BigBasket Grows Bigger with 400,000 Daily Orders on AWS

The online grocery market in India is competitive and value-driven, with some 20 companies vying for shoppers in a market that’s expected to hit $10.5 billion by 2023. The central government’s development of smart cities includes infrastructure to support the growth of ecommerce, and many large companies are making plays to get in on the action.

BigBasket, India’s largest online grocery, has been in business since 2011. Customer retention is at the heart of its strategy, as well as a hyper-local approach to inventory. BigBasket is available in 21 cities and runs its infrastructure on the Amazon Web Services (AWS) Cloud. The business had been growing 20–30 percent each year, but when COVID-19 struck and India went into lockdown, BigBasket faced unprecedented demand. Daily traffic was suddenly six times higher than pre–COVID-19 peaks.

BigBasket subscribes to AWS Enterprise Support and immediately consulted with the AWS team for advice on scaling its architecture efficiently. “AWS Enterprise Support has been our go-to destination for all queries when we’re building something new or maintaining and enhancing existing functions,” says Rakshit Daga, chief technology officer at BigBasket. “Given the breadth of AWS solutions we use, we have questions on all kinds of topics, and Enterprise Support connects us to the right person quickly.”

The company had already begun shifting to asynchronous work queues in 2018, isolating workloads into containers and breaking up its monolithic architecture. However, the onset of the 2020 lockdown shifted this process into high gear. By mid-2020, the company was fully containerized with 60 microservices running on Kubernetes.

Automated scaling features that come with a containerized approach has helped BigBasket accommodate the extreme spikes in traffic the site now experiences. Daily order volume was averaging 250,000 in 2019 and had risen to 400,000 each day by late 2020. Even during its transition to containers, BigBasket maintained 99.9 percent uptime on its website, with response times under 350 milliseconds for its customer-facing APIs.

To automate database administration, BigBasket has been using Amazon Relational Database Service (RDS) for MySQL for five years. Daga estimates the company is saving about one week per month by using Amazon RDS instead of managing databases manually. This approach also came in handy in meeting increased demand in 2020, with engineers quickly scaling up Amazon RDS instances as needed.  

Furthermore, due to efficiencies gained with containers and automation, BigBasket managed to bring the monthly cost of Amazon RDS down to pre–COVID-19 levels, even though its database was serving a much higher volume of traffic.

“With AWS, we can focus our energy in areas that are ripe for optimization. We can understand costs in a very clear way and can take advantage of various AWS tools to move to a different architecture quite seamlessly,” Daga says. “The fact that it’s a very open environment allows you to tap into services at multiple levels, so we can decide to go with a managed service or operate with simple Amazon Elastic Compute Cloud [Amazon EC2] instances for each application. This gives us a lot of flexibility in defining and designing our own architecture.”

One of the reasons for BigBasket’s success is its focus on analytics to expedite and personalize the customer experience. The company uses Amazon Redshift as a data warehouse and Amazon Simple Storage Service (Amazon S3) as a data lake. Using clickstream analysis, the data team gains insights into each consumer’s behavior that help predict what that individual—and similar consumers—are likely to buy. “With Amazon Redshift and Amazon S3, we’re able to carry out data crunching that drives every aspect of the customer journey on our website,” Daga says.

The fundamental aim of BigBasket’s data and analytics activities is to make the online shopping experience more intuitive. Online grocery shopping appeals to customers who are short on time, so the ability to log on, make the day’s purchases, and check out seamlessly is key to gaining repeated customers on BigBasket.

A feature called Smart Basket, for example, makes recommendations that help customers quickly navigate to the products they want, largely drawing on past purchase data and on shopping trends in the area they reside in. A year and a half ago, BigBasket introduced its BB Star loyalty program, powered by insights on a city level, to offer members discounts on delivery as well as preferred delivery slots, among other benefits.

Sales forecasting is another critical application of data insights, enabling BigBasket to stock its local warehouses with the right products to cater to local demand. The company has achieved a high degree of customer segmentation based on location factors, using Amazon Elasticsearch Service to perform geo-analysis and define service areas.

“A core value proposition of our business is being hyper-local,” Daga says. “We’re able to customize to very small geo-polygons within cities to strategize demand-based inventory.” Data also informs BigBasket’s decisions when setting up new warehouses based on customer density, helping drive the right product assortment for the area.

Whether scaling infrastructure to accommodate rapid business growth or developing new data-driven initiatives, BigBasket relies on AWS Managed Services (AMS) to reduce the time and maintenance burden on its engineers.

“The comfort of operating on AWS is there are certain aspects of our infrastructure that we don’t have to think about,” says Daga. “In an on-premises infrastructure, we would have to consume a large bandwidth of our engineers’ time, but on AWS we get new capabilities straight out of the box. With AWS, we have all the pieces of the puzzle to smoothly run a large ecommerce architecture.”