Ocado Group plc

Stuart Rose - Non-Executive Chairman Timothy Steiner - CEO & Executive Director Duncan Tatton-Brown - CFO & Executive Director

Xavier Le Mené - Bank of America Merrill Lynch Daniel Ekstein - UBS Investment Bank Stewart McGuire - Crédit Suisse AG Niamh McSherry - Deutsche Bank AG Dusan Milosavljevic - Berenberg Bruno Monteyne - Sanford C. Bernstein & Co.

Robert Joyce - Goldman Sachs Group Inc. Craig Howie - Shore Capital Group Ltd.

Andrew Porteous - HSBC James Tracey - Redburn James Grzinic - Jeffries James Lockyer - Peel Hunt LLP

Slight change in venue this year. This is the presentation of the 2017 year results.

I think I know a lot of your faces in the room, but if you don't know me, I'm Stuart Rose, and I'm the Chairman of Ocado. And before handing over to Tim and Duncan, I just want to make a few remarks this morning about where we are with Ocado.

I don't think it's an exaggeration to say that I think Ocado is now in a very exciting point in its history. Our role as an agent for transformational change in the food retailing industry is now, I think, coming into focus.

And the recent deals that we've done with Groupe Casino and Sobeys, in particular, have validated our platform. And we are confident that we'll be able to do further deals with the momentum for new deals accelerating over time.

We're growing internationally through Ocado Solutions without losing focus on the hundreds of thousands of our retail customers in the U.K. who depend on us for their retail deliveries, grocery deliveries every week.

But we're doing this with industry-leading standards of quality, industry-leading standards on on-time delivery and industry-leading standards in terms of order accuracy. Now the key factor that underpins both our Retail and our Solutions business is our revolutionary and our proprietary technology, which delivers outstanding service to customers and profitable growth to retail partners.

And today's agenda is, of course, about last year. It's about the progress we've made over the last 12 months.

But it's worth reminding ourselves that it's over 16 years now that we've been going with continuous growth. But the focus now really should be to looking forward.

As a company, we have an unprecedented number of options to grow our business. And this board, which I chair, really supports Tim and Duncan and the whole Ocado team in their efforts now to drive this business forward to the exciting opportunities which we think are ahead of us.

We've got a well-resourced and resilient business. And part of this effort, as you will have seen this morning, is that we have the right financial resources in place, hence, the equity raise that we announced this morning.

Now at Ocado, our board focuses on the long-term net present value creation rather than quarterly trading. The long-term opportunity here is clear.

Grocery is the world's largest retail category, and channel shift to online is definitely here to say. It's only just beginning.

And we, I believe, are a critical agent in enabling the disruption that will continue. Ocado provides the market-leading customer proposition in the category, and it does so with a model that is demonstrating sustainable, long-term economics.

If Ocado continues to get the customer experience right, it will continue to take share and it will continue to support grocers, like Casino and Morrisons and Sobeys, build scale and create shareholder value. So I hope that you'll get a real sense of the potential of the business when you hear Tim and Duncan talk in a moment, the potential to add customers and grow market share in the U.K., the potential to apply new and innovative technologies to continue to set the bar in terms of customer experience and operational efficiency, and the potential to sell our know-how to retailers around the globe.

I think that the opportunity is huge. It's matched by our ambition, and the drive and the determination in the Ocado team is now really palpable.

And our objective, without being too grand, is nothing less than to change the way in which the world shops. So with that, I'll hand you over to Tim and to Duncan, who will take you through last year's results.

Thank you.

Thanks, Stuart. So virtuous cycle drives the whole of our business, Retail and Solutions alike.

Investment leads to innovation in our capabilities which benefits our customers. It also gives us greater efficiency, and both drive growth.

At the heart of the virtuous cycle is our technology, unique, mostly proprietary and difficult to replicate. We continually invest in our proposition because investment in the quality and reliability of our proposition builds trust with our customers.

And trust drives sales, which generates better operating metrics, allowing us to reinvest in our proposition. Every year, we focus relentlessly on making Ocado a better place to shop.

Choice is very important to our customers, and here are two practical examples of what better and more choice mean in practice. Firstly, we're offering customers more products to choose from.

At a time when our competitors have been reducing their range in choice, we've extended ours by 150% in the last six years, leaving us substantially the largest choice in U.K. grocery.

Secondly, slot availability. We've increased our slot availability by 25% since 2011 so customers have more flexibility about when they can take delivery of their groceries.

Today, we offer nearly 40% more slots than our closest competitor. And in online grocery, slots is equivalent to stores.

It's your presence. The business has real momentum.

As a result of the investments we've made over time, the business has high momentum today. Look at the charts that illustrate the scale of our market share gains.

As we grow, we're delivering groceries to more and more households. Over the last 15 years, we've moved our penetration with U.K.

households from 10 basis points to 3%. And in every catchment area where we trade, we're still growing strongly, which clearly shows the market is far from mature.

If you look on that right-hand side of that chart, you'll see even in the areas where our penetration is over 10%, we still have double-digit growth. This has translated into a 20% annual increase over the last six years since 2012 in our platform as we lead the channel shift that is today reshaping the U.K.

grocery market. The virtuous cycle is working.

We're leading growth in this channel. And by partnering with other smart, ambitious retailers in other countries, we're helping their market evolve, too.

The recent deals we've done with Groupe Casino and Sobeys have been a validation of our technology and business model, a fact picked up by the markets, the media and our partners themselves, and some quotes here on the screen. And these deals were signed after our new partners undertook very extensive due diligence.

We hosted multiple visits, and they took - they had much consideration before signing into those deals. Looking forward, we think there are three key questions to answer.

Firstly, are customers going to continue to migrate online with all the benefits and convenience that online shopping brings, that they enjoy in their other interactions online? Secondly, do you believe that we have the best platform to meet these customers' needs?

And thirdly, do you think we can make money from this opportunity? We believe that the answer to all three of these questions is, yes, and that's what's giving us the confidence about the scale of our future opportunities.

I'm now going to ask Duncan, invite Duncan to come and talk you through the financial year results in some detail.

Thanks, Tim. Welcome, everyone.

I'll now take you through the backward-looking part of today's presentation and talk you through our results for 2017, which you already have seen from this morning's announcement and which importantly now includes segmental performance. Note that for all these numbers, we're showing performance on a 52-week basis.

So first, revenue. Retail revenue was up 12.4%, and our reported Solutions revenue, up 16.2%.

The Solutions revenue is almost entirely from our services to Morrisons. As for both Groupe Casino and our European retailer, the initial fees that we earned in 2017 are recognized over multiple years and, as such, are modest.

On the EBITDA, overall EBITDA was flat. I will go into this in more details in a moment, but Retail EBITDA was up 4.5%, and Solutions EBITDA declined.

The balance in other, which includes the accounting impact of our MHE JV Co structure and Dordon incentive cost, was £2.4 million. Net interest was up £4 million due to the transaction fees on our refinancing and the interests cost on the £250 million bond issued in June.

Interest cost for 2018 will also grow as we'll have the full year cost of the bond. Depreciation costs were higher as we continue to invest in capacity and innovation, and we started to depreciate Andover.

Loss before tax was £0.5 million. Now let me take you through the segment results in more detail.

Going forward, we plan to provide the revenues, EBITDA, but also admin cost by segment. With this disclosure, you can calculate operating contribution, so I've saved you the effort and included it here.

So first, Retail. Revenues were up 12.4%, with operating contribution up 6.3%.

I'll explain this in more detail over the next couple of slides, but it primarily reflects cost inflation and the impact of Andover CFC. Retail admin cost grew by nearly 10% due to salary increases and new head office and the full year impact of Fabled launch partway through last year - the previous year.

For Solutions, revenue grew over 16%, primarily due to the growth of Morrisons. Operating contribution, which includes the fees that we earn, grew faster, with the benefit of Morrisons growth, some additional fees for providing store pick functionality and a small impact from our international solutions deals.

Solutions' impacts grew much faster, up 50%, due to the further scaling of the Solutions team and as our technology costs ramp ahead of the revenue growth, both of which are likely to continue into 2018. Overall, Solutions EBITDA was down to £2.7 million.

So onto Retail in more detail. Order volumes are up strongly again, up over 14%, with a strong growth in Ocado.com orders and faster growth from our destination sites, particularly Fetch and Fabled.

Total basket size was down due to the stronger growth in the smaller basket destination site orders. But Ocado.com baskets were also down slightly, with a reduction in the number of items per basket, offset by some item price inflation.

Importantly, we were able to cope with this extra order demand as Andover continue to ramp up now to around 20,000 orders per week and as we're able to squeeze more from our mature CFCs. It's noteworthy that Hatfield CFC, which a few years ago was limited to a maximum of 150,000 orders per week, can now process 170,000 orders per week, and this has been achieved with minimal capital investment.

However, despite this progress, we know that our order growth is being limited as some customers do not see enough convenience slots available when they look to book a delivery with us, and so our sales break was constrained. Now not only are our facilities producing more, they continue to improve in efficiency, with mature CFCs up a further 2.4%, with most of the improvement coming from Dordon now regularly above 180 UPH.

Andover is not included in these numbers, has not yet mature, but UPH there also improved during the year. On delivery operations, they were more efficient, too, as we grew orders in the same catchment, so increasing customer density, and we made some improvements to our algorithms.

Note that some of this improvement in drops per van per week came from extending the use of the van with more deliveries on Sundays, which helps this metric but does not drive efficiencies in people or fuel, the two largest costs. Finally, we're proud that we continue to operate with industry-leading levels of waste at 0.7% in total.

It's notable that we're targeting 0.5% waste in our mature facilities and nearly achieved that this year. And note that this only covers our financial cost.

The actual food wasted is a small fraction of this figure. So given these input metrics, what where the trends for profitability in our Retail business?

Note that we're continuing to disclose all the same information here as before, and I've not seen this level of disclosure matched by another retail business. We maintained our existing pricing policy of basket-matching similar products to the market leader and following the same retail prices of our supply partner despite some moves by both retailers.

With the modest inflation in selling prices, with some reduction in promotional activity across the year and a reduction in operational losses, we were able to grow margins slightly. We continue to see good support from our suppliers as we find ways to help them get their products in front of our customers for the benefit of all stakeholders, our suppliers, our customers and ourselves.

And as such, we saw a further improvement in supplier income. On costs, we're currently seeing wage pressures in the U.K., given the impacts of supply and labor following the Brexit vote and with the increases in National Living Wage.

Wage inflation was between 3% and 4% across our fulfillment and delivery operations, so offsetting the UPH and DPW of efficiency gains. The other main factor was the impact of Andover.

Trunking and delivery costs were higher as we used Dordon to supply some of Andover's product range at lower sales levels in order to lower waste and also from the effect of increased Sunday deliveries. But the bigger factor is the additional fixed cost from Andover, including renting rates, which were higher as a percentage of revenue due to the partial utilization.

And we'll cover this again later when we look at the impact of Erith in 2018. Finally, as highlighted earlier in the year, we've increased the size of the engineering team in Andover to build the resiliency of our new solution faster, given the increased demand for it from international solutions partners.

So overall operating contribution fell 60 basis points to 9.5%, still impressive for a subscale grocery business. Admin costs grew slightly slower than sales, with the growth due to inflation, the cost of moving to a new office in Hatfield and the full year impacts of Fabled, our health and beauty business launch last year.

Finally, a word on CapEx. We spent £160 million in 2017, below my guidance of £175 million, with the majority of the reduction from lower investments in new CFCs.

This does not reflect delays in our program with, for example, Erith still planning to open on the same date as we planned a year ago, but only slightly more cautious spend forecasting. And I'll cover 2018 capital expenditure later.

So in summary, we've seen strong growth in revenue across both Retail and Solutions, and both businesses have real momentum. Our platform continues to drive efficiencies, which help offset cost inflation and the fixed cost of new capacity.

And with segmental reporting, now you can see the underlying profitability of our Retail business, and you'll be able to track the performance of our Solutions business as it scales. So with that, back to Tim.

Thanks, Duncan. Let's now focus on the five key elements to the growth story.

This will help illuminate the scale of the opportunity we have for value creation within the business. Firstly, there's the big latent demand, not just here in the U.K., but in every market where the service is available and priced competitively.

So what is the scale of the opportunity? Well, if we look at the U.K., already one of the highest penetrated markets for online grocery globally, the online channel is the fastest-growing segment of U.K.

food retail. Our sales and those of our biggest competitors clearly show that even in the absence of any meaningful top line marketing, the online channel is growing strongly ahead of the market.

The IGD here are forecasting that the online channel will be the fastest-growing segment of the market between now and 2022 with growth of 54%. From 2019, as capacity constraints ease with a growing number of orders being processed in Erith, we will be better able to meet the strong and growing demand for our retail service, underpinning our ability to win market share.

We're confident that we'll continue our strong record of growth because we're gaining market share even in those catchment areas where we've been trading the longest and where our penetration is highest, as we've already shown you. It's also important to note that food lags behind every major retail category in the U.K.

today. But given the chance, customers will go online to order food in the same way they've done with all the other retail categories.

And if food achieves the same degree of online penetration as the other main retail categories, that would imply a market at least 3x bigger than it is today. Ocado is good at attracting and keeping customers, and this is important to us because the lifetime value of our customers is very high.

And here's how the maths work. It costs in the range of £20 to £40 to acquire a new customer.

We don't turn all of them into loyal customers, but those that do typically shop about every fortnight and with an average basket of over £100. So they spend over £2,000 a year, and they make us over £200 a year each.

So even allowing for those customers who don't become loyal, we have a payback of 6 to 12 months. And once we've acquired these customers, they tend to stay loyal, as the chart on the right shows, which is a cohort analysis by year of acquisition of the total spend in every year.

So if you drop back down maybe 10 groups to somewhere like this, this is the customers probably acquired in 2008, their spend remains incredibly static every single year. And that's a function of our existing customers spending more frequency and more basket size, the longer they stay with us, more than compensating for any customers who drop out of the cohort because they move away or, for some other reasons, stop shopping.

Technology innovation drives better operating metrics for us. Every year, we enhance our revolutionary and proprietary technology.

Every year, we get a bigger, competitive advantage. Technology leadership is a secret sauce for us.

We've seen steady improvement over the years. We'll see improvement in the years to come as Andover and Erith mature and as we look at other innovations like robotic picking.

We have the ideas. We'll choose when to introduce them.

And this also applies on the road. For 10 years, we've been using our knowledge, our proprietary algorithms to enable us to operate more efficiently and better than our competitors.

And yes, part of this comes from scale, but we expect to continue to scale as we continue to take market share. But it also comes from more CFCs.

And our new platform, we expect the algorithms on our new platform to give further efficiency gains. And our Retail business is a cash-generating machine.

Look at the numbers on this slide. This shows how much cash we generate when our four CFCs that are open or about to open are at maturity.

We've told you before, Hatfield and Dordon, with their current operating contribution of around 10% after maintenance CapEx, generate £125 million of cash a year. Our share of Erith and Andover together will generate about £1.1 billion of revenue.

We've said our target operating contribution is 11.6% of revenue, therefore, after maintenance CapEx, for those - the share of those we generate £125 million of cash a year. What we take off that?

£75 million to allow for admin costs, higher admin cost than 2017 to allow for the opening of Erith and replacement cost for vehicles. You still got £175 million of cash from those four CFCs to contribute to the platform and to grow new capacity.

It's a wonderful cash-generating machine.

We like to partner with smart, ambitious companies. When our partners grow their business using the Ocado Smart Platform, we grow, too.

It's a win-win situation. Now for commercial sensitivity reasons, we can't tell you the fees our partners pay or our expected returns.

For that, you'll have to see our segmental reporting over time. But this slide shows that whatever the returns, they will get better as we do more business with solutions partners.

As you already know, our fees increase the more capacity utilized per CFC, the more CFCs our partners open and, of course, the more partners we have. Our costs, on the other hand, do not necessarily grow in the same way.

True, we do incur larger incremental build costs when we sign new deals or when we increase the number of CFCs for an existing partner. However, for the remainder of our inputs, we see either modest or no incremental costs as ours Solutions business grows.

Remember, of course, the unique value of our Ocado Smart Platform is that it is the distillation of 18 years of learnings. So what will 2018 and beyond look like?

Let's look at a few points. I talked at the beginning about investment being at the heart of the virtuous cycle driving our business model.

We're not going to stop innovating in 2018. Quite the opposite, we're going to increase our investments so that we can grow faster and achieve higher returns.

Let me talk you through some of the details. We're making continued investment in robotics.

That's not just the next and following generation of the robots sitting on our grid in Andover that you could see in the video outside, but also innovations in robots that will do things like picking in the facilities, frame loading and other activities that will increase their productivity both for us and our partners. We're investing in artificial intelligence and machine learning, again, not just in the front-end using the consumer data, but also in the warehouses to help the robots operate more effectively, to help the layout of stock in the facilities and in many, many other areas.

We're improving our fulfillment development capabilities. We had set out our store before saying that we thought we were capable of doing 2 or 3 deals.

We're working on how we can currently do many more than that, so we're investing money to improve our capabilities to build multiple sheds at the same time for multiple customers. All this means that we need more people, more software engineers writing code, more hardware engineers developing new kit.

We've already increased the number of software engineers four-fold since 2013 when we did our deal with Morrisons. And we will continue building our team in 2018 with the addition of several hundred new colleagues.

Now into 2018, we'll open Erith midway through the year. It's a major step forward for us.

It gives us extra capacity to meet the untapped demand we've got for our business. It's all the same technology as in Andover, but it's just much bigger.

Importantly, it's inside the M25, and we expect it to deliver better UPH at maturity. But you won't see these all visible immediately.

If you look at Dordon, when we opened it, it had higher fixed cost. As you incur the full fixed cost, but you utilize only part of the capability, it had higher waste.

Because you carry all the range, your rate of sale is slower and, therefore the amount of product that goes out of life is higher. It starts initially with a lower UPH because at low volumes and with the new team, you're not as efficient when you launch.

All of those, we saw in Dordon, and all of those improved. We expect the same in Erith.

But in 2018, we think there's an impact not included by many of you in your forecast of around £10 million, but we expect the EBITDA to grow significantly in 2019. And on to CapEx.

We expect to invest and continue to invest at a pace and expect £210 million of CapEx next year. Three major areas to pull out.

First, we want to develop our platform faster, we want to write more software, we want to improve our MHE, shown on here as technology development and fulfillment development, a total of £80 million next year, a big increase on this year. We're also adding new capacity to meet the demand that we have in the U.K., so £70 million on new CFCs, still at Andover and at Erith.

And lastly, included for the first time is a £20 million on the initial spend for CFCs in Paris and the Greater Toronto area for our partners there.

Ocado Solutions' tech platform has been validated internationally. We have an important opportunity to grow the business by signing more Solutions deals.

A typical CFC deal has an expected £30 million peak cash outflow, and so the 5% placing will facilitate signing of several new partnership deals globally. It will also allow investment to step up capacity at both Erith and Andover.

The time is right to accelerate our growth opportunities and drive scale, and this means having the right financial resources and the right people in place to get the job done. In conclusion then, Ocado is already well-known as an iconic and much-loved British retail business.

Our role as a tech innovator and disruptor, however, is now coming into focus as Ocado Solutions begins to scale its customers. The deals with Groupe Casino and Sobeys are a major validation of our business model.

We are in a significant number of meaningful conversations, and we expect to do further deals with the momentum of these new signings building over time. Now is the time to take advantage of our growth opportunities.

We need to invest to ramp up the output of CFC3 and 4 as quickly as possible to meet the growing demand from U.K. customers for our service.

We also need to have the right resources in place to meet demand for the Ocado Solutions offer, and this includes both the right people as well as the necessary working capital. Taking advantage of these opportunities will make our virtuous cycle turn faster.

The opportunity is now. We mustn't squander it.

Our addressable market is huge. It's matched only by our ambition to create value for our shareholders, for our customers and partners and for the communities in which we do business.

Achieving our ambitions will take a combination of focus, energy and skill. We believe that these ingredients are in place at Ocado, and we look forward to the challenges and opportunities of 2018 with confidence and energy.

Happy to take your questions. We've got a microphone, so...

Xavier Le Mené from Bank of America Merrill Lynch. Three questions, if I may.

The first one, just looking at the new digital signing in terms of CapEx, you mentioned actually a peak at £30 million. For instance, you mentioned a £15 million investment for this year for the Casino's deal, for instance.

So can you explain us a bit more the - how it work over time? That's the first one.

And second one is about the features you are bringing or the robotics or these features into the warehouse. What is the life actually for these assets?

What should we assume? And the last one, you take about a huge market opportunity, so how do you see that?

Do you look at the number of countries? Or can you explain a bit more what is your ambition and how you look at that?

Do you want to answer the CapEx, and I'll talk about the next theme?

Yes. So we talked about a £30 million peak outflow.

The peak outflow is between at the end of the second year to the end of the third year. So it's not in the first year.

We typically receive fees from partners at signing and over the expected about two-year build program before the facility opens. And that means that the peak requirement comes later than you might expect.

We were clear to say that's a typical one CFC deal. So if it's a larger CFC, obviously, that number changes.

If it's two CFCs, that number changes. If it is a CFC where the retailer wants the same size ultimately but wants less capacity, that figure comes down.

If it's a CFC where it's the same size and they want more capacity at launch, the number goes up. In all cases, if they're committing to more business that drives us long-term more revenues, the short-term consequence is we need more capital.

Exactly why we talk today, what we're putting in place today is a financing structure, which will enable us to do more of this. More of this is a sign of more income in the future.

So the life of the assets in the new plant, you can probably look at the old plants to give you some idea, we tend to have three different types of asset in a facility with three different kind of lives that we depreciate them over and they last over. The software is usually the shortest-life asset in terms of we can rewrite it over years as we learn more as techniques improve, as processing power improves, and we can get more out of the hardware with new software.

So that probably has a life that's less than 10 years in terms of the software running the robotics. In terms of the things that have got electricity running through them and are moving, there are obviously some certain amounts of wear and tear on them.

And usually, in our facilities, we look for those assets to have a life of around 10 years. If you went into Hatfield today, you would see some assets that have been operating in there for 15 years and will continue operating in there.

But on average, I would envisage that the moving parts would, on average, have a life of around 10 years. And then you've got some fixed stuff in there, so whether that's mezzanines or whether that's grid or stuff like that, that we normally think has a life span that's usually equal to the life of the lease in the building that we're operating in.

So just to quickly add to that. Note that when we have engineers on sites, if they're replacing parts that are broken, we're not capitalizing the new parts.

So the assets that we bought are being maintained without any further capitalization.

So we run them a little bit like you run an airframe effectively. So we preemptively change components in them and run those through the operating cost of spares.

In terms of how do we gauge a huge market, look, we are in talks, as we said, with quite a number of counter-parties. We visit a large number of markets.

We know what we're being asked to investigate price, et cetera. Interesting that I've said to a number of you before that the U.S.

was a market that seem to be the least ambitious. Actually, that appears to have changed quite dramatically.

And I was the keynote speaker, not the weekend we just had, the one before, in Miami, FMI Global, where the kind of theme of the conference was it's not going to be 8% to 10%, it's going to be 20% to 25%. Not our words.

I think those were the consultants that the Board of FMI had brought in. And it definitely seemed to be a theme that have been picked up by a number of the retailers there, which to me was fascinating because in previous visits to the U.S.

over the last few years, retailers have been talking about 1% to 3% moving online. So some of them are starting to see such significant growth from quite nascent offers and quite basic offers that they've put in place.

And so we are seeing interest from all over the world in the solutions, and that's why we believe it's a huge market. We believe that consumers want it.

And everywhere they're being offered it, it's being taken up. And sometimes with not great execution and quite high pricing, people are still seeing quite significant growth.

And so we and they believe that if they can have a better proposition that was economically viable with a token delivery cost that they will see significant market shift.

We go to Stewart - fine, okay. Ownership is everything.

It's Dan Ekstein from UBS. I've got two questions, please.

The first one is on the EBITDA profitability of the Solutions business. You've done about £3 million of EBITDA on primarily the Morrisons capacity, about 90,000 orders per week.

And I know there are some one-off sort of front-loaded admin costs that you highlighted. But I guess adjusting for those, we'd probably be kind of in the mid-single digits of EBITDA.

That relationship between EBITDA and capacity in the Solutions business, should we expect similar sort of run rate from the Sobeys and Casino Smart Platform deals? And the second question is on gross margins.

Your gross margin stabilized last year in the Retail business after three years of quite significant declines. I wonder if you could just talk a bit about the competitive environment there and how you see the outlook as we move into 2018?

Sure. Just on Solutions EBITDA going forward, so what we're now doing is we're showing you the admin cost split between the two businesses.

And I think for those of you who could make it to the segmental discussion we had a couple of weeks ago, what's clearer now is that we have to allocate some costs. And the classic example I always use is, immaterial, is the cost of developing the Ocado app on the Apple Watch.

Does the Retail business incur the cost or the Solutions business? Both businesses will benefit from that because both can use that capability.

So what we did is we allocate technology costs. And clearly, as the Solutions business grows, our existing total technology cost will do whatever they do.

But the allocation proportion to Solutions will grow because they're using more of the platform. So that doesn't, unfortunately, give you the answer, but it gives you sort of a direction of travel.

On the Sobeys and Casino, what we've talked to you about is earnings neutral in the year and the year ahead. What we haven't given you is a specific by contract, by customer of the impact into 2019.

And that, we're not going to do for obvious commercial sensitivity reasons. But what we have said is we expect our Solutions business to grow in profit significantly over the years to come, and that still would be the case or expectation.

Sure. Admin costs as a percentage of sales in the Solutions business, we should expect to move higher?

The admin costs, you should expect to move quite a bit higher. The question that you all want to know, and obviously we can't tell you exactly, is how much sales will grow.

So absolutely, we're investing heavily to enable us to do more deals. The top line will be driven by the number of deals we do.

So no, I'm not saying it's going to go up as a percent of sales because it depends on the pace at which we do it. Clearly, we've got the ambition to invest heavily because we expect that revenue to come, whether it's in 2018, 2019, 2020 or whatever.

We're absolutely expecting it to come.

I'd just remind you on that one that with Morrisons, we actually operate a lot of the business for them, so the pickers, the drivers, the call center are all employed by us and recharged by us with a small margin onto Morrisons. In the international deals, we will be running the software hosting at 24/7 support.

We will be running the on-site engineering for the robotics and the other material handling solutions. But our partners will be running the actual operating staff, the pickers, the dispatchers, the drivers will be run by our partners.

So the numbers, obviously, in terms - will look quite different on that basis. In terms of gross margin, we did see gross margin pressure last year.

We did a very good job on reducing the lines between buying margin and gross margin, and so that obviously helped us. I think, in particular, we largely are price followers, as Duncan explained, and we tend to follow largely two retailers in terms of Tesco is the biggest in the U.K.

market on branded goods; and then on the Waitrose own label, we're a follower of Waitrose pricing. They both try to become, I think, more competitive during the course of the year relative to the market.

But I think we have seen stabilization overall in grocery margins. And I'm not probably - we're not the biggest player, as you know, by a long way in this market, so it's probably best to go and ask both of them.

But we have performed well, I think, relative to their positioning that they've done overall through tight focus on the other lines in that area. Do you want to hand it over this side and let Stewart have a go, and then pass it back to Alex.

Stewart McGuire from Crédit Suisse. A few questions from me.

I'll take a different angle on the Sobeys and Casino transaction. You've historically given us mid-single-digit guidance for fee revenue for third-party transactions.

Does that guidance still apply? Question number one.

Question number two, you talked about range. And on Page 6, you actually showed your range decrease slightly this year.

Can you give us any color as to whether you've reached the optimal ranging? And also, does that have an impact on the ultimate size of your CFCs?

Like how small can your CFC become if you figured out what the range is? And maybe the third question, on your bots within Andover, can you give us any metrics on those?

As costs come down, what's the reliability like? How many do you have operating?

Okay. Tough set of questions here now for me to make sure I don't say what I shouldn't.

Guidance, in terms of mid-single digit as a percentage, I presume you mean of TTV that the client can put through the platform, I think is a reasonable guidance. So yes, range, our range slightly came down, driven by our partner in Dordon, Morrisons taking up some more of their entitlement in that facility.

But we are always churning through our range to try and get the optimum and most attractive offer at any given point in range for our customers. And so we think we did that with minimal impact on the customer offering.

The new infrastructure, Andover, Erith and beyond is very range-friendly. So a significantly smaller warehouse than a Dordon can carry significantly more range than a Dordon can.

And I think at some point, we will move to be able to carry different range in different sites. So at the moment, we're tied.

What we sell in Hatfield, we sell in Dordon, we sell in Andover. And I think we can move that and then if we choose to, to sell even more in any of the new sites because of the capability.

So in Andover, that's something like 250,000 different storage boxes. And in Erith, they'll be like 0.75 million, which gives us great range flexibility.

Your last question was about bot reliability. But we've got just over 500 bots on the grids probably this morning in Andover at a run rate of over 20,000 orders a week.

We've seen significant improvements in bot reliability, bot resilience and uptime on the grids in the last six months, ahead of our planned improvements. And we're really, really excited.

We took the board down there, was it last week? It's going very well.

I could show you the MI screen, but I won't.

Alex Mistrov [ph], KPS Global Management. Just a few questions.

If you were to break out software and hardware piece of your CapEx, how does one think about, going forward, once you develop software for yourself, for Morrisons, Sobeys, et cetera? Margins on the software business, should one think of it as a margin in a similar software businesses around the world, reaching 80% to 90%?

And then on the hardware side, in the last few years, in the last year, I guess, you've placed, what you said, 500 bots on the grid. You now have orders from Sobeys, Casinos and potentially other players, how do you think about cost per bot as you manufacture instead of a few hundred bots but manufacture a few thousand bots going forward?

So the majority of our software is, once developed, it's clearly cheaply replicated for another retailer. Obviously, we are actually hosting it for them as well, so there's some element of cost in the incremental hosting and capacities.

But clearly, if another retailer takes just the same software, it's got a very high margin on it. The first few retailers that we sign up with may have more requirements that we haven't had in the U.K.

Like in the U.K., we have single-format pricing, for example. In France, you have multiple pricing files.

In Germany, for example, you have traceable bottle returns. So there'll be bits of software that we might need to develop for each new customer.

But over time, that will be a decreasing amount for every incremental customer as we've probably encountered that requirement before, effectively. So ultimately, yes, on the software part of our solution, the margins over the medium to long term should become like a software business.

And on the hardware side, one of the attractions and reasons for developing the solution that we have is its modularity and the fact that the differences in each individual site are catered for by very advanced routing algorithms rather than by very complex design decisions and new software, and also that we can achieve manufacturing scale. So yes, we have previously said that Andover will be about 1,100 robots, and Erith, somewhere between 3500 and 4000 robots.

Meaning that just for ourselves and Morrisons, we already had the requirement to build 5,000. Obviously, if we end up building tens of thousands of bots based on significant demand from other people for them as well as our own future growth, then as in any manufacturing process, you'd expect to get some significant economies of scale.

Niamh, you're first then. Quickly.

Niamh McSherry, Deutsche Bank. Just looking at the EBITDA contribution from the Solutions business, which is down year-on-year, and reading that, that includes some upfront sign-on fees, I think that implies that the Morrisons' contribution is down year-on-year.

Can - I guess, can you confirm that, that is the case?

No, absolutely not. Morrisons' fees was up quite substantially.

What is happening in Solutions, as I was saying earlier, we're making decisions to invest across many areas, and the admin cost in Solutions are going up because of that. We didn't see any decline in operating contribution in Solutions.

Operating contribution in Solutions grew at 20% versus revenues growing at 16%. So no, it's not the case.

Okay. So then the new licensing deals were signed on at losses, bigger losses then?

Again, Niamh, no, absolutely not. The contracts that we win aren't loss-making.

We're deciding to invest in our Solutions capabilities, our sales team and in our capabilities generally. And if we write more code and our Solutions business use more of that code, they get more of the admin costs associated with that.

But the code does not - if we sign up a second, a third, a fourth, a fifth, we don't double, quadruple their admin costs because it's just the cost of writing code. If we have one client, if we have 50 clients, you share the cost over the number of clients.

Okay. Can I try this once more, sorry.

If you didn't sign new deals, so there was no sign-on fees, what would that number have done year-on-year?

If we hadn't signed, we'd have seen a very marginal difference in the Solutions EBITDA this year.

Up or down?

Slightly down because we said we recognized a small bit of revenue. We're not making decisions today based on the expectation of signing Sobeys and Casino, and then Luke Jensen is putting its feet up and is going on hold though.

We're making the - we've been quite clear, I think this morning, we're making decisions because we see the momentum of deals building over time. You see what you don't do is say, I'm going to make an investment at some point in the future when I've done a deal, if you are confident about doing the deal.

So we're making investment decisions today because of an expectation of deals in the future.

Okay. Great.

And then just one more, kind of moving away from segment to total Group. I think before you announced these deals and since you've announced these deals, consensus expectations for the Group are down on EBIT and PBT and we're down year-on-year, and it sounds like we'll be down in 2018.

And I guess I was just wondering, obviously, then you mentioned significant growth in 2019, so that would be presumably the virtuous circle kicking in. I just wanted to double check, is that based on the deals that we have signed or not?

And if it is based on the deals that you've signed and if you signed more deals, would that be a reason for profit expectations, Group profit expectations to go down? Or would it be a reason for Group profit expectations to go up?

So we haven't given a formal consensus in terms of a number. I think people have picked up from the message this morning that the consensus estimates for 2018 are too high, and I think that's fair.

I think in terms of what happens if we sign more Solutions deal, I would point you to the consequence of signing the Sobeys deal and the Casino deal, where in both cases, we said the first year after launch would be earnings-neutral. So if we sign a lot of deals at the back-end of this year, I think the effect on 2019 will be immaterial.

There's a completely different decision-making process, which is, if Tim is talking to Luke look at the end of this year and says, I think my expectations for the following year are 2x, 3x, 4x what they were, Tim may say to the board, Can we hire 200, 300, 400 more engineers. So that's a separate matter, which is the pace at which we want to do them.

But we could do deals with the current headcount. We're deciding to accelerate our capabilities so we can do more of them because it's much better to decide to do more of them quickly than keep the same headcount and do few of them.

Okay, so that sounds...

The decision is we have to go faster.

If you sign more deals, then that would be a net positive to profit expectations in 2019, '20?

Again, if we decide in 2019 to invest ahead of the curve because we see the curve getting steeper in the future, not necessarily. But their [indiscernible], the consequence of signing a deal is earnings-neutral.

The consequence of deciding to go faster because your expectations in the future have gone up might mean that you incur more cost. Those are two separate decisions.

If you want us to sign a certain number of deals a year and never accelerate, we won't grow our head office costs. If you want us to sign more deals year after year so we can accelerate and earn more value for shareholders, we might incur more head office costs.

Go ahead.

Dusan from Berenberg. I just have a question on the Retail business actually.

You spoke quite a lot about investing in growth over the period. And generally, would you consider doing what some of your other e-commerce peers have done, which is to flatten margin consensus expectations and reinvesting growth, the efficiencies, the cost efficiencies that you feed through the business?

Look, today, and in fact for the last 16 years we've been trading, we've sold pretty much every slot we've made available to customers. So we're not sitting with 50,000 orders of capacity somewhere that currently is not being used, in which case, that would be a really important question to be asking ourselves.

To date, we've managed with not a significant amount of marketing to fill all the capacity that we've built. Now this year, we've got more capacity to come from further ramping up Andover, from launching and starting to ramp up Erith.

At some point in time, if we were able to ramp up Erith very fast and we found that our existing marketing efforts were not driving enough growth, then that's a question we could ask ourselves. But we're, at the moment, quite a long way from having a challenge of not being able to sell the capacity that we've built.

And just as a follow-up, second half of the year, as Erith capacity ramps up, would you say that you would be least capital growth constrained as you have been in a long time? Is that when we should expect the...

I think that the second half will be easier for us to grow than in the first half. I just want you all to remember that we sold some of the capacity of Erith to Morrisons so that as we start to ramp that site up, they will also be growing their business using some of that capacity.

Bruno from Bernstein. The outlook you've given on growth guidance, the 10% to 15%, is that still largely based on being capacity-constrained?

Is that the main driver of the guidance? The second one is, I think last results, you talked about issues with drivers, picked up a few of other read as well.

Has something structurally change around London or the Southeast where drivers are getting more expensive so it's harder to retain? And the third one is, I think in the last set of results you discussed some resiliencies in Andover.

Is that fully resolved? Can you give any more comments on that?

Yes, if you want.

In terms of the outlook guidance, the 10% to 15% range, 15% is, effectively coming back to the previous question, is the capacity limit in 2018. If the marketing teams said I'd like to spend some more marketing and I can promise you more sales, we'd say, I'm sorry, no, because we won't - we don't expect to have additional capacity.

Now that could change in the future, but in 2018, that's why there's an upper limit. The 10% is just to give a range because we don't want to give a precise forecast.

Look, September is always, year in, year out, a challenging time for drivers because we see exceptional growth coming out of the summer period. I think that this year, it was more challenging than it has been in some previous years.

I think that is a reflection of some of the changes in the labor market on the back of Brexit. We have recruited now as many drivers as we want, so we dealt with it.

I think that hiring drivers has always been something that we spent a lot of time and effort on because our CSTMs, as we refer to the - our colleagues in the business, are the customer face of Ocado. They are the person that our customer interacts within their kitchen, and we pride ourselves on finding the best and training them and enabling to do a great job for our customer.

So we will always strive to find the best, but we do believe we will be able to hire, retain and continue to motivate those people.

Has the markets changed in London there because of all the growth of delivery? Is it materially different?

I don't think in the area that we've been involved in. So I think a lot of the growth is moped drivers with an L plate which is a slightly different role to our employee, benefited employment rights, stable, 40 hours a week, rostered.

We're not part of the L plate moped gig economy. Resilience in Andover has improved significantly beyond our expectations, but it's a never-ending opportunity for us.

So until we get to the point where we have Boeing or Airbus resiliency, it will be an opportunity to improve the throughput capabilities of any given facility. The more facilities we expect to have in the world, the bigger the prize is in achieving it for us, and yet it will always remain the same investment to get there.

And therefore, that's partly what we're talking about when we're talking about trying to go faster in Andover and Erith. And it makes more sense for us to try and do more of that work earlier because we now expect to have more international facilities earlier than we had expected to before.

But I would say we've made very significant improvements, but expect and are challenging ourselves to continue to do so.

Rob Joyce, Goldman Sachs. Three for me.

First one just on the capital raise today. Does the size of that reflect the number of deals you may consider to be imminent?

Or does it reflect the amount of capacity you in the business have, the number of deals you could deal with at the moment? Second one is just, probably hopefully, clear up an earlier question.

In terms of the significant EBITDA improvement in 2019, FY '19, given both Andover and Erith will still be ramping up, can you say what were the major drivers of that - your base of that statement are? And the final one is just to confirm on Slide 21, the cash flows from existing assets, that's for CFCs 1 to 4, excluding the contribution from Morrisons?

So I think I'll do the first one, and I'll leave Duncan 2 and 3. I don't think it is - we're not saying that we're imminently about to do whatever - 140 to 150 or whatever it is divided by 30 in deals tomorrow.

That is not a statement we're attempting to make today. Having said that, I don't want to be sitting in our board or in a board call debating whether we should sign a contract for which we haven't got sufficient financial resources because I think that, that is just a wrong constraint to be operating under when we see an opportunity of the scale that we do.

On the flip side, it is neither a ceiling to the amount of deals that we think we can do or that we hope to do. We think it's just a comfortable or sensible amount to raise now.

If we found ourselves having utilized it, we think that our shareholders will be quite happy if we came back and said, Guys, we've used this in a shorter time frame than you imagined. The business is not yet at a point where we're going to finance these deals in the debt markets, and we'd like a bit more because I think that will be a positive message.

So it just feels like a sensible size, neither too big nor too small, to add in terms of our reserves and capabilities at this point in time.

Rob, on your other two, in terms of 2019, clearly, the issues that the sort of as it were sore tooth of performance when you open new capacity, start reducing as you annualize that. Now we talk about Erith as midyear, I think it's probably fair to say mid-calendar year, which is - but still it's about.

So there is some impact next year. There will still be some impact of Erith into 2019.

But if you think about it, there was more impact in 2018 from Andover. Come 2019, Andover will be, one will presume, getting much closer to maturity.

So the combination of those factors means that as long as the other trends continue, which I think you've seen today about our results, continued growth on the top line; stable, slightly growing margins; improved supplier income; improved operating efficiencies. All the key operating in our metrics have been incredibly consistent over the last few years, and we expect them to continue.

With that continuing, it's quite relatively easy to model that there'll be improvements in EBITDA next year. And last one, yes, on the cash slide.

Yes, that slide said, assume today all 4 CFCs were running at full volumes. So yes, it did assume that we're getting £1.1 billion of revenue across Andover and Erith.

But only our share of them, not the...

So it doesn't include cash flows from the Morrisons?

It doesn't include any cash flows from Morrisons deal in that figure at all. So that's all would be in our Solutions cash flow.

It's Craig from Shore Capital. Just two on the Retail business.

Could you talk about the Ocado Retail promotional participation? And could you just remind us of the supply agreement on Waitrose in terms of the break terms, et cetera, et cetera?

So our promotional penetration is down slightly. I would say in line with the market, but it's not in line with the market.

We're down less than the market. So I think the market in terms of the Big 4 have contracted their promotions at a faster rate than we have.

And your second part of your question was?

Waitrose.

The Waitrose, we were - sorry. 2020, same as we've said before, in terms of the current contract, which if my memory serves me right is the fourth one, having its termination date, yes.

It's Andrew Porteous, HSBC. Three, if I may, just coming back on a couple of earlier point.

When you talk about the £30 million peak cash outflow, is that the cumulative impact? Or in a given - of the year at which there'll be a £30 million cash outflow?

And if that's...

It's cumulative.

Cumulative.

Okay, that's helpful. And then just coming back on a couple of early points around driver wage increases, that's something that happened towards the back end of last year.

Should we expect an impact into this year from that as well?

I think driver wage increases have been happening for two years now. So I think we've been [indiscernible] that the National - the introduction of the National Living Wage, I think, has driven driver wage increases as well as the competition in the market for, from memory now, at least, two years.

So I don't think you're going to see any - it's not a like it's a sudden jump suddenly at the back end of last year. And the majority of our pay rounds occur in our fiscal first half.

Okay. So we should expect some impact into this year?

Just so you understand, that wasn't across Ocado's whole estate we had a wage inflation issue. Apart from the normal ones that we already talked about, what we had is in a couple of sites we found coming back post the summer break when we need to hire more drivers because we have more business because our customers spend less when they're on the beach than they do when they're back here.

So we needed to hire more drivers. We started hiring, found in a couple of specific locations that we weren't able to get them.

Put more effort into it, still weren't able to get them. And so it took us two weeks, three weeks to react to increase salary rates in those two sites and then catch up.

So the capacity shortfall was a temporary blip that effected that quarter and influenced that quarter rather than underlying dynamic, apart from the general dynamic, National Living Wage and that.

That's very helpful. And then just on last one quickly.

Just coming back on the resiliency point, I think you talked about putting more, effectively, labor into the CFCs to help resiliency in your kit. Is that something that as resiliencies improved you've been able to take out?

Or it just requires a bit more, a few more engineers around to make sure you're at those levels?

I mean, without getting into specific numbers, I think we probably have roughly the same number of engineering resource that we had six months ago or 12 months ago, but we probably have a significant increase in the amount of equipment that we're using. So as the equipment is getting more resilient, we're able to add more equipment into the warehouse with the same amount of engineering.

But some of it's in there not to keep it running, but to do the root-cause analysis. So some of the work that we're doing is around what is causing that resiliency issue and what reengineering do we do, what prototyping do we do, what testing do we do in order to drive it out permanently.

So that's really where there's a lot of focus, which is on being able to produce a bot that has significant, incredible statistics in terms of its reliability. And we are launching, imminently, a second-generation production bot, and next year, we'll introduce a third-generation production bot, both of which have been - are in testing, are in subsystem testing and prototyping.

So there's a lot of work going on to produce bots that will be at lower prices, as somebody mentioned before, but also most significantly, lower, long-term operating costs and higher resiliency and reliability for us and our customers.

It's James Tracey from Redburn. A couple of questions for me.

The first one is a technical one. Can you break down the Solutions revenue into the OpEx recharge and fees and also give the EBITDA contribution of MHE JV Co.

Haven't that split up anywhere. And the second question is on - for the Solutions business, could you give an indication, going forward, of operating and capital costs as a percentage of TTV?

So you've given the mid-single-digit guidance of revenues, what about costs? It looks as though operating cost is around 5% of the TTV in '17, if you assume that Morrisons' sales capacity is around £480 million and you divide the admin cost by that?

And the final question is on CapEx, £205 million guidance. Can you split that into how much you think will go toward Solutions and how much will go towards Retail in a total sense?

I'll show you the first one. I'm not sure I picked up the second one, but - and I can do this one.

I might have to ask you to ask the second one again, but I didn't quite hear that. In terms of Morrisons' split of costs in revenue versus fees in revenue, we're no longer providing that split.

If you look at the last year, there were just under £20 million of fees and the balancing costs. Assume this year that just under £20 million is now a bit over £20 million.

So we've earned more in fees from Morrisons because their business has grown and we've provided store-pick functionality that I mentioned. So assume a, let's call it, £mid-20s million fee income, so the balance in operating costs and a quite modest Solutions revenue.

MHE JV Co impact, from memory, was about £11 million. Might have just between £10 million and £11 million.

I'm sure I'll find it somewhere, and I'm pretty sure it's in the full accounts, which were published on the website today. Second question, I didn't hear.

To your second question around Solutions' operating margins, I think I'll just remind you is that the original deal we did at Morrisons that covers Dordon is a different structure to the deals we struck with Morrisons in Erith, and it is - which is more similar to the standard Ocado Smart Platform deals and the deals that we specifically done with Casino and Sobeys. But that's worth remembering that those two retailers are not live yet.

And so you can't read anything across from the numbers that we've presented here in Solutions business for when those deals are in any form of maturity. I think the other thing just to bear in mind is that there's a fixed element of the Solutions business.

There's the development of the product, which scales extraordinarily well as you put more customers on to the platform. Each customer or each facility will have an operating cost in terms of its incremental cloud hosting fees, its incremental engineering spares costs.

But obviously, we wouldn't be in the business if we thought that those fees or those costs are - they were going to be equal to our fees.

Follow-up on my question. Are you able to strip out the sort of elements which are for future growth?

And then talk about sort of like-for-like basis, what you think the operating costs are related to the TTV? And clearly, you also capitalized a significant proportion of costs, so can you say what the split is between going-forward capitalizing cost and expense and cost for these deals?

I mean I think we're giving a lot of disclosure, and I don't think we're going to give - to open up our entire finance department. I think what I'm trying to do is give you a flavor that these deals will have an operating cost, the future deals.

It will be around incremental cloud hosting. It will be around incremental engineering costs, and the rest of it will flow down to the EBITDA line.

There will be at the center an ongoing cost for ongoing development in the platform to ensure that it is not just the best platform today in 2018, but it's the best platform in 10 years' time in 2028. So we will be making ongoing investments in the software part of the technology and also into the robotics and MHE part of the technologies, and then we'll have a sales effort.

So those will be the central costs, of which the operating margins from each of the deals will contribute towards. Clearly, as we scale the business up, they'll become a smaller and smaller percentage of the total contributions.

James, on your CapEx figure, obviously, including CapEx is only assets, well, I mean capitalized assets, we wouldn't capitalize, hosting costs is not asset, it's a P&L cost, so any assets get capitalized. Of our £210 million in the year ahead, about £100 million is for U.K.

across mature CFCs, new CFCs and delivery vehicles. £20 million is for Solutions because that's the line in there for Solutions.

£80 million is for Retail and Solutions because all the investments in our platform benefit the Retail business and benefit our Solutions business because they're both intended, do and are intended to use the same platform. And there's a £10 million other, some of that's in Retail and some of that I'm not going to tell you about because that's the next thing we might be talking about at some point.

James Grzinic from Jeffries. I had two very quick ones.

The first is can you just say what Andover - the Andover dilution was last year relative to the £10 million we were going to get from Erith this year and how that flips this year? And secondly, I think I heard you saying you think Ocado Retail still subscale business in the U.K.

What does a scaled-up Ocado U.K. business look like?

At what point do you reach scale?

I'll let you talk about Andover dilution.

Yes. So call it half a year for Erith, and it's effectively - our proportion of it is effectively 2 to 3x of the scale, so just do the maths back from there and you'll get to somewhere between £3 million and £5 million depending on how long.

So that's a reasonable figure for Andover. The subscale comment was my comment, and effectively just saying that it's quite difficult buying Coke from Coca-Cola when Tesco is buying £40 billion equivalents worth and we're buying £1.5 billion, and we're selling at the same price as them.

So we shouldn't really be in business because we have that disadvantage, yet you can see the economics of our Retail business, which I think are pretty attractive. So that's that comment.

We shouldn't really be here because nobody would say, globally, try and run a grocery business at £1.5 billion worth of revenue when you're competing with people who've got £40 billion worth of revenue. So you're crazy, and I think that's the point.

So over time, as we grow in scale, we can take out buying power, we can take out the fees that we're paying to our partners. As we increase scale, we'll improve the density between our customers, we'll build more facilities, we'll be closer to our customers, we'll have less trunking, we'll gain economies of scale in the manufacturing of our plants, we'll obviously be able to leverage the central costs over much larger sales base.

So there's a lot of feature advantage to growing - continuing to grow the Retail business.

Sorry, can I just clarify something, Duncan, is that the 3 to 5 from Andover this year?

I think you asked the last year.

Yes, last year. And how does that flex this year, the 3 to 5 becomes?

Hardly anything. Hardly significant at all because you've annualized the vast majority of it.

It's James Lockyer from Peel Hunt. Just two questions from me, please.

Your supply income has improved this year, and you said it was because you put more products in front of customers. Previously, you spoke about how personalization and AI has helped drive this.

Now I was wondering if you could talk about how this is progressing, and how the customer experience is improving? And obviously, given that you're using all this data, how does GDPR impact you?

And then secondly, given in the new CFCs you're fulfilling orders in sort of 5 to 10 minutes, can you talk about how your same-day capacity potentially will increase or change over time? And obviously, once Erith opens in London, what do you think is the same-day potential in London?

Okay. So you're touching on the right things, and as customers are shopping on smaller interfaces and more than 50% of our business, I think 60% today is going across mobile, then our influence as a retailer is greater than it is on a computer screen, which is already greater than it is on a store.

And therefore, it's important for the supply base to work with us as the retailer to make sure that if they've got the right product for a customer and they get that product in front of the customer at the right point in time, particularly in instances where there are two products that satisfies the customers' demand, one from them and one from their competitor. And so that's why you see that growing income line.

As we introduce more and more functionality, that makes it easier for our customers to shop, then our importance continues to grow. So if we're predicting baskets for customers in their totality or giving customers the ingredients to a recipe as part of the meal planning exercise, obviously the choice of olive oil or the choice of butter or the choice of salt is something that our software will suggest one to those customers.

And therefore, if the customer we feel is indifferent between two suppliers' product, it's up to us to choose which one of those two, and therefore, suppliers want to work very closely with us. So I think it is an area where as more intelligence is added into the user interface is it's more important for the suppliers to work more closely with us.

I'm not an expert on GDPR, so I probably not...

Obviously, we have a plan in place to ensure we comply. From a customer perspective, we don't think it will be a material impact.

Customers like to shop their favorites. If they can't shop their favorites, they'll be unhappy.

To shop favorites, we've got a whole data around their previous shopping history, so we don't expect it to be a problem.

There was a last part of the question, just remind me what is was?

Same day.

Same day. So yes, obviously, as our new facilities are able to process orders in a fraction of the time that the historical facilities are, it significantly increases the possibility of doing more same-day orders.

In addition to which, as our scale grows, enabling us and requiring us to open more facilities, then we have more facilities closer to more of our customers. So if you just decided that, for example, you would do same day in 1.5 hour driving radius from a warehouse, if you went back six years, that was 1.5 hours driving radius from Hatfield.

By the end of this year, that's a 1.5 hours driving radius from Hatfield, from Dordon, from Andover and from Erith. So a combination of more warehouses and the shorter processing times will significantly increase our capability of doing more same-day business.

I'm conscious of the time. I'm going to suggest actually to finish it there.

And actually, nobody's hand is up, so it's probably perfect timing.

Thank you very much.

Thank you.