mPokket centralises digital intelligence analytics and telemetry

Founded in 2016, mPokket is an instant loan app providing financial assistance to students and young professionals in India. Users can download and install the instant loan app from the Google Play Store. Borrowers can then take out instant cash loans anywhere between Rs.500 – Rs.30,000 depending upon their profile and credit history. The money is transferred directly into their bank or Paytm account. Repayment plans are flexible with borrowers having the option of paying back the money over a period of three months along with a nominal interest charge.

Traffic volatility and a lack of visibility

With 1.3 billion people, India is one of the world’s largest and most complex credit markets. Increasingly, incomes and consumption are the driving demand for retail credit. The fintech market is estimated to reach US$150 billion by 2025, according to Invest India, making it a highly attractive prospect for new entrants. Instant loan app mPokket is one startup hoping to capitalise on this growing opportunity. But with competition becoming increasingly fierce, providing a flawless experience is getting all the more critical as it relies on customer trust.

"The company is experiencing rapid growth right now. We have drastically grown in terms of our business as well as in terms of engineers and other areas. And we are maturing in terms of process as well. With all this growth, we started looking at how we can get proactive to resolve issues even before end-users report them. We wanted to be sure whether our applications are working fine," explains Ajay Garga, VP, Technology, mPokket.

Traffic volatility is a particular issue for mPokket. Users typically run low on funds in the second or third week of each month, which results in a rush for loan applications during this time, with most repayments made in the first week of the following month. mPokket runs a microservices-based architecture, 100 per cent on AWS, with Kubernetes orchestration and a predominantly PHP-based stack. But it was starting to hit a few pain points, both with its internal application and the cloud solution in its tech stack.

"We ran into a couple of production issues, and at that point in time, we really felt that we didn’t have observability in place, which was causing a delay in the entire debugging process. We didn’t know what was causing the bottleneck, and it took a few days to identify the issue because the visibility wasn’t there," Ajay recalls. “We had to deal with customer complaints and we took all feedback very seriously. Thankfully, now with New Relic, we can identify these pain points very quickly before they escalate.”

A single solution with multiple dashboards

Ajay has been an advocate for New Relic for many years, being a power user in his previous role. He says New Relic was amongst several services evaluated at mPokket and was ultimately determined to be the best fit for many reasons, including its user-friendliness and compelling dashboards.

"We wanted a unified observability solution to avoid tool sprawl. New Relic was able to demonstrate these capabilities, ultimately enabling us to eliminate some of the existing tools that weren’t being fully used," Ajay says.

He says implementation was “very easy” due to his familiarity with New Relic. Team members were also able to take advantage of the free training courses provided by New Relic. “With the engineering team continuing to expand, being able to get new staff up to speed quickly was vital,” Ajay says.

He explains that New Relic was set up to accommodate three types of users – himself as the administrator, engineers - a pool of assets were created where they could obtain data about every single app that was on-boarded to New Relic, and the third group comprising dashboard users, including senior executives, who want a snapshot of the main metrics and performance over certain periods.

Hitting KPIs, ready for rapid scaling

“One of the biggest benefits of New Relic is the impact on mPokket’s KPIs,” Ajay says.

Recommendations provided by the New Relic team for tool consolidation were immediately implemented under Ajay’s leadership. This ultimately led to mPokket improving on their KPIs–including identifying the availability and latency of apps that were difficult to determine previously.

“With New Relic, all the data was there. We could narrow it right down. We could see what the latency was in peak hours. Previously, we didn’t know what was happening.

“We are a B2C company so definitely the scale will keep increasing. Today, we have about one million active users but this could grow to 10 million active users within a few months. We need to be confident about what the latency is,” Ajay says.

Another KPI was the error rate, which mPokket needs to be “very measurable about, ideally, less than 0.5 per cent”, he says, accepting that this may fluctuate as new apps are rolled out but “to be able to tell what they (error rates) are in the first place is a big improvement”.

According to Ajay, MTTR (Mean Time To Recovery) is another KPI and shows how quickly mPokket can identify and fix an issue. With New Relic, engineers can finally take a proactive approach to troubleshoot and resolve issues before they escalate.

“When an issue comes up, we immediately look at New Relic to figure out what is happening. When I talk to stakeholders and senior management it’s like having a solid line of defence. A problem might occur at any point in time so you need to be constantly monitoring your infrastructure and applications. With New Relic, this is easily achievable,” Ajay says.