Much like the other areas of finance, the payments industry can benefit tremendously from adopting the latest techniques in data storage and analysis. To get an insight into practical use cases in the Payments & Fintech industry we interviewed Daniel Linder, Data Scientist at Adyen this month.
PCM: Why is Data Science needed?
Daniel: There are a few reasons why data science is important for any organization. First of all, the amount of data that most organizations are dealing with and could leverage for their organizations is not scalable to analyze in a manual way. Data scientists have the ability to free up valuable resources by automating the known solutions and working on the unknown. For example, as the data increased at Adyen, querying one day of raw payments at Adyen took 5-6 minutes. As a result, we started using a customized version of big data streaming platform - Apache Spark and can now query 3 years of raw payment data in less than 3 minutes. Without changes like this, analyses of all of the analysts in the company slow down, leading to a big decrease in efficiency in the system as a whole.
PCM: How did the rise of Data Science influence the Payments & FinTech space?
Daniel: More and more Payments and FinTech companies are using data science to solve a number of problems across the organization, for example uncovering insights around customer journeys, fraud prevention, corporate compliance, and overall service quality. Data science creates products that often become a differentiating factor in every field. Most of the best fraud prevention tools today use some form of artificial intelligence, many times leveraging expert knowledge in the payment industry. This combination of AI and expert knowledge is able to obtain much better results than any traditional tools. This combination is really where the future of the Payments and FinTech space is going, and you won’t be able to compete for very long without a strong data team.
PCM: How is Data influencing decision-making?
Daniel: Frankly, any FinTech company without a data team is making decisions in the dark. This is not just for fraud detection, but also for questions such as: What markets should we enter? Should we block the payment as fraud or route it through 3D Secure? Once the data is clean and visualized in front of you, the decision will be so much easier to make. At Adyen we have a very large reporting system that is constantly being improved, as we don’t only want to make good decisions internally, but also educate our merchants to make their own data driven strategic decisions.
PCM: How do companies make sure the Data is legit and can be used efficiently throughout the organisation?
Daniel: Cleaning data is a skill that every data analyst and scientist must have, and it’s not just removing outliers, incorrect/null values etc. At Adyen we make sure we define metrics that when looked at together will paint a more accurate picture of what’s going on and have those available whenever anyone looks at the data. Every data scientist should always be skeptical of the data they’re looking at should try to leverage their technical skills and domain specific knowledge to really ask themselves “Do these numbers make sense?” Especially when the numbers seem too good to be true, many data scientists will allow their confirmation bias to blind them to the faultiness of the data. This will unfortunately lead to incorrect decisions and a loss of time and money that could have been saved by looking at the data through a skeptical eye. We also do a Data Roadshow at Adyen where we send data scientists to other parts of the company every few months to look over how they’re using data and to come up with data driven solutions to make sure they’re using the correct data in the best way.
PCM: What purpose does Data serve in an organisation and on how many levels is the information used?
Daniel: Data should be involved in as many decisions as possible throughout an organization. I think a data driven organization can really be seen by how many people in the company actually see the data. Often, companies will create data “silos” within the company, hiding the data from all but one or two teams, relying on them to make decisions and share their results in reports. When a company is really advanced, the data is spread throughout the company, allowing anyone to easily double check the work of another and many times look at the data in a different way. At Adyen data is really used from top to bottom. We have built dashboards for our account managers to break down all of the data on their merchants in real depth, which not only allows them to make smart decisions, but also frees up resources from the data team to work on the more advanced projects. Management uses other dashboards to review growth of merchants and markets, compliance for AML alerts, and marketing for understanding the effectiveness of a campaign.
PCM: As a service provider, how does Data help you stand out of your competition and what’s in it for Merchants?
Daniel: Adyen has a true end-to-end platform, which allows us to have a full overview of every part of a transaction, in all markets and channels. Old payment systems are a black box to merchants. With them, transactions are either approved or declined, and that’s the end of the story. We have developed a number of data-driven automated tools, which work in the background of each payment to drive authorization rates. One of these tools is RevenueAccelerate, which by analyzing data to increase authorization rates for our merchants lead to an average revenue increase of 1.43%. Since our platform is fully omnichannel, these insights across channels also help merchants offer a more unified experience to their shoppers. Next to the business benefits, having access to these insights also helps our merchants think data first when it comes to their payment and commercial strategy.
PCM: How will Data Science help shape the future of the industry and what developments are we going to expect in the future?
Daniel: There are a lot of exciting things that will happen around risk management. In the near future ML risk systems are likely to be standard, and there will be many new data points when it comes to fraud preventions, such as identifying fraudsters by mouse movements on a webpage or scanning the dark web for breached card information. I also think that we can expect an influx of a lot of smaller players into the business focusing on one problem and doing it better than the big players by using data science solutions.
PCM: Any words of wisdom for aspiring Data Scientists or companies looking to branch out a Data Science department?
Daniel: For budding data scientists I would say just start. I studied music, biology, bio-medical engineering and fought in the special forces of the IDF. Later I worked as a Bio-Medical Engineer and as a CEO of a start-up, then moved to The Netherlands and decided to be a Data Scientist. I think my background makes me a better data scientist than if I had just studied computer science or something similar. The rest of our team is made up of Astrophysicists, Aerospace Engineers, Mechanical Engineers and Computer Scientists. Don’t worry that your background doesn’t “fit”. In the end, diverse backgrounds mean you have a more open approach to the same problem. For companies looking to branch out a Data Science department, remember that good data scientists can improve processes all over the company, not just in classic data science areas. Make sure your data scientists work on gaining as much domain knowledge as possible. A data scientist that doesn’t understand his domain can sometimes make great solutions to things that are technically difficult, but aren’t necessarily the most impactful. It’s also important to make sure your data scientists are merchant facing, as they can often be a key factor for opening up discussions in many merchant meetings that may lead to big decisions.
About Daniel Linder
Data Scientist at Adyen
Daniel Linder is a data scientist at Adyen, the technology company reinventing payments for the global economy. After starting his career as a Bio-Medical Engineer and CEO at the start-up GreenOn, he moved on to become a data scientist in 2014, joining Adyen in 2016. Daniel works on building models to improve authorization rates of payments across the Adyen platform and consulting merchants in solving data problems.
Adyen is the technology company reinventing payments for the global economy. The only provider of a modern end-to-end infrastructure connecting directly to Visa, Mastercard, and consumers’ globally preferred payment methods, Adyen delivers frictionless payments across online, mobile, and in-store. With offices all around the world, Adyen serves more than 4,500 businesses, including 8 of the 10 largest U.S. Internet companies. Customers include Facebook, Uber, Netflix, Spotify, L’Oreal and Burberry.