Earnings Transcript for OCDO.L - Q4 Fiscal Year 2019

Stuart Rose: Good morning, everybody. Welcome to the meeting this morning. I was only just reflecting last night on what words I might use to describe the year that Ocado had last year. And without over-egging it, I came up with a few

Timothy Steiner: Thanks, Stuart. We're pleased to report results, which show strong momentum in the business. Although the statutory results reflect the culmination of factors, including the impact of the Andover fire, which as Stuart just mentioned, I had a call before I stood up here last year that said, We think there may actually be a fire in the building. We might need to put some water over it. It's been a busy year. The underlying performance of Ocado Retail and the successful growth of Ocado Solutions are both very encouraging. Our progress over the last 12 months, which includes signing our eighth and ninth Solution clients, Coles in Australia and Aeon in Japan, and successfully maintaining the strong growth post-Andover has demonstrated many of the Ocado Group's most important characteristics

Duncan Tatton-Brown: Thanks, Tim. Good morning, everyone. We've had a busy year, and we have a complicated set of results to talk you through this morning. You've seen the numbers already, so I'll be brief. But I'll try and help you understand the underlying performance. The key headlines are that the business is operating as expected with our joint venture with M&S well established, a strong and resilient performance of Ocado Retail after the Andover fire and a great deal of progress in Ocado Solutions with significant investments for our clients, in our platform and in our capabilities. Now just a couple of points on the presentation of the numbers. First, on this slide, we're showing the statutory results before the impact of exceptionals. Also, 2019 numbers include the impact of IFRS 16. You already have the impacts for 2018, which we shared at the accounting seminar a few weeks ago. Note that we indicated that the impact would likely all be in UK Solutions & Logistics segment. But because Ocado Retail has rights of use on properties still leased by Ocado Group, the impact is actually in both segments. The balance of the presentation will exclude the IFRS 16 impact, where relevant, from 2019 numbers to make them comparable. So the headlines. Group revenues were up 9.9%. Group EBITDA, which includes a £25.4 million benefit from IFRS 16, was £43.3 million, and loss before tax was £120.4 million after a £9 million IFRS 16 impact. As reported in the first half results, there's been some significant exceptional impacts from the fire in Andover with the write-off of the property, the stock and from additional operating costs. We received £74 million in 2019 from our insurers. We recognize this as exceptional income when we incur the capital expenditure for the rebuild or as business interruption costs are incurred. Up until the end of the year, we'd recognized nearly £24 million. The other notable exceptional cost is £1.3 million of legal cost as a result of the theft of our intellectual property. We will vigorously protect our IP and challenge anyone who uses illegally obtained information directly or indirectly. We've made a claim against the company and a number of individuals. We strongly believe in the merits of our case. So back on to performance. Here are the results in our new segments. Firstly, revenue. Retail revenue grew 10.3%. UK Solutions & Logistics grew slower, up 7.8%. The slower growth comes from the part-year impact of Morrisons holiday, where we agreed with them to take back Erith capacity to allow us to continue to grow. Reported International Solutions revenue was flat, although this doesn't tell you the true story or the full story. More on that in a moment. Other includes Fabled, our health and beauty business that we sold during the year. Looking at the EBITDA variances pre-IFRS 16 impact. Retail EBITDA fell to £20.2 million with the declines largely due to the impacts of Andover on reported performance. As you'll see shortly, the key operational metrics continue to move in the right direction. UK Solutions & Logistics grew 10% due to some small admin cost efficiencies. International Solutions EBITDA fell as we continue to invest in the platform and incurred increased costs ahead of the launch of our first overseas CFCs. Within Other, share scheme costs were up, with the costs were up, with the cost of schemes reflecting the strong growth in value over the last few years, offset in part by research and development tax rebates. Also included, although not material, were some trading losses in the ventures we consolidate, at this point only Jones Food. Our venture investments are all at an early stage in their development, and there's much to do before we can be confident of any significant value. But there are encouraging signs. And to date, around £140 million of fees have not been recognized in revenue. We expect this strong growth in fees to continue with capacity commitments from our partners. Though, as you know, it will take a few years before all the fees we receive can be fully recognized in profits. At the same time, costs will also grow as we make the necessary investments to deliver this future growth. As a reminder, the UK and Solutions Logistics segment includes the services that we provide to Ocado Retail and Morrisons. This includes both a standard solutions offer with the hardware and software to operate using our platform plus logistics services, so the actual operational roles of picking the product and delivering to customers. Revenue was up 7.8%, lower than would have been the case, save for the Morrisons holiday, which resulted in lower fee income, partly due to some discounted fees, where there are further effect to come next year. We continue to be more efficient for both our partners. Delivery efficiency was better with DPV up to 196 and mature CFCs UPH up to 168. It's worth noting that Erith UPH is now regularly over Hatfield. Costs for this segment grew due to a full year of fixed cost at Erith. Engineering costs per order at Erith are higher than in mature CFCs. But importantly, they fell by 1/3 in the year. This is prior to the benefits we expect from our new-generation robot, which we used in Erith and all future CFCs. Now to help understand the retail performance, it's worth picking out some of the key capacity points. In the year, active customers grew at approximately the same rate as revenues and order volumes. We highlight big basket order volumes here, as this excludes Ocado Zoom. Tim will touch on this later, but our first Zoom site in Acton performed ahead of expectations. I am one of many active customers. The order growth was constrained by lack of capacity, hopefully, well-illustrated by the middle graph. We've lost the Andover capacity we had, here shown as year-end numbers. Although this was much, this was higher up by the time of the fire and clearly which we'd expected to be much higher by the end of the year. We continue to grow because Erith's capacity grew, ending the year at over 70,000 orders per week. And it's also notable here how much extra capacity we could squeeze from our existing mature CFCs. Not only did our CFCs grow capacity, they became more efficient, we still have more to do to get to our target of over 200 UPH with the opportunity from further efficiencies from robotic picking to come on top. Now I want to take a moment to talk a bit more about wastage. This is an important issue for grocery retailers, and it's also an important issue for the environment. It's commonly thought that most grocers have a wastage figure in the range of 2% to 3% due to some of the inherent challenges of forecasting customer demand accurately across a large portfolio of stores, not being able to manage stock rotation perfectly and due to the length of supply chains. We have, for many years, operated at much better levels using the benefits of our model, notably bringing stock into a larger location benefiting from scale advantages that come from that. We've also used our platform benefits, better predictive capabilities, a perfect first-in, first-out system and strong visibility of customer behavior. That let us operate at 0.8% waste in 2018. But we've worked and continue to work hard to improve. And with the use of AI and machine learning, we've been able to fine-tune our forecasting even further, so helping us to reduce to 0.4% this year without any significant impact on product availability. Note that even then, only 0.02% of our food ends in landfill. So on to Retail performance. This is now presenting our Ocado Retail as an Ocado Solutions customer. There are no engineering costs for MHE, no technology costs for hosting and supporting our platform and, of course, none of the capital costs for these. These are replaced by fees for both our platform, but also management fees for logistics services and for historic capital investments. Gross margin was down slightly year-on-year with supplier income flat. As I highlighted at the half year, we're not expecting any gains in the second half as the market remains competitive. Trunking and delivery costs were lower due to the improved efficiency, offsetting the wage cost and pressures. Operational efficiency was better, up in mature CFCs and in Erith, but we have higher fixed costs in Erith and some small impact from our Ocado Zoom site. Marketing costs were up as we invested more heavily in customer acquisition after the slower growth period post-Andover fire and from trialing an offline campaign in one of our regions. Fees here are charged by Ocado Group with the income recorded in UK Solutions & Logistics segment. It's worth noting that within the total fees, the OSP fees are lower than for a typical partner, reflecting a discount as the majority of the capacity is coming from less sufficient assets. Admin costs grew in line with revenue. Overall EBITDA was, therefore, down 80 basis points. We expect EBITDA growth ahead of revenue growth in 2020 as Erith sales further. I thought I'd briefly touch on the economics for our partners. So Ocado Retail has some characteristics that make it different from our partners. First, Ocado Retail currently sources under the Waitrose agreement, incurring a fee, but also not benefiting from their buying scale that most of our partners can. And as I mentioned, the legacy assets in the UK don't have the efficiency that we'd expect from our new facilities. Marketing costs were higher because of the Andover fire, noting that any insurance income will be recognized in exceptions. Here again, size matters, leveraging a well-known brand across a larger pool of customers brings better marketing efficiencies. I've already covered fees, but note that our admin costs are higher, one other line that benefits from size. Overall, once at scale, we expect our partners to have a higher EBITDA margin and low capital costs. Now cash flows. We started the year with £411 million of cash. Working capital was positive with significant benefit from fees invoiced to our Solutions partners. This was partly offset by a temporary decline at Ocado Retail, partly due to faster payment to suppliers since becoming a designated retailer under GSCOP. We received £74 million of cash from our insurers, which will help fund the rebuild of Andover and other business interruption losses. We also cash-settled our growth incentive scheme, although this was largely offset by earlier treasury share sales. The largest part of the Other is finance costs, including the interest element of IFRS 16. So overall operating cash flow was £48 million. Given the initial proceeds from the M&S deal, we had ample funds to continue investing, incurring £260 million of CapEx in the year. You can see that this was spread across CFC growth in both UK and for our international plants here, mostly Casino and Sobeys. The biggest investments in the year were on continuing to develop our platform, growing our technology teams by over third. We ended the year with £751 million of cash, and we increased our headroom with a convertible issued in early December. We now have £1.3 billion of cash and over £1.4 billion of headroom to continue to support the growth of our clients' businesses in UK and overseas. So here, on to outlook. I just wanted to talk this through and give you an indication by segment. So assuming normal market conditions, we'd expect our Retail revenues to grow between 10% and 15%, with the upper end dependent on delivering sustained, fast ramp-ups in Erith. We will stay constrained until we can open more capacity expected in early 2021 with the opening of a Bristol mini-CFC with that project on track. UK Solutions & Logistics should grow at a slower rate than Retail as we absorb the full year impact of the Morrisons holiday. They're due to start to return to Erith in February 2021. For International Solutions, as you know, we're only starting -- we only start recognizing fees as revenue when operations commence. We expect less than £10 million with the quantum defined primarily by the opening date for Casino and Sobeys. As you'll hear from Tim, both projects are progressing well. A more important indicator is that we expect International Solutions fees invoiced to grow by over 40%. This would mean we've had nearly £0.25 billion of unrecognized revenue by the end of the year. For EBITDA, Retail EBITDA should grow above revenue growth as Erith scales. The UK Solutions & Logistics, as I mentioned earlier, there will be the full year impact of the Morrisons holiday, which will result in a decline in EBITDA. The corresponding offset from insurance will be in exceptions. International Solutions EBITDA is expected to decline. There are a number of reasons for this. First, there's minimal reported revenue despite strong cash fees. Second, we're growing the size of the technology team even further to ensure that we developed quickly all the features that our partners need. And finally, as I indicated at the seminar a few weeks ago, for the first CFC openings, we plan to have higher levels of resource than we would in the future to ensure success for our clients and to help take the learnings for the significant program of openings that we have ahead. On cash guidance, we expect continued cash receipts to cover Andover rebuild and business interruption costs, recognizing these, when appropriate, as exceptional income. CapEx guidance is for £600 million. The majority of this is on the extensive program of CFC openings in the UK and internationally, remember with Andover funded by insurance. We expect another £150 million on further developments of the platform and on other projects. With our organic cash flows, the prepayment of client fees and our substantial headroom, we're well set to grow our business to meet the demands of our clients. So in summary, 2019 has been a busy and eventful year. There's much more to do and many further opportunities ahead. But I'll now hand you over to Tim to explain more.

Timothy Steiner: Thanks, Duncan. So we're building for tomorrow at a pace. So what you can see here is the 2 facilities that we're going to open in this first half. At the top, just outside Paris for Groupe Casino; underneath, just outside Toronto for Sobeys. So these are real. They're going to go live soon. We are developing, we've got 4 key elements

Q - Andrew Gwynn: So it's Andrew Gwynn from Exane. I'm going to, I've got lots of questions, actually, but I'll just go for three of them. So apologies. On the Zoom, obviously, I think the technology in there is relatively basic at the moment. So I'm just wondering what the plans are going forward? Should we expect the sort of high type model with robots on top? Second one, just to help me understand a little bit, a bit, Tim. You're capacity constrained, but the marketing cost stepped up in the retail business. So I'm just wondering what that was about. And we did touch on it very briefly. And then EBITDA, sort of 2 stage question. So firstly, is it the right way to measure the business, given partly what losses elsewhere in the P&L and we'll see IFRS 16? But second, actually, I think you've got to tie it that consensus on your website. So if you can confirm the numbers because I forgot the number, Tim. We'll see if you're happy broadly where that's coming at.

Timothy Steiner: Let me start on the first 2. You're right. So the first Zoom site, as we said, was to test the customer proposition more than anything else. And what it's shown us is great demand. And what it's shown us is the basket sizes are, the basket sizes is 25% plus larger than we initially anticipated. And it's not that we don't like each other. But, and it's half automated, and it's not automated with our equipment and therefore, the operating costs in it are not wonderfully attractive at the moment, and it's not a sustained, long-term sustainable site without some significant changes. The Zoom site that we're looking at building now in London will share exactly the same automation as all of our other automated sites. It will have full automation in the Chelan and the Ambient, and it will operate at materially better levels that will lead us to a profitable model that we expect to be able to operate long term. Okay. And we're ready to do that now. The second one was on the marketing cost. There were 2 parts to that. The first part is that one of the very first things that we did when we realized that, that site was not coming back live in a very short period of time was to turn off our marketing. The problem is, is there's an element of momentum to your marketing where if you turn it off for a period of time and then turn it back on again, you've lost the momentum, and you have to overinvest to try and build it back up again. The second part of that is you'll have noticed from Duncan's slides that some of the loss in the Andover volume, the gray part of the box that we lost was taken up not just by the growth in Erith but also by growth in Hatfield and Dordon. And some of that growth has come through further flattening of the weep through driving volumes through the days when the kit is less constrained. So we're driving significantly higher growth on a Tuesday, Wednesday and Thursday then we're able to drive on a Friday and Saturday. And so to push the customer demand into those days, above and beyond where we were before is more expensive in marketing terms then if we'd maintain the shape of week that we had prior to the fire. Okay. And I'll let him have the hard question.

Duncan Tatton-Brown: Just slightly to add to what Tim said there. That extra cost of marketing, where we're doing that to ensure that we maintain the customers we otherwise would have lost, that's in the insurance company's interest because it lowers the losses that have to pay for us. So some of this extra marketing we'll actually get back. But, of course, we'll never report it back in marketing. It will get reported in exceptional income. So it's not a cash loss to us. So we're taking the decisions on the cash basis, not on reported EBITDA basis. Your question was is EBITDA the right measure. To a certain extent, I've given you part of the answer there. EBITDA, certainly in the short-term, isn't a great measure because a lot of the things that we're doing might hit EBITDA, and we'll get paid for it from insurance. But accounting says the rules will follow the rules and the rules, therefore, reflect in exceptional income. Obviously, at this stage, there is a question about how valuable EBITDA measures on International Solutions growth business when you're not reflecting it. But I think it's the best measure we can use today. And that's partly why we've given you increased segmental disclosure because you can look at the different parts of the business and consider the best way to understand the value that's there on those different segments. And I think the retail business EBITDA is a good measure. On the Solutions business, ultimately, it's a good part of that measure, obviously, to get a better element of the capital behind it. On the international business, at the moment, it's pre reporting revenue, so it's not a good measure. Consensus, I think the website is showing, on a pre IFRS 16 basis about, £8 million for the year, which -- so that is based on -- coming in here this morning, that was the right, I think, consensus. Clearly, we've made some indications here today. I think there's a bit of work to do our models, not least because it's a complicated set of results. So you have my sympathy. But that's why we've given you lots of information to help you work on it.

Robert Joyce: Rob Joyce, Goldman Sachs. I'll take three as well, if you don't mind. So first one, on the retail guidance for the year, in which journey you're assuming when you shift from Waitrose to M&S within that 10% to 15%. Second one is just, you're now doing over 70,000 orders a week at Erith. Does that confirm the sites of that size, i.e. the Andover site you plan to roll out? You now have confidence that they will operate the economics you expected them to? Final one, just in terms of the immediacy market you've talked about. I think you said global grocery is around £5 trillion before. How big do you see that immediacy market as being? And within that model, how are you doing the last mile? Are you using contracted riders or fully employed riders? How is that working? Thank you.

Timothy Steiner: Look, the data suggests that it's a low single-digit percentage of customers who are passionate about Waitrose. We see a larger number of that -- of people who are passionate about M&S and don't shop with us today. So we expect there will be some churn, but we expect there will be a replacement from new customers who are more passionate about the M&S brand than they were about Waitrose. But we are budgeting in there more marketing costs just to make sure everyone understands what's going on around that time period overall. In terms of Erith. So yes, it's very significant. Erith is operating over 70,000 orders because that's roughly the scale of the average sites that we're building for our customers, which is useful in understanding that there aren't parts of the routing algorithms that can't handle that scale or the communications infrastructure that can't handle that scale, et cetera. So it's massively reassuring to us and our clients to be able to see Erith operate at those numbers during its growth to know that there aren't new hurdles to overcome to achieve full capacity in their sites. On the economics front, Erith, as we mentioned, is already operating at more efficiency than Hatfield. It's not yet operating at the efficiency that we expect it to get to, so there's ongoing work there. It's moving in the right direction kind of almost week on week. And also, on our economics of providing the site, Duncan mentioned that engineering costs per order were down a third in the year. But we need them to continue coming down, and that is our expectation, and it's something that we're very clear on, and it's been a successful journey in cost reduction so far. And we have the new generation robot coming out this year that we expect in the medium to long-term to have materially lower engineering costs. So everything is moving in the right direction, 70,000, big and important tick for us. I don't know that global stats on what percentage immediacy or what you might define as convenience in the bricks-and-mortar world is. I can look at, it's a material but minority percentage of our sales, where we have it for sale, available at the moment, but it is rapidly growing. I don't expect it to be as big as the opportunity from big sites. And obviously, there's the challenge of it is really only available where density exists. So you're not going to be able to offer it in all the geography that you could offer same-day and next day services in. Currently, we do the last mile working with an outside. A third-party does the deliveries for us. It's something that we could bring in-house at some point. We could use a third party. Most of them are delivered by moped. Some are delivered in the car because they're larger than the standard moped's capacity. It's something that we want to continue to work on to optimize. It's again one of the reasons why we wanted to get a solution live prior to deploying it with our full infrastructure because there's lots of learnings to actually undertake to drive down the operational cost models to make it work.

Robert Joyce: I mean that seems very quick. Are you paying them per drop? Or are you paying a sort of hourly rate per container?

Timothy Steiner: We pay per drop but a different amount per drop based on the size of the order and the distance from the site.

Bruno Monteyne: Bruno Monteyne from Bernstein. A few from me. In the CapEx numbers, you talk about £104 million for platform development versus £71 million international development. I'm a little bit struggling what you call platform development versus international. So clearly, a robotic arm would be platform development. But if you have to make it earthquake proof or the Baguette in France versus a really, where does it sit? My second question is, I noticed you had, like, for Andover 40% of the £225 million CapEx in the UK that would suggest it cost £80 million. But though the old numbers was that Andover was more like a £40 million, £45 million. Has your CapEx gone towards more expensive? My next question is around the admin costs, which you have well over 3%. What's realistic for a point, given what you know is in your numbers and the fact you have to run a commercial overhead, is below 1% realistic for a partner? And then last but not least, clearly, your supplier income is a very big part of your boost to the P&L, 4%. Now how realistically can a retail partner assume, let's say, Kroger, that they're going to get an additional 4%? And are you helping them with the support, they call it the milking of the suppliers' process, of getting there?

Timothy Steiner: If you want to?

Duncan Tatton-Brown: Yes. Maybe I'll do the first two. Platform development versus international rollout. So people involved in designing a change to the platform, and you, I think you mentioned seismic in there. The cost of developing a new seismic solutions in platform development, the cost of installing a seismic grid in a particular facility is in the CFC. So it's are you writing software for use in multiple CFCs? Or are you doing something specific to a CFC? That's the differentiator. £104 million, right, for this year. Expect that to grow next year, part of the overall £150 million. So assume £120 million, £125 million, that sort of thing. Andover's numbers, £40 million to £50 million, was right in terms of your recollections for the mechanical handling equipment in Andover. But remember, the building burned down. So we need to rebuild the building as well. So, and that's from clearing the land through to a construction, a complete new building and all the facilities that go with that. So that project is, in total, about a £85 million project. So it doesn't reflect the MHE costs. It's the type of build cost. Of course, that's all funded by the insurer. And as a point of detail maybe not understood in that £250 million is actually some build costs for the Bristol mini CFC, which Ocado will fund and recover from Ocado Retail through a capital charge because the reason why we're building the first of many CFC, apart from the strategic reasons, is because we need capacity quickly. So it's effectively part of the insurance claim.

Timothy Steiner: So moving on to your other two. I mean, look, I think you can look at what global retailers spend on admin costs. I would have thought they were doing well if it's under 1%. I would have thought most of them were in the 1% to 1.5% type of number, but generally not operating at 3.5%. But again, for you to look at. The supplier income, I think it's fair to say that in a multitude of areas, our clients have not signed up to work with us and are not intending to work with us by saying, thanks very much for this software. Thanks very much for those robots. I'm glad you're here to service them. Now we're going to go and run our business. There's a lot of knowledge transfer going on. So there's a lot of questions being asked not just about 4% supply income and how we raise it. Bear in mind, if you look to the top 10 payers of about 4%, I think there are probably suppliers to every one of our clients globally. So the tools are there for them to earn that money as well. And we will support them in understanding how to do that, where to get it from, who to get it from and everything else. But they're looking for our help in understanding how we digitally engage with customers, how we market, all the different tricks, everything that we tried over the last 19 years, they want to learn from rather than spend the next 5 years trying to do that themselves. So there's an enormous amount of just kind of knowledge transfer, and it's not just about software and technology.

Bruno Monteyne: And maybe just one final follow-up question. You talked a lot about increasing the velocity of the business. I think the way I'd sort of tried to measure the velocity is how many CFCs per year can you annually deploy? Have you increased it from 10% to 15%, from 7% to 10%? How would you quantify your velocity you're cruising at?

Timothy Steiner: Look, I would just say that we have increased our velocity such that we're comfortable to fulfill all the obligations that we've taken on from our existing clients, and we're still allowing Luke and his team to run around and talk to more potential clients. And if he comes back with the right one on the right terms and we want to sign another deal, we believe we've got the capability of doing that. So it's difficult to say exactly how many sheds could I start tomorrow versus how many sheds could I start next year. We're growing that velocity so that we can handle more clients, more sheds per year from the existing clients, more than we've, both what we've currently committed to and more is what we think that we can handle.

Nick Coulter: Nick Coulter from Citi.

Timothy Steiner: We'll try it under 5 questions.

Nick Coulter: I think, really, Bruno actually numbered them. He just rattled them off. I'll go for 3. I'll just do like a, firstly, on the mini warehouse or Bristol Warehouse, could you give a little bit more detail around the economics relative to either Erith or Andover? And are those economics partly contingent on working in tandem with a larger warehouse? Secondly, on the generations of robots. Again, where are you with that with respect to the generation going into the French and Canadian warehouses? What's the development beyond that generation as you continue on that journey to lower engineering costs? I mean, what are you testing on your grids effectively? And then lastly, if I may, kind of an ESG question on your bands, what are you doing with respect to your band replenishment, how you're moving towards natural gas or electric over time? I remember that was a project way back, so I'll assume it's still on the radar. Thank you.

Timothy Steiner: Okay. So the mini in Bristol as an operating site will initially have slightly higher costs than an Erith or an Andover would have done. But there are moves afoot to do something along the lines that you just kind of suggest that will bring it down to much closer to being in line over time to do some unique ways of working together. As I say, there's commercial sensitivity around them. So I don't want to explain it all. But in the future. But yes, initially, slightly higher, and then at some point, not dissimilar.

Duncan Tatton-Brown: Tim, sorry, just to add in there, and you've got the benefit of not necessarily needing spokes, spoke CapEx and spoke operation cost. So 10 months, Tim, I think.

Timothy Steiner: So it will be almost 100% direct deliveries from that site. Your next question was Generation robots. Thank you. So the first two sites in France and Canada will have a combination of the robots that are currently deployed in Erith. That's what's on their grids now testing and ready for go live. As they scale those sites, they'll have second -- third generation robots that will be arriving during the course of this year. Kroger, one, should go live, I think, with probably 100% new robots. Both robots are both CE and UL mark. So both can be used in North America or in Europe. But predominantly, the rollout is going to be next-generation robots. But for now, still on current, which is second generation. Yes. The way that we work is kind of parallel development streams, where, although we haven't yet rolled out next generation, we are working on the generation, the next iteration of that generation, and we're also working on a much more forward generation. And the whole game is to lower the total cost of ownership. So not just the initial capital cost but the total cost of ownership because we maintain responsibility for total cost of ownership and charge the clients for the provision of capacity. Yes. And then your last one was on vans. Look, the -- most of the vans in our fleet are diesel. They are Euro 6. They are significantly under the ULEZ requirements. We do have a variety of tests going on in our network with electric vans. One of the biggest constraints on electric vans is a combination of route lengths, given that we're a heavier user of energy in the van than the parcel businesses because the energy source in our van needs to keep the food cool as well as drive the van. Okay. Just worth remembering that. And -- but electric vans probably have all the [Indiscernible] and the test inside London, in the inner parts of London, show that we can get enough capacity out of them. The issue then is about electrical transmission into the site to get in the capacity to charge that many vans, the availability of enough power, okay? So that's something that we're looking at a number of sites, but that's a challenge going forwards for the country if we're going to fully go electric, basically.

Nick Coulter: And you're using natural gas as well?

Timothy Steiner: We're using natural gas on the LTVs. For example, we recently made a commitment on the site that we've taken in North London to only use natural gas LTVs to deliver to that site so they won't get large diesel trailers going into those, into that area. We also have a trial going on of a quite a unique hybrid model that is more like the kind of P3 BMW was with the range extender. So they're kind of, they're hybrid but not as in they can run on diesel or petrol and run on electric. They run on electric, but you have the range extender capabilities. And then you can do some clever stuff like geo fence, understand where the root is and decide how much power you've got and where on route you're going to actually run your combustion engine to recharge your batteries because the batteries haven't got enough technology. And we're doing that sometimes with chassis that have reached the end of their useful life as a, the chassis itself hasn't, if you saw what I mean, but we said that the vehicle has because it's engine has been used for 5 years. So we take the engines out and we're putting in a different power plants and stuff. So we've got a whole bunch of experimentation going on. I do think it's an important area going forwards, both for us and our clients. And so, yes, it's an area of focus. .

Victoria Petrova: Victoria Petrova from Credit Suisse. On Page 26, you were talking about ongoing zone projects and future zone projects. Can you maybe elaborate how remote your solution project on Zoom could be? What is the CapEx per Zoom CFC or ZFC? And also, given the current agreements with your international partners, if they decide to go with someone else in their micro fulfillment strategy, would they lose exclusivity or not? How do those agreements work?

Timothy Steiner: On the CapEx, it's probably too early to start disclosing on the model apart from the general comment that we're doing this and the model looks good. But it's too early to start giving you specific economics.

Victoria Petrova: Maybe relative to CFC sort of 10%, 5%.

Timothy Steiner: You'd expect it to be a bit higher because you've shrunk it down. And as I said, the end result of having a smaller site in a more expensive location with a shrunk down version of the kit, but you've got to have all the program management and project management to install it. You've got to take the lease on the site, et cetera. And you've got to supply it from a big shed. We still think, overall, the operating costs are going to be 8 to 10 points higher than the operating cost. The total fully amortized, et cetera, et cetera, costs mean that you need to charge more, basically. And you can do that through, for example, our delivery fee, it doesn't look large because the order size is small. But a £2.50 delivery fee on a £35 or £40 basket is materially higher than a £1.5 delivery fee on £100 basket. And you can do that, for example, in your consumer mix by having less promos as opposed to having anything that looks like materially higher pricing. So there are ways of doing it, which make it, it's an attractive proposition. And the pricing can be very attractive compared to the local stores that are convenient. And those local stores can have between 1,000 and 4,000 SKUs in them, whereas the Zoom can comfortably carry 15,000. So materially better range, fast service, not more expensive but has to be more expensive because those costs are a bit higher as a percentage of sales. And the second part you asked was something...

Duncan Tatton-Brown: Exclusivity around...

Victoria Petrova: Exclusivity.

Timothy Steiner: It's all slightly different, and it would depend on volumes and stuff like that. So it's a little bit tricky. Some, I don't think I can give a definitive answer to that. But if one of our clients rolled out an extensive network of micro fulfillment centers with someone else, they'll, I would imagine they'll naturally break their exclusivity anyway even if that wasn't specifically covered. And we would be able to sell our solutions then to their competitors. It wouldn't make a lot of sense, by the way, because the, our micro fulfillment centers won't be more expensive than others, but we'll have a huge advantage in terms of the way they interact with the other assets our clients are building. There's one at the front here, and there's one there. And there's one there, and...

Marcus Diebel: It's Marcus Diebel from JPMorgan. 2 questions, both for Tim, both on strategy, really. The first one on retail. We talked a lot about future developments, one on pricing. When can you actually expect even more on pricing than just the price metric? I can see a proposition that you are the most convenient and the cheapest, which would be quite powerful. Is there anything, or is it just too early given the capacity that you're running? But I think, yes, how do you think about this for the next couple of years? And then the second point, thanks a lot for your remarks on the last mile. I think it was very interesting and very useful. When you talk to some supermarkets who don't go with Ocado in the end, and they'd again talk about their branches, they seem to be concerned that the consumer in the long run would prefer both channels, i.e., buying the heavy commodities in the, through the fulfillment center online but still wants to go to the supermarket to see the meat and touch the avocados kind of theme. Is that valid? Or do you think this is not at all a valid and kind of [take] these basically reasons to keep the branches as they are and to utilize them as much as possible is a valid, at least, approach?

Timothy Steiner: Sure. Look, on pricing, we've made a number of pricing moves over the years and where Ocado.com first started as a price follower overall of Waitrose by starting to, in 2006, I think it was to start price matching Tesco's on all the branded products. In 2007 or '08 by basket matching the whole of Tesco's, by matching Ocado own label to standard tier Tesco's and Sainsbury's products and stuff like that. In terms of going the next step and saying you've got the most efficient operation and taking it under, as you say, we've never had the capacity. We've always sold all the capacity that we had. At some point in the future, if we were able to build the capacity faster to realize the economic benefits of the model, then the Ocado Retail Board of, which I chair and Duncan and I sit on, could decide that that's a smart strategy for them to do. But at the moment, there's plenty of demand for the facilities that we have and the facilities we foresee rolling out. We think there's an, there should be an increased excitement and opportunity for that business to grow based on the improvement in the proposition that's coming in September anyway. M&S currently is driving its pricing down, I think, quite noticeably in the market, to be more price competitive. Waitrose has been driving its pricing up to maintain, in our opinion, to maintain profitability of the supermarkets to make up for holes elsewhere in the group. So the switchover will, in itself, be a reduction in pricing to the consumer. Because right now, M&S Food is cheaper than Waitrose Food. Moving on to the question on the last mile. Look, in 2007, we went out and did extensive surveying of customers, and some customers that like to go in the store and like to look and feel. So we went in the store, kind of, I just don't want, I don't want this to sound creepy, but we went in the store and started watching consumers' behavior, if you saw what I mean. And what they really like to -- they didn't really like to feel much because it's not much you can feel. And -- well, particularly, at the moment, we're not supposed to touch things. But what they were really looking at was the sell by date, was the life. They want to make sure that on the shelf, if part of the shelf was replenished more recently, usually, the back of the shelf, they want to grab that product. And that's when we altered -- effectively all our systems have to be touched to put product life on the site. And today, Ocado.com has the highest fresh food sales of any full range supermarket in the UK. So I think that -- and customer baskets, I think we've said historically, grow by over £10 from the first shop to the tenth shop, as customers start to understand the quality and the freshness of the product they can receive from us as well as familiarity with the UI. And so whilst it's an interesting proposition to say, get your heavy bulky from us and come in and pick your produce and your meat, we don't see anyone -- the quality of produce and meats that we can pick for our customers is great. And we can avoid lots of people handling them in front of you. And we don't see it as an issue. If a client wanted to use -- to put a micro site in an existing store, they could. I would strongly advise them to have the chill and fresh in it. But I wouldn't stop a customer using it and doing a collection at site or from a part of the range and then buying the other part. I think in the explanation I was -- explained before about building a micro site, to get micro site efficiency working, you need to do a material amount of sales, let's call it $20 million plus. And the issue there is if you're going to do $20 million plus and it's just dry Ambienand it's for collection, because it's only good for collection if you've only got dry Ambien in it. They got to come to buy the fresh. You basically got to put 100% of the store sales through that channel. And I'm not sure the market is yet at 100%. And so it's a challenging thing because if you build a -- you can't build automation, an operator and have the engineering on site to keep it live to do $2 million or $3 million of Ambien alongside some fresh food that you're going to sell in a single site. The economics are not going to stack up, okay? So I'm a bit skeptical at the moment.

Marcus Diebel: Maybe just to add on this. I mean, then as a consequence because I think everybody can agree that the online share will go up and it's more convenient. But I still don't fully understand how the supermarkets then think about their branches. I mean, conceptually, they would say, okay, we've given these efficiencies. We just run the fulfillment center. Perfect. Yes. But they are having a lot of branches. So what would you -- from what you can say, how are your customers actually thinking about the branches about closes and how their model really looks like in 5 years' time?

Timothy Steiner: Look, I think let's remember that, at the moment, still, the main channel is from -- through the branches. And some of the branches, particularly usually in the urban, in the centralized urban areas, are still actually delivering great economics. It's just that sometimes, there are too many branches or some branches are in the wrong locations. It's definitely possible, and clients are definitely looking at carving some space out some of those branches. And we can put a micro site in there, and then they can, therefore, give 10,000 or 15,000 square foot of the site to run a micro site to do £15 million or £20 million of annualized sales to serve that catchment area. So that's a better utilization of space where they've got too much space in a catchment. That's one option. If some hypermarkets are of a certain size that if the economics of the hypermarket no longer makes sense and they can do the right planning changes, the size sites actually might be perfect for many warehouses. So there are options of ways to reutilize things, but there may be challenges going forward with some sites that, if more and more people shop online, then, if the market is not growing fast enough, there might be some challenges in some of the bricks-and-mortar sites for sure. Mic is just behind you.

Tom Davies: Tom Davies from Berenberg. Three questions. So regarding the sourcing of the MHE, what proportion of the parts are sourced from China? And have you seen any impact from the coronavirus? Essentially, like, what flexibility are you building into the supply chain of this? Secondly, regarding Zoom. Are you able to provide a bit of an insight into the incrementality of these convenience shops versus the main shops of the customers coming through the CFCs? Essentially, are you increasing the absolute spend per customer, an absolute cash profitability of each individual customer? And then finally, on Zoom, what time frame would you expect it to be offered to partners?

Timothy Steiner: The first question on the coronavirus is a great one. I think it's a little early to be able to give you a fully accurate answer. The majority of parts we don't believe are coming from China, but there may be subcomponents of parts that are. And we don't currently have any delays as a result. There's at least one part where we are looking at alternative sourcing, where we believe we could encounter delays if we don't. It's a non-electronic, just solid casting part that we can manufacture elsewhere that was coming out of a part of China that's hard to get stuff out of at the moment. So we think we can mitigate any effects of the coronavirus. But obviously, that will depend for everybody on how extensively it spreads, et cetera, et cetera. Zoom, look, I think it's fair to say that more than half of the sales are incremental. So there is some cannibalization and that, of Ocado customers, more than half the sales are incremental, and then some of the sales are from nonexisting Ocado customers as well. So more than more than half of the sales are incremental. Right now, given the fact that we're not using it with the right infrastructure that someone around here raised before, it's currently, that's not improving our profitability as some of it migrates. In the long term, we would expect both for our clients and for us to have a similar net margin on sales, whether that sale goes through a big warehouse, a mini warehouse, a micro warehouse, whether, so we would expect to be able to help our clients build a model. That means their profitability, they're indifferent, and our profitability as the supplier of the solution, we are also indifferent to where that sale goes through. Can I just clarify one last point on your incrementality, and it's not affecting the basket size of the overall order. So they're not taking 80% of the order there and 40% there. It's an incremental shop. But overall, there is some cannibalization.

Xavier Le Mene: Xavier Le Mene from Bank of America. Two questions, if I may. The first one, just looking at the pipeline you've got for the new opening for the CFC outside the UK So can you help us a bit to understand what is the time frame here going forward? And potentially looking at the EBITDA, where should we expect the EBITDA to breakeven for intentional solution, in what year? Second, just on the balance sheet and where you are now. Are you happy with what you've got so far? And do you think that's enough to fund all the partnership, including Aeon? Or do you think, at some point, you would have to potentially raise more cash?

Timothy Steiner: Well, on the pipeline for a moment, sorry, you just lost me on the question for a second. What was it? Where are we on the pipeline and...

Xavier Le Mene: Yes. How much CFC per year, more or less?

Timothy Steiner: Look, I...

Xavier Le Mene: For the year?

Timothy Steiner: It's increasing the CFC build per year. Obviously, we're expecting the first 2 to go live in the first half of this year. We're expecting the next ones to go live in the first half of next year. And then eventually, we'd expect to have a more, kind of, standard rollout. As you know, we said, and, the equivalent of, kind of, 20 warehouses of a certain size with Kroger over a 3- year period, so that's 6 to 7 a year. Plus, you can add on to that, so something for Aeon, the UK, France, Australia, Canada and then the one site in Sweden. So, and then potentially more. So we can see that number growing. So whether that's 10, 12, 14, 16, we have a trajectory. And we can see we're really more focused on the number of live projects that we have to manage at the same time. And obviously, that number could grow as people want to add in Zooms or somebody might decide to order 3 minis instead of a medium-sized one or something like that. So we're geared up for that. It's just also worth understanding that our clients need to get planning and need to find the availability of sites. And so some of them have gone very fast, faster than anticipated. Some of them have taken a little bit longer than they had hoped. And so there's all that variability in there, and we have to be capable of managing that and deploying at the right time. There are resources at the right time on it. In terms of the numbers part of that, I'll hand over to Duncan.

Duncan Tatton-Brown: Yes. I mean, if you take the £600 million CapEx this year, take off 85% for Andover because that's being funded elsewhere and isn't obviously going to repeat, and then allow for the fact that if you take a 40% growth on £80 million of cash fees, call it, about £110 million of cash fees, you're back to £400 million cash outflow for Solutions. Ignore the retail business that can generate some cash, £400 million cash outflow. So let's assume that's 3 years worth of funding because we've got £1.4 billion of headroom. So I'm a CFO. So I'm conservative. So I'm going to sit there and say, well, in 2 years ago, I'm going to think about how you might fund the business. I can tell you what? In 2 years' time, the next 12 months EBITDA will look quite attractive. So I think if there is a need for funding there. And frankly, I hope that we're in a situation to say we do need funding because we've signed up lots more plants with lots more capacity needs. Then the profile of the EBITDA in the next few years will look quite attractive. So yes, there may be a need to more funding. Let's hope so.

Xavier Le Mene: Just in the EBITDA in terms of the UK [indiscernible]?

Duncan Tatton-Brown: To a certain extent, I'm not giving you that answer, which is in 2 years' time. The next 12 months should look pretty attractive.

Bruno Monteyne: Bruno from Bernstein again. You didn't mention there was something special about the CapEx that you have to clarify, Duncan. Could you give that secret, please?

Duncan Tatton-Brown: Bruno, you didn't pick up the hint. It was the Bristol. We're funding the building as well. Just to make sure that we pick that one up as well. Tom again.

Thomas Davies: Tom Davies, Berenberg. One follow-up. Regarding the whole funding question, the existing convertible bond was issued in 2017, which is before the Solutions business signed most of your deals. Now obviously, you're a much larger company. Would you consider addressing this bond, potentially replacing it with a larger one where you could get a low coupon?

Duncan Tatton-Brown: So you don't look at the CEO with a funding question because that's my job. We'll obviously keep the capital structure under review and we'd do anything that gave us a more efficient form of capital. But, so I don't think you should assume anything in the near term. But, yes, we'll, absolutely, we'll keep it under review.

Stuart Rose: We've kept you a long time. Any last question? Otherwise, we're going to [indiscernible] now. Andrew, one quick last one.

Timothy Steiner: One, Andrew, not 5 or 6.

Andrew Gwynn: Yes. And it's Andrew, not Bruno. So...

Timothy Steiner: Is it Andrew?

Andrew Gwynn: Yes. No. Sorry, but Bruno with his multiple questions. It's a quick question. It's a hard answer. Next 12 months, next 24 months, massive years for execution. Where do you think the key risks are? And any sort of particular things we should have in mind as you go live with partners?

Timothy Steiner: Look, to be honest, I get asked this question a lot. What is the thing that keeps you awake at night? And I would say if there's one thing that kept us awake at night, then we're not doing a good job of balancing the business properly. So I don't see a single risk, if that makes sense, or one thing. Obviously, it's complex to build in multiple countries at the same time, the amount of code that we're developing, the challenge of new robotic design and bringing down the operate -- the ongoing engineering costs, keeping the clients happy. I need to be in 5 countries at the same time sometimes. But I -- there isn't a single thing that I would say for you to watch out for because that's the thing that we're worried about but we're kind of hopeful, because I actually think that we've got a good balance of resource on all of the challenges that we have as a business. And I think it's just look for continued overall execution. Obviously, if we were sat here in 9 months' time and we weren't -- we haven't turned those sites on live for our clients in Toronto and in Paris, you need to be asking some very serious questions to whoever is sitting in this seat because it won't be me. But -- so obviously, there are some very clear progress items that we need to make. But I don't see that being a thing that we're really worried about, if that makes sense. Plenty of stress and challenges on an ongoing basis, but nothing outstanding.

Duncan Tatton-Brown: And I think all of our investors recognize that we are not a company that you're investing because you want a slow, steady, safe, modest growth. So by its inherent nature, we're taking on more than most would do, which does not mean we're without risk. But I think our, as Tim says, our execution of that and the ability to constantly adapt to that means, yes, sometimes, there might be some bumps in the road. But overall, we are looking with great prospects for the years out.

Stuart Rose: Tim and Duncan, thank you very much. Ladies and gentlemen, really appreciate your time today. Thank you.