Discover how a team of 5 data experts uses Keboola to manage the entire data operations.
GoPay, the fully integrated payment solution, serves over 18.000 merchants.
As a payment gateway, GoPay processed over 100 million transactions, connecting merchants, retail customers, and financial institutions in a seamless payment experience across mobile, desktop, and web.
Their work is challenging - on one hand, the data they work with needs to be thoroughly validated to comply with financial regulations. On the other hand, data collection is challenging due to the volume and speed of the operations.
We talked to Jiri Zezula, the Data Analytics Manager at GoPay, to discover how his team of 5 data experts manages the entire data operations for GoPay.
Jiri has worked with payment data at GoPay for over 11 years.
Back in the day, it was Excel hell.
“Before you had any devoted products, you had to create your Excel reports by getting data from some transactional SQL database. And you actually had a task so you manually sent out Excel reports to all parties. It was tedious, it was slow, and nobody liked it.”
As Power BI made its debut, Jiri and his team jumped onto the bandwagon and migrated analytics and reporting to the new tool.
But migrating the data ingestion to a different BI tool did not solve their issues.
In the early days, Power BI was extremely awkward to work with. You could not format the tables, a lot of ingestion required manual work. You had to export CSV and XML files manually from data sources, and spend a day or two writing scripts to clean and reformat the data, before you could send it to Power BI and visualize it in a friendly manner.
Jiri was thinking of building an intermediate data layer between the data operations and BI tooling, to lower the manual workload needed for the reporting.
By coincidence and word of mouth, GoPay started using Keboola for some simple data exports.
Jiri jokes they thought of Keboola not as an ETL tool, but as a data warehouse. They started collecting data from Google Ads and Google Analytics with Keboola and in Keboola, which was much faster than manually exporting and formatting the data.
And then it clicked.
“Over time we discovered that you could work with data in Keboola in a more complex way, linking individual sources into a single data model. The power of Keboola’s ETL became apparent when we started tackling more complex projects. This developed to where we are today, where we build most of the data models and the OLAP cube specifically in Keboola. And then we pour the cleaned data into our warehouse and other BI tools.”
GoPay works with multiple financial institutions. From banks to other payment gateways. And when working with financial data, you have to be careful. The data is highly regulated and needs to abide by regulatory standards.
“It's about money, so there's not a lot of room for error. And that's where we control or audit what we know versus what the third party, the provider of that payment method, sends us. We know, for example, that we expect there to be XYZ payments and they send us their data and we have to check to see if it meets that expectation.”
So Jiri and his team make sure the data matches up.
When they receive billing or transaction reports from a merchant or a payment gateway, they use Keboola to quickly and automatically match transactions to their backend to make sure everything is processed according to the given reports.
With the volume of transactions, GoPay handles daily, you cannot monitor every process manually. You need to build a system to monitor it.
“There are hundreds, if not thousands of these processes in the system. And somehow we need to control the inputs, we control the outputs, but we also control the internal processes, we control the inconsistencies that arise or may arise, because after all, the human operators that work with that system are not infallible.”
By automating the process monitoring and validation with Keboola, they can scale their operations and save time that can be devoted to other - more interesting and productive - initiatives.
General reporting stems anything from evaluating marketing performance to keeping an eye on financial risks.
The data is used to answer anything from “what was the average payment in the last quarter on the apparel segment” to “how much money is going into a specific bank”.
They use the data to understand the market trends (hint: it is all about mobile payments), opportunities to grow, and to showcase their value to the big financial player and partners.
All of the reporting is automated end-to-end with Keboola.
But for the data to bring insights via reporting, they need everyone on the team to access and use the data. Enter data democratization.
Jiri and his team have a saying:
“Stupid work is supposed to be automated.”
The moment they realize someone is producing a daily Excel report for their work, it is a red flag.
To avoid such inefficiencies, they work hard on data democratization.
“On the basic level, everyone from my 60+ coworkers has access to the data they need. We're trying very hard to democratize the data. For example, if someone asks you in the company what the average payment is last quarter in the apparel segment, that report just has to exist. It has to be somewhere and it has to be maintained.”
With Keboola, they can empower every individual to search and find the report they need themselves. With fine-grained access control, Jiri and his team can give access to each individual in GoPay, while keeping the sensitive data protected.
This is why every new analyst needs to go and complete the Keboola Academy once they join GoPay. It empowers them to answer their own questions, instead of having to rely on engineers to provide the data before an analysis can be made.
When we asked Jiri about the overall value of Keboola for GoPay, he was reluctant to admit that they saved over 2x the manhours to build the entire data operations for GoPay.
“I'm sure everybody will always tell you, Keboola saves us entire people’s workloads. But no, it's not just that. With Keboola you need more people to run operations because suddenly you can do projects that you could never do without it.”
Keboola made possible for GoPay what a team of devoted in-house developers could not build in the same amount of time, at a fraction of the cost. Automated ETL, integrated reporting, the democratization of data, and many other facets.
As Jiri said:
“Keboola is kind of a Swiss Army knife for data analytics.”
Ready to discover what Keboola can do to help your business? Get in touch with us and let us start building a better data future.