Online Fraud Prevention to Keep Companies Safe, Profitable

Q: The pandemic led to an e-commerce boom. As an online fraud prevention and revenue optimization company, how has Bayonet performed?

A: Any company that sells products or services online can use Bayonet’s service to understand who are their trusted shoppers and make sure their payments go through frictionlessly. Improving payment acceptance rates optimizes profits and the user experience, but in Latin America, you have to be careful about fraud (chargebacks) since fraudulent payment attempts have increased fourfold since the pandemic started. 

Before the pandemic hit, we already envisioned this market would grow. In 2019, e-commerce in Mexico grew 35 percent, more than in any other country, and recent dynamics caused this segment to accelerate exponentially. From one quarter to the next, transaction volume through Bayonet increased threefold, but we also saw fraud attempts increase fourfold. Fortunately, all of our customers improved their payment success rates by 72 percent and decreased fraudulent payments by 82 percent on average. We work with companies that have innovative business models, such as Boletia, Neerme, Moneypool, Pagaloop, Dapp, UnDosTres, or even big ones like Grupo Modelo and MIT, the latter being one of the largest payment processors in Mexico. This year, we are looking to strengthen our position in fintech and retail, while also expanding into Latin America with the recent investment closed in 2020.

Q: How does Bayonet prevent digital payment fraud and optimizes payment success rates?

A: What Bayonet does is replicate human intelligence in real-time. It is like having an army of people checking every transaction in detail, things like the type of card used, the product purchased, the sale’s amount, what IP address is coming from and what type of device is the shopper using. Understanding these variables lets us know how legitimate a shopper is.

It goes like this: as the online shopper clicks “pay”, our application programming interface (API) receives the parameters of the transaction. Then we do three things. First, we build the user's profile to see, for example, how many different email addresses, phone numbers, and credit cards they have to detect if he has a fraudulent profile or not. Second, we generate specific decision models for each of our clients to understand what is normal and abnormal for each business model. Third, we connect to payment vendors on behalf of our customers and route the payments via the ones with the highest chances of approval.  All this happens in milliseconds, the shopper is not even aware of it.

Q: How does Bayonet keep pace with innovative attacks? 

A: We build our decision models with the Latin American mindset to ensure our service is flexible and fast. For example, speed of analysis is important to stop sophisticated Bot attacks, and flexibility is important to address any type of business model. 

Nonetheless, when a company opens up to an online sales channel, it will naturally be subject to fraud. There are services that ensure the cost of fraud, but we prefer not to because this naturally increases the friction with good shoppers. Our vision is to never structure a review process with just the bad guys in mind. Ninety-five percent of the users on average are genuine and the rest are fraudsters using multiple identities. It will always be more profitable to make sure good shoppers can pay, than to build a high-friction flow to have no fraud. That’s why although we’re a fraud prevention company, our focus is to increase profitability, and that means having a small percentage of fraudulent payments each month but favoring an increase in payment success rate. For example, is better to take the payment approval rate from E.g. 58 to 67 percent and having 20 percent more fraud by 0.5 percent than to avoid fraud at all costs, which will greatly affect the shopping experience.

Q: What sets Bayonet apart from its competitors?

A: The fact that we solve fraud and increase payment success rate in a single service. Most fraud prevention solutions are only focused on one side of the equation: losing less. But we’re also focused on selling more, that’s why on top of fraud prevention we help clients connect to multiple payment processors, tokenize their customer cards and route payments intelligently to the best-performing vendors, we can even retry failed transactions. Additionally, most solutions in the market are difficult to install. It takes, for instance, three to six months to install a US fraud-prevention service. In contrast, we can install our solution in a matter of days. Moreover, competitors’ costs are designed for US and European markets and this is not an option for SMEs or startups that are just scaling up. A long-time customer of ours told us that the quote he got from one of our competitors represented 50 percent of his total profit margin. Other solutions need full-time employees to manage their tool. At Bayonet, our philosophy is to allow companies to focus on their core business without having to become experts on payments and fraud.

Q: How does Bayonet protect the data it collects from users?

A: In the payment industry there is a very well-known certification called Payment Card Industry Data Security Standards (PCI DSS), which was created to increase controls around cardholder data to reduce fraud. Bayonet has this certification and we are subject to pen tests to check for vulnerable data. In addition, every piece of data we receive is masked (hashed). This means that if someone did access our database, they would not be able to see the actual data.

Q: What opportunities does Bayonet see in the market in 2021?

A: We see a very strong trend, which is fragmentation in the payments market. There are more and more new payment methods in the market, such as Apple Pay, which has just arrived in Mexico, cryptocurrencies or cash payments. Companies are looking to enable as many payment methods as possible and, according to CONDUSEF, of all online transactions in Mexico, only around 60 percent are successful. We have seen that between 3 and 20 percent are declined for no clear reason and could potentially be saved.

Bayonet is an online fraud prevention platform powered by machine learning. The company, founded in 2016, analyzes customer data to create patterns and identify fraudulent users