Ocado realizes six-figure savings as it moves to the cloud with New Relic

The Ocado founders were ahead of their time when they started the UK’s only "pure play" online grocery operator in 2000. Today Ocado exceeds £1 billion in annual sales. Developing the innovative software and systems that power the online retail business is Ocado Technology, a division of software developers, engineers, researchers, and scientists spread across five offices: Hatfield, UK; Wroclaw and Krakow. Poland; Sofia, Bulgaria; and Barcelona, Spain.

A typical Ocado warehouse packs and ships more than 1 million items each day, supported through a technology platform built in-house. Additionally, the Ocado family now also includes Fetch, a specialist online pet store, and Sizzle, a homeware company, both based on the technology platform built by Ocado Technology. Recent e-commerce sites such as Morrisons and Fabled, a beauty destination site in partnership with "Marie Claire" are more recent success stories for Ocado Technology.

In parallel, Ocado Technology is also developing the Ocado Smart Platform, an end-to-end e-commerce solution designed to put other brick-and-mortar retailers around the world online. From its user-friendly mobile apps and webshop to highly efficient automated warehouse technology, OSP offers superior customer experience and a highly efficient end-to-end supply chain solution. Ocado now offers OSP as a managed service capability to partners internationally, enabling them to build sustainable, scalable, and profitable online grocery businesses in their own markets.

Moving to the cloud demands increased visibility

In 2014 Ocado found itself in a transitional period. Demands on its business were increasing and Ocado felt that a move to a cloud infrastructure would give the company more flexibility and scalability. Amazon Web Services (AWS) was chosen as the target environment.

"One of the challenges was that the dashboards we had developed to monitor the various sites and the platform behind it were not compatible with AWS and needed to be replaced. I had used New Relic technology before and wondered if that might be our answer," says Peter Thomas, head of e-commerce OSP at Ocado Technology.

New Relic was introduced to support the cloud migration, including integrating well over 100 microservice applications into the OSP platform. "New Relic APM provided the extra visibility into our systems and platforms, which we needed. It gave us confidence that our migration was going as expected, and, if not, we were in a very good position to address any issues early. We learned lots about cloud architecture, and it was great having New Relic there to reassure us nothing was quietly going wrong behind the scenes," says Peter.

New Relic is used to pinpoint the root cause of any performance issues. When one of the OSP backend systems showed a gradual performance degradation that was in danger of impacting customer orders, a lot of effort was spent on finding out what was wrong. Confusingly, different systems that were seemingly unconnected were affected. It wasn’t until an engineer used New Relic dashboards to show the problem that it became clear that although the database systems were all different, they were hosted on the same physical machine, which was the root cause of the problem.

"New Relic made it easy for us to visualize and track this issue in real time," says Peter. "It helped resolve this particular problem a lot faster than we could otherwise have done."

A better customer experience

Ocado Technology works with a globally distributed team of 1,000 developers, all working on different parts of the estate simultaneously. New Relic provides a central overview and a common communication tool. The features to enable effective team collaboration come straight out of the box with New Relic: "We’ve tried other products which needed a lot of configuration. As soon as you start customizing the solution with so many teams, it becomes very difficult to manage and maintain. New Relic did an excellent job at giving us some very sensible defaults which we could easily build on," says Peter.

The New Relic team also took an active interest in how Ocado used the toolset. A center of excellence was introduced, consisting of New Relic and Ocado experts. As a team, they discuss the New Relic roadmap and what other features could prove useful. Having that deep insight into how OSP operates enables the team to add value to Ocado.

Dashboards to support development

When Ocado rebuilt the frontend of some of its non-food sites to become more customer-responsive, New Relic browser monitoring proved invaluable to the experience. "The general trend we’re seeing is that code is moving from the backend to the frontend," says Peter. "Most other solutions don’t seem to take this into account, but New Relic browser monitoring enables us to see what’s actually happening within our customers’ browsers. It gives us confidence that we are building solutions that work for our customers in the real world, and that if they don’t, we’ll know about it straight away. You can’t get this kind of insight just from looking at your server logs."

The combined data from New Relic APM and browser monitoring has been used to highlight performance issues, and identify where problems originate so they can be addressed. Dashboards, meanwhile, delivered solid ROI in one of Ocado’s warehouses. These highly automated customer fulfillment centers pack and ship more than 1 million items a day from a stock of over 50,000 different grocery items. A part of the warehouse suffered a broken sensor that was reporting an exception, even when this wasn’t accurate. The systems are event-driven so without counting the frequency of these exceptions Ocado had no way of identifying this issue. Through dashboards and alerts, the issue was quickly identified and passed over to engineering for fixing. The storage capacity, which was lost due to the error, is worth £35,000 per year in real labor costs. It showed Ocado the potential for making incremental gains through New Relic. Overall, improving automation and monitoring capabilities with New Relic has realized savings of £100,000 per year.