A business catch-22
Ocado’s technology has enabled Ocado to run an online supermarket profitably, scalably and sustainably for many years. Yet, when we decided to answer international demand, by licensing our grocery technology to retailers around the world, we were faced with some significant challenges we hadn’t met before:
First, the technology had to be easily implementable anywhere, globally. It needed to be scalable to new proportions, and be able to grow with demand simply unknown for any UK retailer. A single US retailer, for example, could have revenue worth more than the whole UK market.
It needed to be sufficiently configurable to allow our retail partners to adapt to their local markets while remaining a single, manageable, high-performance platform - taking into account things like language, currency, and preferred operational models.
Our existing Technology, evolved and refined over 20 years, couldn’t accommodate these requirements. We needed a new platform, built from scratch, covering the entire end-to-end process - from ordering to picking and then delivery. This platform would benefit from the latest cloud and microservice technology and be a foundation for the scalability and flexibility we needed.
The real product conundrum was this: In order to build a mature product for these massive businesses, we had to use and trial the platform; but until we had built this mature platform, our clients couldn’t use and trial it - they were just far too big.
To give you an idea of the scale of these businesses: Kroger, our North American client, reports “Every day the Kroger Family of Companies makes a difference in the lives of eight and a half million customers” with just shy of 500,000 employees. That’s more customers than the population of London, every day.
You might ask: Couldn’t you just redirect an increasing number of customers to this new website over time to build up the scale?
But this wasn’t simply a new website; the shift to this platform represented a new business model for our clients - a new physical operation. This involves massive changes to their business processes, technical integration to existing tools such as Finance systems, recruitment and training of pickers, drivers and online business staff - a significant amount of upfront investment. Retailers wanted a platform with well-developed processes and the robustness to scale as quickly as possible.
You might also ask: Couldn’t we just do exhaustive user research sessions with our clients?
We saw three challenges with this approach:
At the beginning, we didn’t yet have OSP clients. We had to demonstrate the platform to get clients on board.
Even if we did, the temptation would have been to design OSP for one particular business (such as Ocado.com) rather than creating a platform that has the adaptability to be one platform for any grocer in the world.
The processes for such a big business are so complex we would have ended up trying to build the end-game directly rather than iterating - a recipe for failure. We needed a ladder to bridge the gap - to get us from the very foundations of the platform to the point at which the platform was mature enough to be used by our clients and receive their live feedback.
We needed a ladder with many rungs - smaller clear steps to reach where we were going.
Our First Client
What was our ladder? We needed a ‘client’ - retail partner we could sell OSP1 to. We established the Ocado Smart Platform’s first client. An internal client. The customers? Our own hungry employees!
We had to build a convincing retailer persona that delivered a service that our employees actually wanted; only then would they care about the service that was being achieved using the fledgling platform and give us quality feedback.
Each step on the ladder set a new functional challenge for the platform and we had to evolve a service that would engage our staff as customers, and simultaneously trial the relevant functionality.
Our first challenge? To deliver a very basic order.
The very first iteration was to deliver chocolate bars from the vending machine in the office to developers’ desks. This was a particularly popular service.
I literally donned a headcam and a high vis and walked around the business delivering snacks that were ordered on our platform and picked from the vending machine, posing as a tiny warehouse.
Soon our platform had a heart beat - daily orders from our employees demonstrating that the foundations were in place.
On the one hand this feels very simple, as any person can start a basic webshop delivering a product. However, the foundations of this platform were broad - not just the webshop but the order management systems, supply chain systems, fulfilment, customer communications, and delivery.
We had learnt from Ocado.com how to architect these systems for the flexibility that would be required - for instance, how to compartmentalise the processing of fulfilment information so that retailers could support multiple business models in parallel: Home Delivery fulfilled from warehouses in one area, but Click and Collect fulfilled by Store Pick in others.
Demonstrating delivery of basic products proved that these technical foundations were sound - like demonstrating the basic plumbing when building an entire city.
On the other hand, it quickly pointed out a few rough edges. During that first order we had a damaged packet of Maltesers, and there was no way yet to refund or placate the customer. In practical terms, this was more important in producing happy customers and providing a basic service than, say, improving the look and feel of the webshop. This live feedback immediately enabled us to prioritise better.
Step Two - Growing the Platform
The next rung set the challenge of supporting a range of products. However, we couldn’t yet manage chilled produce. The service we developed to trial the platform was breakfast delivery; we could provide a compelling range that filled a need for employees and didn’t require chilled produce (milk was already on hand for tea).
The rungs became very tangible and told us immediately where we needed to adapt - if a feature missed the mark, no-one would get their breakfast.
Out of the building
We needed to demonstrate delivery on the road as well as the ability to deal with the real life challenges delivering grocery-sized orders. So we got our own vans and began delivering full customer orders to the homes of our colleagues. We even teamed up with a local supermarket to try out our in-store picking technology and offer a more realistic range.
We had developers picking and packing, living the experience of running the platform, as well as being customers; all the time trialling and iterating the platform.
In our most recent iteration, we even built a micro version of Ocado’s signature highly automated warehouses to deliver these orders and trial the entire platform for our internal business.
Reaching the Launch
This ‘ladder’ technique became a primary way to help us understand how to prioritise before we had external clients giving us feedback. It became a sales tool to attract clients by demonstrating our platform in action, and it became our way of emulating our clients’ businesses ahead of successful launch.
While we could never demonstrate the scale we could sufficiently emulate the first stage of their roll-outs to build critical confidence.
Our ladder bridged the gap - it led us up, rung by rung, until we created the platform of a scale and maturity that our clients have adopted and can now grow organically.
Even once we are live, we continue to run our internal retailer - we use it to find ingenious ways to trial new features ahead of sharing it with our clients.
It's a good example of how we are constantly innovating and optimising.
Change your world with us
Across Ocado Technology, we have a diverse, rich mix of teams and expertise working to solve complex problems. Learn more about our full range of opportunities here.
Java - a key language for OSP
Java helps us develop intelligent systems and high-performance applications serving our business from webshop to warehouses. Matthew Cornford explores the role of Java in our techstack and why it’s still so important.learn more
The SecondHands project
Ocado Technology and partners have been pioneering a collaborative robot that can proactively assist humans in maintenance tasks. Dr Graham Deacon describes the amazing The SecondHands project.learn more
In store smartness
While our smart warehouse robots are the ones that often grab headlines, our in-store-fulfilment (ISF) solutions allow retailers to meet the growing demand for online orders using existing stores.Learn more