Ocado Group plc

Good morning, everyone. I'm Stuart Rose.

I'm Chairman of Ocado. Welcome to this meeting, Tuesday results meeting for the 52 weeks ending the 2nd of December 2018.

I think it's probably fair to say, for me, I've been involved with this company for 6 years. It's been the most exciting six years that I've had.

But in the last year, it's been more exciting, still. Very transformative, as you have seen yourselves, and you're going to hear more from Tim and more from Duncan on that in a moment, but there's plenty more to come.

Few comments for me before I hand you over to Tim. We've established Ocado Solutions as the go-to partner now for grocers looking to create a sustainable and profitable online businesses for their customers, also, at the same time, as offering unrivaled service and value.

The Ocado Smart Platform will soon be powering the e-commerce business of 7 of the world's most innovative and forward-looking grocers, including Morrisons, that you know about; Bon Preu; Groupe Casino Sobeys; ICA in Scandinavia; and of course, Krogers, which is the most recent announcement that we made. And over the coming years, we'll be building 23 new CFCs for our partners and one new one for ourselves in the U.K.

And if you remember, that we've only opened, prior to that, two since 2016. We'll continue to engage with multiple retailers in a variety of markets with a view to adding more partnerships to this platform, and at the same time, we will continue to reset the bar for our Retail operations, improving our shopping experience, improving and winning new customers up and down the country and leading the channel shift online.

This growth has been enabled by a new generation of Customer Fulfilment Centres, and the most recent of which is the one we opened in Erith in Southeast London, which opened last summer. And Erith will, at maturity, be the largest such facility of its kind in the world.

And our success in ramping this up has shown that we can do this for ourselves, and shortly, we will be doing this for our partners. So having done all this in 2018, you might say to yourself, "Well, this transformation story has played out.

What's to come next?" Well, nothing could be further from the truth.

And in many respects, I think and I believe that our story is only just beginning. We are committed to the successful execution of our various Solutions deals.

We now have the resources in place to deliver world-class execution, and we will help our partners lead channel shift in the markets in which they operate. And we'll be doing all this while progressively reducing the cost of ownership of our assets.

As we reduce these costs, so our return in capital will improve year-by-year, not just in the short term but also in the longer term. We're also taking advantage of the growth opportunities we have available in the U.K., responding to customers' increasing demand, which there is on a daily basis, for increased value, increased quality and better service, which we believe the Ocado model is the best-in-class to provide.

We're constantly striving to raise the bar. And this is why you'll be hearing from Tim and from Duncan shortly about our work we're doing to develop a more immediate offering.

And while the hard work in building our solutions in retail businesses, we're also innovating for the future. And we will be exploring new ways that we can create value in different markets through our broad and deep understanding of technology.

And all of this is yet to come. But rest assured, the Ocado team remains focused.

We're focused in serving our customers. We're focused on building our business.

We're focused on creating value for our shareholders. We are very, very busy, but we are, as I said, focused.

Now I'm now going to hand you over without any further ado to Tim, who's going to take you through the bits and parts of the results presentation. Thank you.

Thanks, Stuart. Good morning, everyone.

Our model remains as it was before. It's around our investment and our investment that allows us to innovate.

And we innovate to improve the customer proposition, the customer experience, which helps us to drive growth, helps us to improve our efficiency in our operations, helps us to reduce the costs of increasing our capacity; so higher productivity at lower cost, more customers, better service, drives growth; and the growth allows us to create more investment through more profitability. We're changing the way the world shops.

So when we stood here this time, last year, we were able to show ourselves as a customer of our own infrastructure: Morrisons who signed on in 2013, and we were able to announce that Groupe Casino had signed on in November '17. During the last year, we were able to add Sobeys in January, ICA in May and then Kroger for the 20 buildings they've asked us for in the first three years in May as well.

So to now, OSP will be powering seven of the world's most innovative retailers. I'm going to hand over now to Duncan.

I'm going to come back on after he's gone through the numbers.

Thanks, Tim. Good morning.

I'll now take you through the financial results we announced this morning before Tim talks to you about our future. So the key headlines.

Our Retail business continued to deliver market-leading growth. We improved our offer and we made some gains in efficiency.

And for Solutions, we're hard at work building the teams and delivering on the significant commitments from our existing clients, with further prospects ahead. Finally, we're well-funded to enable us to invest at pace.

So the financial summary. Now just a couple of points on the presentation: First, the slides we'll be sharing compares against a 52-week year for 2017.

Second, as you already know, we have adopted IFRS 15, and so the 2017 numbers have been restated for the impacts on the Solutions segment, and hence, overall numbers. Now we've continued to grow revenue strongly with our Retail business up 12%, well ahead of both the total and the online grocery market.

Revenue growth was even stronger in Solutions. This reflects the revenues primarily from our arrangements with Morrisons, as the fees that we own from other Solutions clients are not yet recognized until the relevant performance obligation is met, typically, on opening of the CFC.

Overall, our revenue was up 12.3% to nearly £1.6 billion. On EBITDA, the story is different for each segment.

Our Retail EBITDA grew to £82.5 million, as expected, below the level of revenue growth, largely due to the additional costs of Erith and Andover, which are operating at low capacities. Solutions EBITDA was a loss of £17.9 million as the growth in reported revenues was offset by significant further investments to bring on the resource to deliver these facilities, further invest in the platform and expand the commercial and relationship teams.

Our Other segment includes the accounting impacts of MHE JV Co, the cost of the board and of share-based incentive schemes. I highlighted at the interims that we should expect up to a further £9 million cost given the significant increase in share prices, and this has come through as predicted.

Overall EBITDA, therefore, fell to £59.5 million. Net interest costs were down to £12.5 million, given the interest income from increased cash balances.

Depreciation costs increased to £91.3 million, where the biggest increase is from new CFCs and from our new software platform. As a result, we reported a loss before tax of £44.4 million.

As I explained a few weeks ago, the impact of the new standard IFRS 15 on the recognition of revenue from contracts is to defer revenues that would have been reported in earlier years to later years, with no -- overall, no change in total revenue, cash fees or value. Here, you can see that there's a reduction in revenues of £9.3 million for 2017, and we estimate £15.2 million for 2018.

This impact only in Solutions and flow to the bottom line. In 2018, there was also some reduced costs, as cash costs relating to some specific sales bonuses have since been capitalized.

So the total estimated EBITDA impact is £14.4 million. As promised, at the back of your slide pack, there's details on the restated first half numbers.

As usual, I've split up the Retail and Solutions EBITDA in a bit more detail, showing admin costs for each segment. And I'll cover Retail operating contribution in further detail later, so for now I'll focus on Solutions.

Revenue growth grew 15.8% on the prior year, with operating contribution up fast to up by nearly 31%. Note that revenue includes Morrisons' income from both recharges of cost and for management fees, so effectively, with low margins.

As we now have our first new style Ocado Solutions contracts, I'll explain how we report the costs. The costs of maintaining the CFC remain in operating contribution, but the balance of cost comes through in admin costs.

This include both the noncapitalized costs of developing the platform, but also the cost of hosting and support. So you should expect, as the proportion of Ocado Solutions contracts grow, to see improving operating contribution margins but also higher admin costs, and of course, higher depreciation.

We've disclosed for the first time the cash fees that we earned during the year. So whilst reported revenue for Solutions grew nearly 16%, cash fees were up 37%.

Currently, the majority of reported revenue relates to cost recharges and management fees, including those for the original contract for Dordon. For Morrisons' Store Pick and Erith, both are now live, and so we're reporting revenue from these cash fees over the expected life of the contract.

Revenues which correspond to the cash fees from CFC contracts for Groupe Casino, Sobeys, ICA and Kroger are not yet being reported as their facilities aren't open. So onto Retail performance.

Retail revenue was up 12% on the year, with total order growth slightly ahead at 12.1% to 296,000 orders per week. Hypermarket baskets were slightly down as we continue to see a bigger mix of mobile orders offsetting some price inflation.

The growth in orders was an increase in average order volumes of 32,000 orders per week. Our new CFCs, Andover and Erith, provided most of these extra capacity and are both continuing to perform as expected.

In Erith, we now have a peak week of 35,000 orders, and we expect further growths over the year ahead. Mature CFC efficiency, meaning for Hatfield and Dordon, was stable at 164 UPH.

Dordon actually improved marginally, but this was offset by the operational impacts of severe weather in the first half. Andover UPH is improving, and we expect it to be ahead of Hatfield by the end of the half.

In fact, last week, it was already at 140 UPH. Erith UPH has also started positively and for the first 6 months is approximately 30% more efficient than Andover was for its first six months.

Our delivery efficiency, as measured by Drop per Van per week, improved by 6.7% to 194. There'll be further opportunities to come, and we revised our target again, now to 200.

Wastage costs remained at industry-leading levels, although slightly up on the year due to the higher waste percentage at Erith, whilst operating at lower volumes. And note that the actual product that is wasted and ends up in landfill remained at 0.02%.

So as promised, back to Retail operating contribution. Gross margin were stable across the year as a whole.

The market remains competitive, but we've annualized against the major price moves, and so in the second half saw some modest margin gains, which offset the declines in the first half. We continue to find ways for our suppliers to increase their sales with us, and this helped us grow our supplier income by 40 basis points.

Trunking and delivery costs were lower due to the improved efficiency I mentioned earlier, offsetting the wage cost pressures. The biggest cost movement was for CFC costs.

Operational efficiency was flat in maturer CFCs; and despite improving UPH in the newer CFCs, as more volume was handled in these facilities, their currently lower UPH impacted CFC costs by around 30 basis points. In addition, despite engineering cost per order being less than 1/2 of the previous year, they are currently higher than in the mature CFCs and there were higher fixed costs from a full year of Andover and a part year of Erith.

The combination of these impacted CFC costs by 60 basis points. As an aside, it's important to note that the engineering cost per order in Erith are already lower than Andover as we benefit from our learnings, such as with better reliability of their latest generation box.

Other operating costs and marketing costs were both up slightly. We focused on improving the quality of our customer acquisitions, with retention of new customers was 3 percentage points better than the previous year.

So overall, operating contribution was down around 30 basis points; retail admin costs were up by less than 20 basis points as we invested in some central teams to improve margins; and overall EBITDA was down 40 basis points, with the differences only in the roundings. We spent £213 million of CapEx, with less than £30 million on maintenance-type expenditure, primarily in maturer CFCs and for delivery vehicles.

Including in this were investments to improve our operations, such as the installation of further bagging machines in Dordon, which should help improve efficiencies this year; and from uninterruptible power supply in Hatfield to lower the impact of problems with electricity network. We spent £80 million on new CFCs, mostly at Erith.

The majority of the balance was on further development of our technology and fulfillment solutions as we scale our teams to improve our platform further. And we've begun to purchase materials for international CFCs and spent almost £11 million in the year.

Finally, within the other are the costs of the second general merchandise distribution center and the cost of our immediacy trial that Tim will cover shortly. Our cash position is strong.

We ended last year with £150 million of cash, and we've had two equity issues since then. And despite spending over £200 million of CapEx, we had underlying cash outflows of only £65 million.

So with our undrawn RCF, we have over half -- we're over £500 million of headroom, giving us sufficient funds to keep growing our U.K. capacity, improving our platform and meeting the current needs of our Solutions clients.

So to finish, I wanted to give you some indications or our outlook for 2019. Assuming normal market conditions, we'd expect our Retail revenues to grow up between 10% and 15%, with the upper end dependent on delivering sustained, fast ramp-ups in our newer CFCs, Andover and Erith.

Although this year we'd expect to commit to a 5th CFC, we would not expect a material increase in fixed costs as the last year. We'll have a full year of cost for Erith, but Andover has annualized and both facilities should be operating at higher capacities, so reducing the drag on EBITDA.

As a result, our profitability should benefit from the scaling of the business. For Solutions, I explained in our recent session on IFRS 15 that there'll be an estimated impact of £40 million on revenue in 2019.

I know that many of you have already adjusted for this in your forecast. On costs, you may not have sufficiently allowed for the additional costs that we'll incur as we continue to invest in our platform, build the support teams and execute on the CFC projects.

I'd assume a further £15 million to £20 million of costs. It's worth noting that our Solutions teams are busy and involved in a number of conversations.

Whilst making no promises, there's clearly the prospect of further deals to come, which, as you now know, would result in cash fees and some further costs, but as no revenue would be reported with impact on profits in the short term. Now lastly, onto CapEx.

Given the significant number of CFC projects underway internationally, there's a big growth in CapEx, up to an estimated £350 million. Details on the split are in the back of your packs, but the underlying CapEx is down and international CFCs drive all the increase.

2018 was a transformational year for Ocado. 2019 will be busy.

We have much to do. And now, hopefully, Tim will give you a sense of the ambitions ahead.

All I can say is there's still much to come.

Thanks, Duncan. As time is moving on and we're progressing our platform, progressing the development work of our new series of robots and our other infrastructure, our risk is actually coming down.

So despite the fact that we're scaling, time is passing, risk is coming down. As we roll out our new CFCs for our clients and as we build more and more of them, we get to benefit from the ever-improving nature of the platform, the more reliable solutions.

We get to benefit from manufacturing economies of scale, and our return on capital will significantly increase. So right now we're enhancing our platform with the aim of reducing the long-term cost of ownership.

We're focused on reliability, efficiency, manufacturability, serviceability and throughput. We're working all over the place so I can only touch on a few of the small points that we're doing today.

During the course of 2019, we'll introduce live robotic picking into production in Andover. We will introduce during the course of this year our third generation of robot.

We're currently building the first prototypes of those in our facility in Welwyn. They have over 75% new components compared to our second generation of robot that we introduced during the last year.

That robot itself had 25% new parts compared to the first generation that many of you will have seen running around in Andover. For those that have been to Erith, you will have seen a number of the second-generation robots as well.

As we make changes to these robots, we significantly increased our testing capabilities. So today, when we introduce new software onto the robot, we performed over the equivalent of 1,000 hours of testing for every software release that we do.

We built a new facility this year for where we design and test our robots, and we increased the amount of physical testing we're able to do by a factor of twentyfold. We went from 1 rig that we run five days a week for 10 hours, to 6 that we're running 24 hours, seven days a week, 20 times more testing capability, so that we can get those robots working the way we want to much faster than historically.

We've written a lot of cool software that means that we'll be able to run multiple facilities across the world at the same time and manage them from a single room in Hatfield. Examples of that would be machine learning software that reads the talk coming off of the wheels of each of the robots and is able to spot any discrepancies, compare them to historical issues and tell us that we've got a water spillage on an exact location in the grid or an oil spillage, it knows the difference; or that a piece of the grid has started to move slightly, and there's a small ridge that needs adjustment of one of the legs in the columns.

This is the type of thing that means that we don't have to have permanent experts in every site that we operate in the world. We're able to have software that can tell us what's going on.

We developed software this year using vision systems that can look inside our bins and tell us if they're dirty. But that sounds really simple.

But actually, they can tell us whether the dirt inside is innocuous or whether it's chicken juice and that tote needs washing. And they can do all that using machine learning.

We've got software that can now use the camera systems, the CCTV systems, that tell us exactly where every bot is. So if a bot looses its location, which we used to have to -- manually, a human would then work out where it was, which was the single cause -- single largest cause of any collisions that we had, was a human error there, we use machine learning to look at where they are, interpolate the grid from the fisheye lenses and tell us exactly where that bot is and tell it where it is so it can reprogram itself; so lots of software being developed, not just to enhance the actual operation but to allow us to run multiple operations at the same time; so all around, continually reducing the long-term cost of ownership and operation of all of these facilities.

We want to grow our Retail business, and our Retail business is core. The U.K.

online grocery market is set to grow again from £11 billion to £17 billion by 2023. It's set to capture 20% of the growth.

We've been outgrowing not just the U.K. grocery market, but the U.K.

online market as well with 12% growth versus 9%. We expect to announce the location of CFC5 in the first half this year.

But also, today, we're announcing that we are live-trialing to actual customers. We were live-trialing to friends and family before Christmas our new immediacy offering, which will allow us to tap into a part of the market that we've not addressed before, the sub-£60 basket market, which is over 50% of the U.K.

grocery market today. So enabling us to build new CFCs to capture, continue gains in market share as well as operating and capturing previously unrealized opportunities.

So we'll be launching Ocado Zoom, built to deliver the best possible service within 60 minutes. So the key difference from Zoom to many of the previous immediacy offerings that you may have seen are: one, it's a genuine, long-term sustainable and profitable opportunity; two, we will sell customers over 10,000 lines, so a range more equivalent to an actual Tesco's or Sainsbury supermarket, not the 1,500 lines you'd find in one of their convenience stores; and it will have competitive pricing and low delivery fees.

We're able to do this, because it is a unique model with micro-automated centralized fulfillment; being fed itself from our large-scale; automated centralized fulfillment with a common asset and modular common infrastructure; and part of a complex delivery network operating between the two with clever software that enables us to replenish it very cheaply and efficiently in small-batches, daily delivery but allowing us to carry a large range in a small site with freshness and immediacy. We've built a site, a test site that will launch to customers in March in West London.

And over the coming years, we expect to mature this offering and obviously make it a part of our offering to our international customers to enable them to add more immediacy in their core markets. We're focused on innovating for the future, and just to give you some idea of the scale, we've grown our headcount in technology in the last three years by 76% from 750 to 1,300 people today, that 750 was itself a 150% growth in the previous two years.

So we were 300 heads in 2013; 750 by 2015; and 1,300 today, and we're continuing to grow. We generated innovations and protected them.

We filed now 395 patent applications, a fivefold increase in where we were three years ago, covering 80 separate innovations. And we've got a really broad and deep technology real estate, our talent bases across multiple areas, including those listed: real-time control systems for our robots; robotics, a variety of different areas in robotics; computer vision systems, I mentioned before looking for totes, but also for the picking systems that we'll be rolling out.

We're using machine learning and AI in all manner of places, whether that's for replenishment forecasting, whether it's for fraud, whether it's for recommendations on the websites and stuff like that; a lot of data science, forecasting, routing systems. We're improving our routing system algorithms, which will drive further increases on our deliveries per van per week; inference engines for voice and understanding, which we're using on our voice interfaces, like Alexa; the Internet of Things, all over our warehouses today.

We built private cloud in our warehouses so that our bot routing software that had -- that can't suffer the latency of the public cloud is able to operate, they're still in cloud-based software that enables us to run it on cheaper infrastructure, more reliable infrastructure; and an obviously, enormous amount of big data in our business. And the big data that will allow us to run so many sites because their modular solution is a digital solution.

Before we build a facility, we design the facility. We put it into our simulation software.

Our simulation software will tell us what we're able to achieve from that building. The same software is what will actually run it.

There is no fine-tuning that you have do in traditional warehouses. It's build the grids, put the robots on, put the peripherals in and get going.

Well, maintaining and then increasing the velocity of our innovation is crucial. We're primarily doing it organically, but we're not closed off to inorganic growth or through partnerships as well.

We're innovating as well, and we have been for a little bit of time, outside of our core. This is not an incubator where we throw in some smart people and then throw in some money at them.

This is all-around innovations that stem from the core technologies that we're developing for our grocery business. But as we have developed of those technologies, we can see some crucial and critical uses for those technologies in other sectors where we think we can be equally disruptive.

And so we just put a little bit of that information down here in the likely possible and aspirational columns. And to say that they're likely and most similar to what we do today, things like general merchandise.

The possible are things where we modify slightly from where we are today, possibly in scale. So think of a very large version of Andover: instead of having groceries in each of those boxes, pop a car in them.

You've got a car park that has significantly more density than a traditional car park at significantly lower cost. And so you've got a whole load of opportunities here.

Vertical farming, something of great interest to us as a retailer and our partners as retailers. And vertical farming is an area that's making huge strides at the moment, but they're all in the science of growing inside as opposed to in terms of making those buildings more efficient and more dense, which is what a solution similar to our automation could achieve.

And so as our CTO Paul likes to say, we're not just changing the way the world shops, we're changing the way the world stores, assembles, sorts, moves and sells atoms. There are no guarantees and no time lines on any of these projects.

But we just want to explain to you that behind the scenes we're working on a large number of very exciting opportunities. What's absolutely critical is we have a separate team working on these ideas.

They can take IP out of the business, modify it and use it, but they cannot distract us from the absolute critical mission of delivering on our core mission and delivering for our customers that have signed up to work with us today. So in conclusion, our market remains an inflection point, as we highlighted earlier this year.

Now is our time. We've spent 18 years becoming the world leader in our field, in a field that most people didn't think existed.

But today, all over the world, all grocery retailers are focused on the channel shift that's occurring in their markets, on the need to apply robotics and automation to generate greater efficiency in a market that has always been won by the person who can cost-dominate the most. Outstanding execution is our priority for our retail customers every day, and for our retailer customers as we build out and help them to launch their own platforms.

We have many opportunities in grocery and beyond to create future value. And we still believe very firmly at Ocado that our story has only just begun.

Tim, thank you very much. Ladies and gentlemen, I'm going to invite Duncan and Tim to come the stage.

We're going to take questions. If you could use the microphones in front of you.

There's a red button. If you press that, then you'll be live.

If you could just say who you are and who you work for, that will be great. Thanks.

Nick Coulter from Citi. Could you expand on how the delivery economics might work for the immediacy offer?

I get it from the micro-fulfillments and how the robotic picking will come through, but the delivery economics and the drop densities that you assumed and how that might work.

So in a market with sub-60-minute deliveries, we're talking initially, certainly, about courier-based deliveries, so third-party delivery services. The economics of those are fairly well-known in the market.

They're mid-single-digit pounds per delivery. The limiting thing is the size of the delivery to the customer.

And I guess the way to think about it is, if you look at our core retail businesses, you saw before the percentages that we spend on delivery. If you think that, that's what our core retail business can contribute on a smaller basket, that same percentage is a smaller number.

But if you add that with a small delivery fee, you've ended up at a number not dissimilar to what a courier will charge to deliver that. The key is, is that we can run the warehousing, the micro-fulfillment center, and most critically, the supply chain into the micro-fulfillment center for not significantly more than we can run the large facility.

And that's where we dramatically differ from a number of other solutions that people are trying to sell in the market, in that they're all focused only on once the goods are in this machine, how quickly or how efficiently they can get them out the machine. But in the grocery market, you got to focus on all the costs: from the supplier all the way through to the customer.

And we think what we have or what we will be trialing here and later rolling out in full glory is the most efficient way to do that, some of which has some patent protection we filed a number of years ago on the thinking here of how these different facilities interact with each other.

So effectively, you're saying that you can manage the frictional cost of an extra touch points on the inventory and existing costs of courier delivery, making economic on that? Basically, you don't need -- it's irrelevant for drop densities, effectively?

The drop density is largely irrelevant because the person that's going out is doing one or two drops. He's not doing 20 drops.

So the density is irrelevant. Where density comes in is to achieve those delivery economics, you can only deliver in a radius of, say, 4 kilometers or 5 kilometers from a single site.

And so you need to be a big enough retailer to be able to generate the amount of sales you want to generate from that single unit in that radius. But I mean, if we looked at that radius from the site that we actually have secured, our existing sales are more than 3x what we would want that site to generate in immediacy.

So we know it's an area of rich density. But the thing with immediacy is it's not something that you would go and put in a low-density area.

It's something that will work very efficiently in a high-density area.

Will you use smaller totes and will you partition the totes and keep the...

I think we -- we won't deliver to customers in a tote. The bags will be what's being delivered to customers.

So there are going to be a number of changes. But you'll see when we -- if you're lucky enough to be in the area.

Duncan's in the area and I'm not. So I'm not sure who in Ocado chose the site, but I'll be having words with them.

And maybe just one thing on that. You talked about additional frictional costs.

Two thirds of our deliveries go through a spoke network. And this instance, you can go from our CFCs direct to a micro-fulfillment facility and not go through a spoke network.

So actually, it's not necessarily -- it doesn't have the impact that you would consider.

So will the trunking cost offset those frictional costs? So would you still expect additional cost?

Not going to give the detail. I think Tim's given you all the answers you'll get today.

Come back and remind the point. We're doing this thinking that the economics of these are attractive.

We're not doing this as a way to lose less money than our competitors are doing the same thing. We think this is an attractive opportunity, if proven.

And it's a trial, but we're highly confident.

Bruno Monteyne from Bernstein. First, on this new solution.

If any of your international partners decide to take it up, I presume there would be additional licensing fees. This will not be covered by your current deal.

Would that be correct?

So with our partners, anyway, every additional facility, they pay again for the new facility. So this will just be another facility.

Obviously, the exact amount of charging might be slightly different as a percentage of sales or volume or something. But yes, they would be additional sales to our existing partners.

And how quickly can it contribute to the U.K. growth?

Is this something to expect in 2019, or is it just a trial and we should just hold our breath?

I think in 2019, it's more likely to be a cost than it is a huge contribution because it is largely a trial, and then we'll start to roll out a more advanced facility. So there are a number of changes in the software and the hardware that we'll make.

So the trial site isn't -- we know it's not the endgame, but it's something that we want to start using to trial various aspects of the service and the proposition and the customer reaction and things like that. But we very clearly have a vision of what it is that we want to end up with.

And we'll work to get that in our second or third or our fourth site, whether that's in the U.K. or international.

And the last one is about -- or the second last one is about the outlook. On the one hand, you talked about revenue and EBITDA to be £40 million based on IFRS 15.

Now that's a step-up of about £25 million versus this year. So I'm sure there's a step-up in fees you were going to get, so is with the costs you're going to get.

These additional costs that are implied in £40 million guidance, are they the same cost as your second guidance, which is cost up an additional £15 million to £20 million? Or is that second bake an additional cost on top of this?

So is that a double guidance or one of the same thing, really?

So one deals with revenue solely. So the first one is cash fees that we were receiving, expecting to receive, had received, and we had expected to report as revenue we would no longer be reporting as revenue.

So that's one. The second thing is, we don't think the market had fully allowed for the pace at which we're going.

We're very clear that there's huge opportunities ahead. We're developing and investing in the platform.

The CapEx guidance I give you in the pack will show that there's about nearly a £13 million growth in development CapEx. Just as a guide, expect about 2/3 of the actual cost of the people doing it get capitalized.

One third are, therefore, straight hit in the P&L. That gets you to about £15 million anyway.

So this is just about going faster, developing our platform faster.

Duncan, I'm probably just thick, but in the beginning years, you always said your deals are roughly P&L-equal before they went live. So if you have an expectation of £25 million more fees you're going to get, you must have also expected £25 million more cost to come next year in 2019.

So this step-up in IFRS 15 adjustment already implies there shouldn't be a neutrality previously that was going to be a cost increase. Is this the same cost increase or above and beyond that one?

Well, Bruno, I'm starting now to get confused from you. I think the simplest way I'd say is, take what the consensus was for this year, take £40 million off for revenue, and versus the average consensus, take £15 million to £20 million off for additional costs.

Why are we spending more money? It's not that we think the costs of operating a CFC for Kroger, for us, have changed.

It's not we think the costs of operating Casino in Paris have changed. It's the fact that we want to accelerate the pace at which we are innovating the platform, as Tim was saying earlier.

And last, any more international deals that have progressed in the last 12 months? I know you're talking to many people, but is there any material progress going on?

So we're making a lot of progress in a number of places. But as we said before, all the deals we've announced, we won't comment explicitly on any one potential customer or region.

And we've made the mistake once before of putting a time line on saying when the next deal is going to get signed. So we won't do that again.

We're not in a -- we don't feel in an urgent rush to announce the deal because we've got quite a lot of work to build the 23 sites that we've already committed for our customers, which is probably 10x what we said we would do. And we have been working incredibly hard to better increase our capacity so that we can do more.

We feel confident we can do more. Our solutions team has grown quite significantly and are all rushed off their feet working extremely hard in a number of places.

So we do expect to do more deals, but we're not saying more than that. Yes?

Yes, it's Andrew Gwynn from Exane. I guess I have 3 questions, if I can.

So the first one is a bit of an open goal, but talking about your Retail sales growth growing materially ahead of the market. I mean, Tesco grew, I think it was just over 2.5% over Christmas.

What's in your data that is telling you why you're growing so much ahead of the market? Second question, Waitrose-M&S, discuss.

And then the third question, just coming back on Bruno's question. And just to be clear, you're talking about EBITDA of around about £50 million for next year, something in that kind of range?

So I'll take the first two, but I'll be brief on the second. I got to remember the first one now.

Why we're outgrowing the market. Look, we have a big advantage, let's be frank.

We're small. So we're 1% of the U.K.

grocery market. We do a good job for our customers.

As Duncan mentioned earlier, we've seen a significant increase in the retention of our new customers, which is very meaningful. And we're always striving to offer the best value, the best ranges, the best products, the best user interfaces.

And we make just too many changes in a year to just stand here and outline them all, but it's all about trying to make sure that we can attract and retain customers. And it is enabling us to continue to outgrow in the market and gives us confidence that we can carry on doing that, subject to having the capacity to fill it.

The second question you asked me was a discuss. I'm not discussing.

We're in the business of talking to retailers. It's a part of what we do.

We talk to retailers all over the world. And we don't comment on speculation about any one of them and we're not going to start now.

On the second part, which was Waitrose, look, from my memory, I think I've done four contract renegotiations with Waitrose. The last time was in 2010.

The contract has a current end date of September 2020. We've been on 18 months rolling notice for, I can't even remember now what it is, three or four years.

There was quite a lot of speculation for years. The first day that Waitrose could give notice under that contract, they would.

Obviously, they never did. What I can say with great confidence is that, today, we sell over 50,000 products to our customers.

In September 2020, we will still sell over 50,000 products to our customers, including a high-quality, own-label range. Whether that's called Waitrose, Ocado or something else, right now, I can't be 100% sure.

What I can be 100% sure is we will sell over 50,000 products, including high-quality, own-label range to our customers. Yes?

And on the last one. Yes, Andrew, you're right.

The math is right around £50 million, just under £50 million is in the guidance.

So it's Andrew from HSBC. Can I just come back to IFRS 15 points again?

You've talked about a £40 million impact for FY '19. You talked about it being based around five CFCs that you're planning, three for Kroger and Casino and Sobeys as well.

Just thinking beyond that, should we -- given that this is a two-year build process, I assume that continues into FY '20. But then obviously, in FY '20, you're going to be presumably ramping up with Kroger.

So is there a bigger impact there as well? Just to help our planning.

Andrew, good point and a valid point to make. The greater the success we have in signing up more clients means in the early years, the more revenue we're deferring.

So the greater impact it is for us. So success for us, a definition for success for us is that 2020 profits in Solutions are much worse than 2019, that's correct, which has no effect obviously on the underlying value, which, actually, is the complete reverse, obviously.

And just one other as well, you talked about the build of technology headcount going into admin costs as well. You've grown to 1,300.

What's the ambition there beyond that? I mean, is that effectively, you're building a capability to support everything you've signed up now?

Or do your ambitions extend beyond that?

So one, we are still growing because we have taken on a huge amount of commitments in the last year and we expect to take on more. We are probably approaching a peak in our required development rate, but that doesn't mean that we will decide to slow down.

It just means that it gives us more leeway to do stuff that we don't have to do but we choose to do. So we do have a lot that we have to do in the next two years to match our contractual requirements for all the customers without listing them all again, and to get the platform live and resilient, reliable and international and multicurrency and multilingual and multicharacter-set and all the stuff that we have to do.

But we probably will carry on growing it, unless we didn't see an opportunity to use it and generate long-term shareholder value, in which case we would let the numbers go the opposite direction. But at the moment, we're continuing to grow, and we're focused at the moment on growing the existing five development sites that we have and taking them all to a slightly higher number.

We've done a lot of restructuring in technology during the last year. And we're seeing very positive results from that with a significant increase in the throughput of feature deliverability, giving us increased confidence around the deliverability on our contractual requirements and also in delivering the world's best platform to our customers.

Sreedhar Mahamkali from Macquarie. So just three quick questions then, please.

Going back to your innovating for the future slides and the likely possible aspiration bucket. Are you in any conversations with customers in those potentially, likely possible areas already?

Is it that advanced, or still sometime to go?

Yes. I mean, no, we definitely have conversations.

There's definitely interest in us helping people in some of those areas. Sometimes, the debate is whether we want to do it, whether we -- is it distracting or not distracting?

Can we set it up in a way where it's not going to -- if it's distracting, we won't do it. If it's non-distracting, we're looking at it.

And we look at some of these things as spinoff opportunities or partnership opportunities. And so sometimes it's like, Could we get together?

We've got this skill and this IP. Someone else has got that.

Someone else has got that. Move it all over here and let it become its own unicorn.

And so that's the way that we're looking at stuff. But no, we do get approached and we do talk to people about a range of those opportunities.

Not all of them, but a number of them. And we file patents on quite a few of them as well to protect various IP that we have.

As I say, there's no immediate expectation of doing something. But we have organized ourselves so that the people who are good at coming up with those ideas and seeing how the IP we're generating for grocery can do other things can sit separately and aren't distracting their delivery of what we have to do for the core Retail and the core platform business, but also give us the opportunity to monetize those ideas.

In terms of stock [indiscernible], it's gone up again to 3.9% higher. What is the range on the [indiscernible] and how do you [indiscernible]?

Look, I keep saying -- and I say it every year and I'm going to say it again this year, do not expect that number to rise in the short term although it keeps rising, okay? And obviously, it's a reflection of a changing -- effectively, a changing amount of influence about who influences the consumers' behavior.

So over the last 20 years, the CPGs have massively increased their influence over the consumer at the cost of the retailer, and you've seen them make significantly higher profitability than grocery retailers. Because when you walk into a physical environment at a grocery store, as a customer, you've got a wide periphery of vision.

You can see every product that's listed in the store. When you start shopping on an interface the size of a mobile phone, then which one of those products that you want to buy gets put in front of you first, which ones get suggested to you, those are decisions being made by the retailer that has more influence now on what you're seeing, and therefore, you're likely to buy.

And therefore, the suppliers who have spent a lot of money in media with Google or Facebook or billboards or radio or TV needs to spend some of that money with us to make sure they get the right product in front of the right customer at the right time. We won't let -- we won't take a supplier's money to put the wrong product in front of a customer because that influence -- that would affect our long-term value and our customer experience.

But often, there are 2 competing products that a customer might want to see, and therefore, we can play a part in deciding which one of those 2 to show. I think in the short to medium term, that number is unlikely to grow, but I said that last year.

I think in the long term, that number is likely to grow.

How much of that is pure advertising?

Pure?

Advertising.

We don't split it out, but it refers to a whole host of activities that we do with our suppliers from physical sampling through to buying a banner on our site. But we only sell a banner on our site to our suppliers.

So we never advertise anything on the site other than beyond the point where we confirm your checkout. But before that, we'll never put -- do anything that would take you off of our site.

So we don't, like, put an advert for a car up there and you go, "That looks nice," click it, and you forgot you're in the middle of grocery store. We don't do that.

So it's all from our suppliers for things that are being sold on our site. But we don't split out exactly where it's coming from.

It is unrelated to volume or transactions. That's the difference between it and overriders or retros or that type of thing.

Robotic picking, what are your expectations in terms of cost improvements over time?

Look, in the long term, they're very significant because the main thing is it's very, very hard to program, effectively program, I would describe it, to teach the robots how to pick groceries because they're not uniform. They're not coming in neat uniform trays, ready to be picked.

If you do that, you can pick everything. If you take a tray and you divide it all equally, and you put one here, one here, one here, but you actually -- and you can see them do that in trade shows and say, look, it's an amazing machine I've got.

And I would say, well, okay, how long did it take you to make the tray for it to pick from? Well, I can pick it faster than you can make the tray, so what's the point?

So actually teach them and -- is the hard part. Once you've taught them, the replication of it is cheap.

So a robot vision system and gripper on its own costs about the same as, roughly, give or take, as having a person for all the operating hours that you would normally on average use a picked location for a year, all right? So in time, we'll have a very significant return.

And we'll offer it to our customers, obviously, as an add-on service. The issue is about training the robots to be capable of handling 50,000 products, not just in terms of picking them because picking them is easier than the placing them, because you've not just got to pick it.

You've got to put it in the bag with the other items and not just break, damage, et cetera, the other items, right? Okay.

Sounds like this is the questions corner.

It's Dan Ekstein from UBS. A question on your general merchandise business.

It seems like the holy grail for online grocery retailers has always been to kind of add these high-margin, general merchandise categories potentially so that the delivery can have a big impact on economics. How do you feel you have progressed over the past year or so in terms of uptake in general merchandise?

And then when you think about the pipe you've got from fulfillment center to the customers' home, would you ever anticipate a scenario where that could be opened up to potential third parties? Or do you think it's best unenhanced?

So the first part is we -- our GM, which is a combination of what we call the general merchandisers inside the grocery store; plus, the destination site volume outgrew the grocery business again this year. So it's becoming a larger percentage of our business, and therefore, it's adding value in that way.

It's not always higher percentage margin, by the way. Some of it is and some of it isn't.

But it's usually always higher cash margin per kilo or per liter. So in terms of thinking about what it's -- if you've got a full van, what it's muscling out of the van, it's usually got more cash margin involved in it.

In the future, could we allow third parties to utilize our delivery service? It's a possibility.

It's one of a number of strategic ideas that people have, all of which require software development, and you have to just prioritize what is it that you're trying to do. Do you want to try and get your first warehouse in an international territory remote from your networks live?

Or do you want to open up your service to allow third parties to do a just-in-time delivery and put them on your van? But it is something that we could do in the future.

It's -- we don't have any immediate plans to do so. Anyone outside the question corner?

Well, no one in question corner. No?

Great. Thank you for coming this morning.

Thank you.