Ocado uses advanced analytics and cloud storage to deliver groceries right to your kitchen table

Ocado, the world’s largest online-only grocery store, picks and packs 1.4 million items every day and then delivers them to its customers kitchen tables in one-hour slots. Ocado is using a wealth of data about customers, orders and trends to provide customers with a better shopping experience.

Hosting an online shopping platform

Ocado, based in Hatfield, U.K., currently serves 72 percent of U.K. households and continues to expand and compete against brick-and-mortar grocery stores. Ocado is leveraging big data analytics to convince customers of the benefit of buying food online, rather than in physical markets.

Since Ocado started shipping orders in 2002, the company has continuously written and re-written the software underpins its retail business. Recently Ocado started building a new hardware and software platform that will host shopping experiences for large third-party retailers around the world to move online, called the Ocado Smart Platform. The company also launched an initiative to leverage its extensive data about its supply chain, customers and shopping trends in order to improve operational efficiencies and the customer experience.

To support these initiatives, Ocado needed a partner to host and scale Ocado Smart Platform as well as enable advanced data-driven analytics. The company also wanted to move business, product and transactional data out of inaccessible silos within its own data centers, and place it in the cloud where it could help drive business decisions and inform the supply chain to help with controlling inventory, predicting demand and improving logistical efficiencies.

The cloud’s analytics and scalability

Ocado migrated nearly all of its business information — currently more than 100TB of data — to Google Cloud Platform in order to knock down silos and enable advanced data analytics which in turn could improve customer service, empower its supply chain, speed decision-making and reduce costs.

Ocado initially used Hadoop and Apache Spark software on Google Compute Engine, but then saw the benefits of using Google BigQuery to create and display complex real-time analytics. GCP’s data solution lets Ocado’s business analysts avoid the complex setup and workflow required to run queries against Hadoop and Spark, and also helps them expand data analytics out to its supplier and partners.

“Our ambition was bigger than what we could handle,” says Paul Clarke, Chief Technology Officer at Ocado, explaining the company’s decision to move away from Hadoop and Spark, and instead embrace Google’s data platform. “If BigQuery can solve the problem for us, it’s our first port of call. We can throw the problem at BigQuery and get an answer back right away.”

Ocado also chose Google Cloud Platform for its scalability. “We’re building a platform on top of a platform, and we’ll be rolling it out as well as dogfooding it ourselves,” Clarke says. “We want to build it once and then turn it on for multiple business-to-business customers.”

With Ocado Smart Platform, each time Ocado adds a new hosting customer it will launch a new instance to fit that customer’s requirements. The capacity and performance of each new Ocado Smart Platform instance needs to be able to scale from zero to huge fast, especially since it’s the backend platform for established retailers with a large number of products, customers and transactions.

Focusing on business opportunities, not IT problems

Clarke and his more than 650 software engineers and technology specialists appreciate that they can focus on high-level strategy while Google Cloud Platform handles data queries. “We don’t have to do a lot of work to get an answer, as Google does most of the heavy lifting and scaling with the data,” Clarke says. “Since Cloud Platform has data centers around the world, we can launch anywhere without having to build our own data center there.”

Since moving to GCP, Ocado has saved on IT operations costs and headcount, compared to implementing additional services within its own data centers. To host Hadoop or Spark analytics as well as Ocado Smart Platform, the company would have needed to massively scale out its internal infrastructure, as well as hire hardware engineers, database administrators, operations employees and experts in on-site software platforms. Says Clarke: “Thanks to Google Cloud Platform, we can hire engineers to add value by solving real business problems and leave the cloud data infrastructure to companies like Google."