A peep at Ocado's proprietary technology
Our sky-walking reporter Sabrina Schadwinkel jumped on her space buggy for a short hop from Frankfurt to Erith, just a tad south-east of London, to have a look at Ocado's latest and largest so-called Customer Fulfilment Centre. Its two hives are each the size of three football pitches. When fully ramped up, they will have 750,000 storage locations* managed by a swarm of more than 3,500 robots.
Given that Ocado has also just agreed a joint venture with British retail icon Marks & Spencer, let's challenge Mr Jensen on why international retailers should buy all this classy high-tech...
"We will innovate
whatever comes next"
I used to work for Sainsbury's, so I know both sides of the equation. The problem is that most grocers use their existing infrastructure and add online grocery on top.
If you have low bottom-line margins of two to three per cent, pay for staff to pick online orders in your stores, and then deliver these items to customer homes, the cost will be substantially more than any consumer will be willing to pay.
So what are the consequences?
You end up offering a service cheaper than what it costs, and you will lose money. You need ways to radically optimize three things: the web shop, the preparation, i.e., the fulfilment of the order, and the 'last mile' (home delivery).
So what's the Ocado way?
To pick a 50-item order in one of our automated warehouses, we need only roughly seven minutes' labour. Doing that manually in-store would take about an hour. On an order of 100 euros that will save you 15 euros.
We receive our goods directly from the suppliers. There is no upstream supply chain that costs normally 3 to 3.5 per cent of sales. In our case that cost is 0.5 per cent – a big gain of another 3 euros.
Another thing is waste. A typical retailer throws away 2.5 per cent of sales. At Ocado the rate is 0.8 per cent. With three warehouse hubs and 17 spokes it is easier to manage your inventory than stocking hundreds of individual stores.
With regard to expiry dates we use Artificial intelligence to manage products on an in-and-out basis. We can predict if we will have excess stocks or if many products are close to their sell-by date.
We had to – due to the level of complexity of 300,000 orders per week. Basically every time someone places an order we recalculate the distribution algorithm. We use algorithms to manage our fleet and have roughly £1m per van per annum of sales. That is 35 per cent higher – we estimate – than other players that are doing this from stores.
There is also an incredible amount of 'smart' in our web shop. Everything you experience will be personalized to reflect who you are and your past shopping history.
To what degree?
Customers see, for example, different products in a different order and different promotions. We have integrated multiple steps before you check out, which seems counterintuitive at first.
Everyone says that one important rule in e-commerce is to make things as easy as possible to checkout, but groceries are fundamentally different. You buy a large basket of products and not individual things Amazon-style.
We have this very intelligent 'Haven't You Forgotten Page' that knows when you last bought a product and at what kind of frequency you buy the product, which genuinely helps you not to miss out. Our stock ordering and planning are literally coming from this prediction of the sum of our individual customers.
All this data is surely very coveted by suppliers and manufacturers?
Yes, 4 per cent of our sales in the UK are from media income. Take, for instance, samples. In contrast to presenting samples in a physical store, a supplier can target a certain group of people according to their behavior and gather information on rebuy rates.
So, all in all, you argue that it's your technology which makes online grocery profitable. But don't robots also require huge investment
It's a trade-off between operating costs and capital costs. On the one hand, we have the UK retail business, and, on the other hand, the technology part with Ocado Solutions. Three years ago, Ocado UK retail was not only profitable at an ebitda level, but also on an ebit level. So it works. Ocado is still a small player despite our buying alliance with Waitrose.
If you took Tesco's buying power and margins and combined them with Ocado's automation, you'd have the most profitable player in the UK.
Isn't that exactly what you are trying to do with your international partners such as Kroger, Casino, ICA, Sobeys, Bon-Preu?
Exactly. The goal is to partner with leading grocers around the globe that have the local scale to buy effectively and who understand their customers. We know that's a much better model than launching Ocado Retail in the US, France or Germany. We've packaged up everything we learnt from our UK business and provide it to other retailers.
We bring the technology and tools.
And who owns what in the end
Our partners continue to run their own businesses. They own the customer data, they decide what they sell and at what price, and they define the service they offer. We just provide the warehouse technology and equipment.
We guarantee a throughput per hour for the facility. We will have 50 engineers on the ground who make sure the equipment is functioning to full capacity.
Yes, we do. We even have a special recruitment programme with Korger in the US and will end up with several thousand engineers. Currently, we employ 1,400 software developers and 200 hardware and robotics developers.
When can you cash in on partners' deals? Do your partners have to pay up front?
Our partners pay us partly up front, depending largely on the capacity and utilization of the units we have built for them. So we are dependent on the success of our partners. We will make most money, if they use the facility fully.
How are your big partnerships with Kroger and Casino going?
The contract with Kroger states a minimum of 20 Customer Fulfilment Centres, and the first three have been ordered. Now we're planning the next ones. With Casino, we are building a facility outside Paris, and, if it's up-and-running in the first half of 2020, we will replace the existing web shop. We will launch with the Monoprix brand and a second brand from the facility.
What about new deals?
We don't discuss who we're talking to or not talking to. As you know, we are in the business of talking to retailers all around the world about opportunities we may have with them.
We will do something in the German market at some point. It is a more complex market than others. Discounters have a big market share. It is still quite fragmented. However, I think our model is also adapted to the distribution of the population and the medium-sized cities there.
The German market is surely also difficult because it is heavily stored, and shoppers don't have very far to their local store?
That is also the case in the UK. We don't believe that online grocery will replace all shopping in stores. Different people have different needs at different times and want different levels of convenience. If you come home from work and want to buy something for the same evening and a banana for tomorrow, the chances are that e-commerce is not right for you.
However, if you are a family with three kids, you may well want to save two hours of your precious time and have food delivered to your home for only three euros.
Your prestigious Customer Fulfilment Centre in Andover/Hampshire has just gone up in smoke. Are you worried that the blaze could scare away potential retail partners?
Accidents happen. I can't comment further at this point. Investigations are under way. We will analyse what happened and make sure it will never happen again. We will move some capacity and continue to serve the same geography, i.e., 75 per cent of the British population.
Do you plan an expansion of your Customer Fulfilment Centres in the UK?
Yes, we plan ahead for our capacity and therefore plan future CFCs.
Ocado has announced a new one-hour delivery service called Zoom. Is this a direct response to Amazon Fresh?
Why then now?
We want to be the leader in all categories and look at how we can bring automation to the convenience market. We want to see if there is an economical offer for smaller baskets delivered in a short time. The trial starts in March in West London.
There is no doubt that the fear of Amazon is putting more pressure on traditional retailers. Their buying Whole Foods was a clear statement. There have been so many doubts about online grocery.
Take the US as a low-penetration market similar to Germany: Online grocery is now doubling in growth there and it has already reached a market share of 3 per cent. So the market is all very offer-driven.
Why were British consumers such early adopters?
The UK has a high population density, say some. However, Tesco and Sainsbury's started very early, and then Ocado lifted the bar in terms of quality and service. Online has a market share of around 8 per cent in the UK, and 50 per cent of households use it from time to time. People are mixing use, just as they go to both Aldi and Lidl and other supermarket retailers.
What's next in e-food and tech?
We will innovate whatever comes next. We're with voice ordering on Alexa. Picking will become robotized at some point in the future. There is a lot of talk about driverless vehicles. We did a trial last year with a robotics spin-off from Oxford university.
But we haven't yet solved how to get groceries from the delivery van to the customer home: You still need someone to do that. But with the current pressurised labour market, work costs will only go up, while automation costs continue to fall.
How is Ocado preparing for a potential no-deal Brexit?
As you would expect, we are preparing sensibly for a number of scenarios. We have not seen any panic buying yet – bear in mind that fresh and chilled food represents around half of our sales and that these are the sort of products which you cannot stockpile. At present, we've seen no impact to the supply of any products from the EU.
*storage locations in tech-speak are slots within the grid pictured above where the bots move products