Where Do You Turn for Help? A Top Fintech Learns a Hard Lesson in Customer Service

A business typically purchases a product to make a specific job easier or more efficient. Choosing a solution can be arduous, and sometimes there is little thought about what happens after deploying the tool. At Konfío, we recently learned that even the best tools get you nowhere without support from the right people.

Created in 2013, Konfío is Mexico’s largest online lending platform for small and medium businesses (SMBs). We offer support services like enterprise credit cards, capital loans, payment solutions and productivity tools, and our work is vital because traditional banks won’t lend to most of our customers.

Our operation is what you might call data intensive. We use a variety of complex algorithms and analysis tools to match our clients with the right loan products. Since the majority of our clients are micro, small, or medium-sized companies, most of them lack an in-house data department. That means the business intelligence (BI) we provide is often the only source of analysis they have.

I joined Konfío in 2021 as the Director of Data Architecture. Almost immediately, I noticed that the biggest challenge of my job was database management. We currently use Neo4j, DynamoDB, Achieve, DocumentDB, and Aurora, and simultaneously running these platforms requires a lot of data replication. The real-time nature of our work also poses a challenge. Our end users (our colleagues) often have to toggle between multiple dashboards to gather information for proposals. This action is very time consuming, and any interruptions in database access could be incredibly problematic. What’s more, whatever we do on the backend cannot impact the transactional database.

Managing multiple rapidly-growing platforms was also becoming problematic. While business growth is generally a great sign for any organization, it provides the ultimate test for data architecture. As the amount of collected information expands, even the most perfectly designed system can quickly become unwieldy. That can eventually complicate replication processes and make it hard to ensure the correct data is in the right place at the right time.

We worked around these challenges, but data architecture is about looking forward. We didn’t want to choose a solution to solve today’s problems. Instead, we wanted to procure a replication tool that would continue to serve us for years to come.

Qlik Replicate: The Dream

Our operation is seated entirely in the cloud. Once we realized the need for a replication tool, the difficulty was finding something that works within our internal cloud application. Additionally, any software we selected had to work alongside our other platforms without impacting the transactional database. For better or worse, the data team started searching online marketplaces for a solution.

Some of the tools we encountered were designed to pull information directly from the transactional database. These sparked the fear that the new solution would impact the performance for the end users.

In our research, we also encountered promising tools that charged based on the number of records replicated. Given the size and growth of our data operation, that cost model could get expensive very quickly. Eventually, we found a winner in Qlik Replicate. It ticked all the boxes: 

• It only required one license.
• It operated in the cloud.
• It could work independently of the transactional database.

Once we clicked to purchase, we were excited about the great product that would soon be in our hands.

Qlik Replicate: The Reality

That excitement turned into worry as we realized that this ticket might not be as golden as we thought. 

Our team was thoroughly confused during the configuration stage, which led to serious issues with the replication process. By purchasing through an open marketplace, we did not receive training or support, so we were on our own. Nothing is quite as frustrating as having a great tool without the know-how to implement it properly. We knew nothing about best practices and had no knowledgeable sources to ask questions. 

It was time for another approach. We reached out directly to Qlik for help. To the outsider, contacting the initial vendor may seem like the first step in seeking help. However, given that we did not purchase the product directly from Qlik, we had no clue how they would respond. Why would they help us when we purchased the product through a third party?

To our relief, Qlik was more concerned that we get the most out of Qlik Replicate than where we bought it. Qlik quickly contacted our technology partner, BAW Systems, and everyone put their heads together to set us on the right path. As a Qlik partner, BAW stepped up and provided incredible assistance without any signed contracts or hidden charges. BAW shared critical guidelines and provided valuable consultation, and with that, we began to unlock the full potential of Qlik Replicate. 

Through BAW, we were finally able to lock down task configuration and get a better handle on how to use specific features. It was always our plan to move everything into a data lake, but the Qlik team finally helped us figure out how. We are now implementing a very aggressive strategy to move everything to a data lake based on S3 buckets on Amazon. We are testing a feature where we can load all of this information directly into S3 instead of moving through other databases before moving into the data lake. We learned how to implement more data sources and different instances within relational databases. We are even thinking about moving our data to target repositories.

The more Qlik innovates, the more information they share with us and the more options we have to keep moving forward.

We Had the Power All Along

Qlik Replicate has a user-friendly interface, and once the BAW and Qlik teams helped us overcome our configuration hurdles, we began to recognize its power. We saw how to quickly and easily change columns and apply filters, and for the first time, we were able to make it work for our needs. Now the platform is part of our onboarding process, and people are encouraged to explore.

Qlik also helped us move away from so many real-time operations. Our previous approach seemed great—after all, who doesn’t love real-time information? But it took a lot of resources to provide that information, and it wasn’t always required. Qlik Replicate allows us to only run the real-time applications that are strictly necessary. Now we define a process in order to identify why someone needs information and when they need it. Based on that, we have more control of the requirements and also of implementation to provide information at the right time. We improved our batch processing capabilities, requiring less input from our end users. In doing that, Qlik Replicate saves us time, effort, and compute resources every single day without sacrificing data quality. And as a result of our better data pool and sharing process, we have sharpened our data collection process and learned to ask better questions of our clients.

Our work affects the business analytic teams and the data science teams, who develop models used by the rest of the business. Thanks to Qlik’s efforts, we now have 2,000 tables in our data lake, and nearly 1,500 running directly on Qlik Replicate. Based on those tables, our business analytics, data engineering, and data science teams run almost 200 dashboards and 20 machine learning models that inform our colleagues so they can make intelligent decisions.

The Product Alone Is Not Enough

We are thrilled with Qlik Replicate, but we learned a valuable lesson: the product alone is not enough. The huge investment of time and expertise by Qlik and our technology partner rescued our organization in a big way. Attempting to get support from before we reached out to Qlik was painful, often lagging three or four days before getting unsatisfactory answers. We suffered hours upon hours of lost productivity and made more mistakes in the meantime. We get answers from Qlik in a matter of hours, enabling our team to keep our business running smoothly and efficiently.

We paid in many ways for convenience. While online marketplaces are a convenient path to purchasing products, it’s only half the picture. Once you have the software, the proper service will maximize its potential. Working together, Konfío, Qlik, and BAW managed to rescue our plan and achieve a complete solution that could help us reach our goals. That’s what being a good partner is all about.