Building a Fintech Product in a Paper-Pen World
Back in 2016, at a time when fintech had just started making its mark in India in the retail space, Drip Capital set an aggressive vision — to take technology into commercial lending, particularly in the SME space. We wanted to do away with the prevailing deep-diligence practice followed by banks; remove the subjectivity, strip the process down to its basics, and create a quick and objective way of assessing applications. While this had never been done before, we knew from the outset that it was the only way to scale rapidly in the sector.
Step 1 — Drilling for data
They say data is the new oil — the most valuable commodity of our time. If we wanted to remove the subjectivity from risk assessment, we had to pre-define the data we wanted to assess and then drill for it. This data needed to have three key attributes — to be from a reliable source (for managing our risk), to be electronically fetch-able (for scaling up) and to be continuous (for real-time monitoring).
We identified and partnered with nine different data sources to paint an instant and real-time picture of an applicant company’s performance and KYC information, eliminating the need for a paper application.
The fact that customers were so accustomed to the archaic process that they would send us 200-page bank statements anyway is a whole other story; many did so simply because they didn’t feel like their application was complete without sending the papers.
Step 2 — Knowing our customer (behavior)
While data integration helped us eliminate the need for a KYC process, we still needed to design the customer conversion process. We needed a keen understanding of customer behavior, most importantly the ideal pricing for a product-market fit.
Finding this ideal pricing was a series of trial-and-error efforts, looking at the reception to a range of pricing options for companies of different sizes and industries. The value of a credit line for a commodity trader, for example, is very different from that for a seller of branded garments.
For the first, supply-demand equations are usually not a limiter for business; rather it is their access to credit, or lack thereof. For the latter, the value may partly lie in protecting their receivables with the operations of a lender. The idea was to be able to divide the customer base into segments and accordingly modulate our value proposition, pricing and internal processes. Thus, we built a modular architecture for the product to be able to sift customers into segments and prompt different processes and protocols as needed. This modular framework has helped us manage risk and grow the business 100x in 30 months.
Step 3 — Building the product
While the company’s position and state today may seem intuitive, the process of learning was far more gradual. You often don’t know the full expanse of the product you will need till you have gone through multiple cycles of market feedback and discovery. We had started with a simple, one-product-fits-all model that worked for a section of the market. However, as we discovered nuances in trade operations, banking and logistics, we had to create new products, and add pieces and controls into specific parts of hefty risk control modules. These large additions had to be made with utmost care to protect the wider product ecosystem from any ripple effects, much like a delicate transplant surgery. This served as a constant litmus test for the scalability of our architecture.
Step 4 — Scaling up teams
A critical component to scaling up the business was the scaling up of internal teams. The foremost requirement was to establish compliance and controls within the system. This involved classifying the various backend functions and parameters based on who should have access to change them under what conditions. We then went on to break the organization down into teams and levels accordingly and thereby enabled user access controls on our entire system.
An interesting challenge that cropped up when we were scaling teams was knowledge bias. As we scaled, some functions scaled better than others. Some areas started to lag in terms of turnaround times or internal service deliveries. The first versions of our internal CRMs were very functional, created by a small team of users who had in-depth knowledge of our business and knew the ‘whys’ behind every task we would perform. Processes were far more instinctive to this set of users. When teams grew and new users were introduced to the system, we found ourselves increasingly identifying and streamlining the less intuitive workflows based on areas that were affected. This, in turn, was a valuable lesson in the requirements and challenges posed by scaling up products as well as personnel.
We always wanted to maintain a lean organization, to let automation drive processes. Where manual intervention was inevitable, we had to drive individual efficiency. This meant creating internal user experiences with ‘task flows’ that identified actionable tasks for each user from every team.
We also layered in a task assignment and commenting system to ease inter-team coordination. As a result, we now have teams in five offices across the world constantly working together to maintain Drip’s same-day service delivery and aiming to scale the business five times per team member.
Step 5 — Going Global
A validation of our stable and growing operations in India got us thinking about other markets. Most developing geographies have a steep demand for capital; many have growing SME sectors largely under-serviced by existing lenders. Mexico was one such market that presented large potential.
As we debated and strategized the launch into this new market, it prompted us to think of our existing system in abstract form, as three individual platforms — an end-user portal, a risk management tool and a customer management platform — each of them vital to our business, irrespective of the geography. Thanks to the existing modular nature of our system, all we had to do was replace data sources and adapt the pricing, and we were essentially ready to go to market.
Having now validated the product in multiple countries through strong growth and healthy loan performance, we can say we have built a product that is truly scalable. While building a good product requires a strong understanding of the market and listening to user stories, at the same time, it is very important to isolate and focus on the things that really drive business impact.
We have done a good job of balancing growth and risk management in this phase of growth. Exciting challenges lie ahead as Drip takes on its next 100x growth phase and further drives innovation in new areas of the trade ecosystem.
