The pandemic has ravaged grocery chains, but tech is helping Ocado thrive

When historians are looking for images to illustrate the U.K.'s COVID experience, they could do worse than this Ocado screenshot:

At the height of the pandemic, with the entire country in a strict lockdown and supermarket shelves stripped bare, virtual queues to access the online grocer's website stretched to over 200,000 people. After years of slow but steady progress, Britons were finally adopting ecommerce en masse, and the providers were struggling under the newfound burden.

Ocado executive Luke Jensen told Protocol that the company was better placed than most to handle the surge in demand. Jensen is CEO of Ocado Solutions, the tech arm that underpins Ocado Retail and other third-party retailers, such as Kroger. With a platform that powers everything from vehicle routing to warehouse robotics, Ocado is arguably the closest thing grocery has to an AWS. And it's delivering the results to suit, with its stock price up more than 77% this year.

"Ocado Retail has grown sales over the period by close to 50%," Jensen said, "demonstrating the scalability of a model that actually works." And the company has big plans to scale it further, rolling out a number of robot-equipped fulfillment centers next year.

In a conversation with Protocol, Jensen explained how Ocado's tech platform has allowed it to scale profitably, and outlined the company's ambitious robotics plans — and why he's not worried about Amazon.

This interview has been edited for clarity and length.

How has COVID-19 impacted the Ocado Solutions business?

We follow markets around the world, and if you look at different markets they were — pre-COVID — all at different stages of evolution in terms of the proportion of grocery markets served by ecommerce, and in terms of the level of engagement of boardrooms in how important ecommerce is to their future.

If you look at the history in most countries, it sort of starts with a degree of denial, to realizing that it's going to be a major part of their business in the future, and therefore they need to find viable sustainable solutions that work economically and deliver good customer experience.

With COVID, nearly all markets have now tipped into that mature bucket. Ecommerce is going to be a very important part of the grocery market, and every major retailer needs to have scalable workable solutions to serve it. Of course, that is highly relevant to Ocado, given that we are recognized as having already been the best around the world at cracking a very difficult problem, which is taking a super-low margin industry and [running] ecommerce in a way that doesn't just add cost and take value away.

What formed Ocado was reinventing the grocery model to make it fit for ecom[merce], versus the standard way, which has been largely in-store picking, which essentially takes the full cost base that already exists in the retailer and adds extra cost on top for a service that customers want, but they're not necessarily prepared to pay for in full.

Are there particular areas of the Ocado tech stack that you're seeing interest in?

The reason why we sell our solution as an end-to-end is actually because every time you start slicing those bits into separate components, you lose something along the way. A lot of the power of Ocado is how the data flows across those different elements of the proposition. For example, the fact that our systems run the routing of the delivery vehicles means that your ability to offer slots and to price the slots appropriately for customers can take into account what routing options you're going to have, in terms of optimizing your routes.

In the same way, one of the things we do for Ocado Retail in the U.K., but also for our partners, is the ability to show accurate sell-by dates on every single product on the web shop. You can do that because your web shop is closely integrated with your fulfillment.

Ocado is probably most famous for the [customer fulfillment centers, or CFCs] and the robotics, but the power of what we do has a lot to do with that full end-to-end system.

How has Ocado Retail dealt with the huge increase in demand? Did it make it harder for you to scale up quickly by having to install new robots rather than just hire a bunch of new people?

If you look at the data, Ocado Retail has grown sales over the period by close to 50%. Considering some people said "you're capacity-constrained, you can't grow," that's quite a lot of growth.

Probably the even more important fact that we reported was the EBITDA contribution of the U.K. business was up 87%. And that's critically important, because that demonstrates that the model, as you scale it, has a powerful conversion-to-profit rate.

The reason why I raise that is not so much making a profit out of that exceptional period, because it wasn't particularly our aim, but it's demonstrating the scalability of a model that actually works. If we're saying demand sticks, it becomes all the more important to be able to have a profitable model today that will serve it. Otherwise, frankly, I think some of these retailers are actually going to probably voluntarily cut back on capacity as they find that they can't serve that demand profitably.

Do you think the reason that the model is scalable for you is because of the tech?

It's the whole business system. If you're doing the prevailing model, which is store picking, [to] do more of it, you just have to throw more bodies at it. You gain nothing as you throw more bodies at it, but actually you also start running into constraints. The typical benchmark in retail is if you're running a store and more than 10% of the sales of that store come from online, it starts to have a big impact on the store operations and starts having a big impact on customers. So, when Sainsbury's announced that online is currently making up 17 to 18% of their sales, they will be running into that challenge in a lot of places.

The benefit of CFC-based fulfillment is, of course, you're taking that all out of the store, and you're running it in a way that is very scalable within the CFC. Of course, you need to build it to a certain size and you can use capacity within that constraint, but certainly within that constraint it's very scalable, and the more volume you put through, the more efficient it becomes.

How is the Kroger rollout going? Was that affected at all by COVID?

We have projects around the world, not just in the U.S. with Kroger, and of course one of the challenges was making sure that those projects stay on track.

Most of the resources that actually work on the physical implementation of projects is obviously in-country, so that wasn't necessarily affected by movement and travel restrictions. But obviously we had to work our way around travel restrictions, and we spent a lot of time on screens. We've trained people with GoPro cameras going around the CFCs to show people in America how they work over here.

But during that period, we went live in the spring in Paris and in Toronto, with Casino and with Sobeys, and those facilities went live on time and have now been successfully operating and ramping up.

With Kroger, those projects are on the way. The first two will be coming [online]` in the first half of 2021. And those are close to Cincinnati and in Florida. There's been some COVID disruption, but in the main we've managed to make up for it with extra work, extending shifts, to manage to make up for any impact on the timelines.

You made some acquisitions recently in the robotics space. What was the driver behind that — what's the long term plan?

The robotic solution we have that's at the core of our warehouses currently is robots moving stock. We've completely optimized the movement: the storage, retrieval, placement of stock and that stock is then brought by robots to picking stations where human beings take an item from a bin that's presented to them and put it in a shopping bag.

Among the jobs that still require labour in a CFC, that is the one that consumes most labor. It consumes a lot less labour than picking in a store, because one of those pickers will pick a 50 item order in six minutes, versus around 40 minutes equivalent picking in store, so it's super efficient — but there is still a lot of labor there.

We had already started implementing some robotic picking: effectively robotic arms that pick those items and put them in a shopping bag. It's a super challenging technological problem. Sometimes people think of groceries being a relatively basic sector, but actually if you think of it, we're running a range of 45,000 products and they literally come in every shape, size, degree of fragility. So the whole apprehension of the product, seeing where it is, catching it, moving it is a big, big challenge, compared with robots that are being asked to move just square boxes. And then of course you need to put those items into shopping bags, without breaking, squishing and do it in a neat fashion. So it's a super big challenge.

We have a team of robotics engineers who had been working on robotic picking within Ocado. But we decided that we wanted to be absolutely at the front of the world in terms of capability in the area of robotic picking. So some time back we went on a global search for who were the best businesses in this space, and ended up finding Kindred and Haddington. The two businesses that are just absolutely fantastic: Kindred in creating all the software that powers the vision systems, that powers the arms, and Haddington actually in the design and manufacture of the arms themselves. And fortunately we were able to buy those businesses.

With those businesses plus Ocado's capability, we have put ourselves in a position where we have the very best talent in the world at solving the problem and we are confident that within two to three years, we will be able to implement in a widespread way robotic picking in our facilities and those of our clients. Initially, it won't pick the whole range, but gradually it will pick more and more of the range. There'll be enormous economic benefits and productivity benefits.

The other thing I would say is that Kindred up until now have been developing applications that are mainly used in the apparel industry — they work with Gap — but also in parcel moving, people like UPS. That's very exciting for us too, because it's effectively taking Ocado Solutions into new sectors beyond grocery.

You talk about wanting to be the world leader in this space: Amazon is also spending a ton of money trying to crack this problem. How worried are you about them and their presence in the space?

It's interesting, because obviously Amazon has become a pretty formidable enterprise, and they have very major positions in a lot of sectors of retail — but less so in groceries. They've been trying to do grocery in various shapes and forms for many years, and their position in grocery is still pretty small. They've got a bigger position in some subsets of grocery like health and beauty but because of course that's more of a kind of cupboard thing.

I think sometimes people say, "How come Amazon doesn't do this better than Ocado does it?" And I think the reason why Ocado does it well is because there's been a single-minded focus on this very complex category.

The same way I was explaining the complexity when it comes to robotic picking, here's another thing: In virtually every every other sector of retail, so most of the other things Amazon does, when people place an order they're placing an order of about two items, and they don't actually care whether those two things arrive together. So if I'm buying wipes for cleaning my glasses and a mobile phone, I don't really care whether those things arrive separately, and it's just two items. Logistically, that's a very, very different problem to solve from "I'm shopping for grocery and I want 50 things and I want them at same time." Nobody who's shopping for groceries wants their carrots in one delivery and then their ham in the next: They want one delivery. When you start thinking of what that means in terms of warehousing, you end up with a very different-looking warehouse.

Ocado has been able to completely focus on that problem, which hasn't really been the problem that Amazon has been focused on solving. The same applies to last-mile [delivery]: Amazon, in most of their businesses, outsourced last-mile, which works fine if you're moving around single packages. When you want to deliver fresh food to somebody's home in refrigerated vans, it's a whole different setup, that you're going to set up for that purpose.

Amazon will continue to drive their presence in grocery. I'm sure they will become a more and more significant player in grocery. But we're confident that we can actually stay ahead because of that single-minded focus on what is probably the most complex category of retail.

Human labor is much more available and cheap than it has been given the general economic environment. Does that change the equation for automation technologies?

COVID may create a bubble of unemployment, but if you look at the 10 and 20 year trend, we're definitely trending in all major economies towards less and less labor, and with aging populations that's going to continue to be the case.

And within that, the drive of most advanced economies is the drive towards higher education levels, higher value-added jobs, so you get a tighter market in terms of availability of labor. I think there's no doubt that over the medium to long term, the cost of labor continues to go up and the availability of labor continues to get tighter. You don't want to build a business on opportunistic availability of labor because unemployment is going to go up next year. You want to build a business for the long term.