"Scaling was the bottleneck in our growth. But, thanks to Confluent, we’ve been able to re-architecture with ease and harness real-time data to create better, more seamless experiences for all."
Alok Sharma
Director of Engineering, Meesho
Meesho is India’s only true e-commerce marketplace, with over 140 million average active users that visit its platform annually.
After starting off as a B2B-focused reselling platform, Meesho has evolved to encompass a single ecosystem that connects millions of sellers, retailers, suppliers, and customers.
As part of its mission to democratize e-commerce, the company was experiencing substantial growth—owing in part to its transition to a B2C business model. But this immense growth was creating scaling problems, limiting bandwidth, and putting a strain on its Apache Kafka® deployment.
To cope with increased site traffic and demand, the platform needed a new systems architecture that was scalable, resilient, and agile. Meesho’s team wanted to ensure that its main focus remained on solving business and engineering problems, rather than managing its infrastructure and platform.
Today, Meesho uses Confluent Cloud to offload the burden of manually scaling systems and managing infrastructure overhead. As a result, the company can expedite time to market, minimize operational headaches, and push forward toward its mission of democratizing data.
“Scaling was the bottleneck in our growth. But, thanks to Confluent, we’ve been able to re-architecture with ease and harness real-time data to create better, more seamless experiences for all,” says Alok Sharma, director of engineering at Meesho.
A system that struggled to scale
After moving from a B2B business model to a B2C model, Meesho experienced extremely high growth and had to re-architect its backend systems to cope with an increased volume of data.
After certain events—like the festive sale period—Meesho would frequently see up to five times more traffic to its platform and app.
“The problem we started facing was more on the engineering side, as the amount of throughput we were handling on Kafka increased,” says Sharma. “There were limits in the bandwidth we could handle on the installation, and it meant we had to do system-level changes to ensure we could work around limitations.”
During these sales, Meesho’s systems would struggle to keep up with demand and this was causing downtime, which, in turn, was impacting the customer experience.
