What I would look at is how much data a given fraud solution has in your vertical.
For example, Riskified has a lot of players in the fashion industry while Sift has a few in the crypto-industry and that definitely helps crosslinking although for privacy and legal reasons we don’t have any access to it, it does affect the fraud engine scoring.
