Ocado Group plc (OCDO) Earnings Call Transcript & Summary

Well, good morning, everyone, and welcome to the Ocado Group Fiscal Year '24 Full Results Presentation. A new face to you here. I'm Adam Warby, delighted to have joined as the new Chair of Ocado Group. I'd first like to start by just paying tribute to my predecessor, Rick Haythornthwaite, and Rick's leadership -- strong leadership and counsel to both the Board and to leadership was certainly extremely valuable during his tenure, and we wish him luck in his new endeavors as he moves on. For me, the first few months have been absolutely fascinating and engaging. This is certainly a business with visionary leadership, world-leading products and a lot of depth of talent. And of course, we've got outstanding partners that number some of the largest and most prestigious names in the global sort of logistics and grocery business. And these are credentials that any business would be very glad to have. And of course, we're determined as a Board and a leadership group to make sure that we take advantage of that strong position. So today, we're going to be reporting a business that's made good progress on its financial results with strong results across the different -- and profitability across the different reporting segments. And of course, from a product perspective, we've reached important milestones in getting the reimagined products deployed in a number of our partners. We're seeing growth across many of the global CFCs. But of course, we recognize that, that is behind our expectation and the expectation of partners. And that's the key focus of both leadership and us to make sure that we engage in accelerating that growth, driving more efficiency and ultimately driving the benefits that our partners expect out of our platform. So I think in the year ahead, leadership is certainly going to be focused on that on driving increased efficiency, increased performance from our partners. We're deploying lots of resources to that, improving our existing offering and additional growth. I think that's what we all want to see through both new markets in the grocery world, but also outside in the non-grocery markets as well. So look, I'm very happy to be here. Lots of work to do. But for this presentation, I'm delighted to hand over to Stephen, who's going to take us through the results. Thanks, Stephen.

Great. Thank you, Adam. Okay. Good morning, everybody. Thanks for joining us. Okay. Next slide, please. Last time I'll be saying that. Okay, financial progress. Good financial progress across the group. Headline numbers, revenue, Adam alluded to this, growing at 14% to GBP 3.2 billion group revenue. Adjusted EBITDA traveling to GBP 153 million, up GBP 102 million in the top right there. I think a number -- a very good number and a positive number is the improvement in underlying cash flow, a GBP 249 million improvement year-over-year. We guided at the start of the year to GBP 100 million improvement. We raised that to GBP 150 million and finished the year with a GBP 249 million improvement, putting us in a really good position on liquidity, healthy liquidity of over GBP 1 billion, which includes cash on the balance sheet and the access to the revolving credit facility. Worth pointing out, and I get into the detail of this a little later with the strong second half cash flows benefiting from upfront fees from some of those reimagined orders and the CFC design fees that we have for the new CFCs that are going live. We also saw a reversal of the working capital outflow that we had in the first half of this fiscal '24 financial year. Next slide? Here's the run through from revenue down to loss before tax, revenue and EBITDA across the group. You can see each of the business segments there. I'm going to go through the detail of each of those business segments, so [ I won't dwell ] on that performance. Depreciation and amortization at GBP 460 million. Now that's at a peak level. You should start to see that flatten and ultimately decline as we're putting less capital into our asset base. You're going to see that shortly. Finance income stays about the same. Interest costs will creep up somewhat with the introduction of the high-yield bond that we introduced in July of last year and the coupon that's attached to that. I'll go through that detail later. And then other finance gains and losses is the GBP 10 million there. That's the increase in the valuation of our investment in Wayve. That's the key item in that particular number. And in the adjusting items there's a number of items in the adjusting items number. The 4 key ones that I think stand out to get to that GBP 5 million number, we benefited from a GBP 44 million gain from the refinancing exercise that we carried out in the summer of last year. We deployed GBP 650 million of cash to buy in GBP 700 million of gross debt. So there's a benefit there. Then we also had the Dagenham and Coventry Spoke sales, there's GBP 12 million from that. On the other side of that item, though, of costs that we've got within that number, we've got the GBP 23 million we've invested in our logistics business around the finance, IT and HR systems to improve that business as it increasingly becomes more of a stand-alone business and less reliant on the services from Ocado Group. And in there also is the GBP 29 million that we've written down our contingent consideration due from Marks & Spencers down to zero to take the most appropriate and conservative accounting treatment for that particular item at the year-end. More on that later. So running through our businesses, Technology Solutions; strong improvement in EBITDA, as just highlighted. Starting at the top there, average number of live modules, the key driver of the revenues and contribution from this business, growing to 116 average modules during the course of the year. That generates that recurring revenue number of GBP 416 million, and I'm going to show you some more detail on that shortly. Revenue of GBP 496 million when you take on the recurring revenue plus the previously received upfront fees that we amortized to the P&L account over time. Direct operating costs, I'm going to talk through that number. That's flat at around 30% of revenues, reporting a 70% contribution margin. We expect that to grow to 75% over the next couple of years, more on that shortly. Technology costs of GBP 93 million, pretty much flat with last year and our support costs coming down materially by GBP 17 million to GBP 174 million, delivering that GBP 81 million number. It's worth highlighting that within the Technology Solutions EBITDA number, there is GBP 10 million of cost overhead attached to OIA, Ocado Intelligent Automation, which is not yet revenue generating. I'll talk about that in a second. And there's also GBP 6 million of Jones Food losses within that number as well. So there's a GBP 16 million drag on those numbers, which we would hope in both those instances to reverse over the next 2 or 3 years. Costs well controlled. I won't go through the detail on what's on this slide here because I've got a few slides that are just going to go into that cost reduction in more detail. So recurring revenues, now representing 84% of total revenues for Technology Solutions and driven by that average live modules evolution that you can see on the left-hand side of that slide. The annualized recurring revenue per module now at GBP 3.6 million as an average in fiscal '24 will grow as indexation kicks in for each of our partners, but also will grow as we add more reimagined kit to our partners' portfolio of technologies that they're licensing from Ocado. We saw a lot of reimagined orders in the second half of last year across Kroger in particular, and that's another driver of those upfront fees that we got that I talked about earlier that benefited our cash flows. I'm going to talk in a little while about what the outlook looks like for recurring revenues as we look forward as we add to the module base from the 123 live modules that we had at the end of the fiscal '24 year. CFC go-lives and drawdowns will drive that module growth. So here, where we are at the end of fiscal '24, 123 live modules. You can see 69 of those in the U.K., 54 internationally. We're calling out today that we expect to get to 150 live modules on average at that year-end over the next 3 years. So that's an increase of 27 from where we are today. Around half of that number will come from the 7 CFC sites that we have good visibility of going live over the next 3 years, in '25, '26 and '27. Warsaw in '25. Phoenix and Charlotte are now going live in early fiscal '26 rather than in '25 as we'd earlier anticipated. Kroger have placed an order for the Auto Freezer for each of those 2 sites. So we're pleased by that order, but it does push the CFC go-live out by a few months into early '26. We also have 2 sites going live for Lotte in Busan and Gyeonggi across '26 and '27 and then 2 for Aeon in Hachioji and Kuki, again in '26 and '27. Those are the 7 sites. So together with those 7 sites opening, we typically open with 2 or 3 modules per site and with the drawdowns on further sites, we expect to -- we hope to actually beat that 150 number. But today, as it stands, it's a good guide for fiscal '27. So that continued cost control that I talked about. Our direct operating costs, we saw made great progress in fiscal '23. Fiscal '24, still at that 30% of revenue number, we are actually adding cost to direct operating costs as we're increasingly utilizing modules that were live last year, but underutilized, if that makes sense. So there's an increase in costs that are attached to those direct operating costs. We do expect, though, to get this down to less than 25% over the next 2 or 3 years. That's consistent with our less than 1.25% previous guidance of live sales capacity. Our technology spend is reducing as we complete our reimagined innovation. So in total, we spent GBP 280 million. We spent that in fiscal '24 across Technology, GBP 93 million of which we expense and GBP 197 million we capitalized as R&D. Our guidance for fiscal '24 for this number is GBP 250 million. So we'll reduce that by GBP 39 million from GBP 24 million to GBP 25 million. We'll come to a slide shortly that will also highlight that by fiscal '27, we are targeting for our tech R&D spend to be around 20% of recurring revenue. So that's how we'll be managing our tech R&D spend in that context. If recurring revenue grows, we'd be ready to invest more, but that's the target that we're looking to allocate our capital, 20% of recurring revenue towards tech R&D. Support costs. Very good progress made on support costs, GBP 208 million in fiscal '22 and now GBP 174 million. We've highlighted there the corporate functions in particular, and how we've driven down costs there. I think I was chatting with Tim the other day, and I thought we'd made very good progress, but he thought we could do an awful lot better than this over the next couple of years. That's the plan. I think there is opportunity. We're engaged with -- well, first of all, we've employed a third-party to look through our procurement activity and contracts in particular. So at the moment, we're retendering a number of contracts and getting good benefits there. We have employed a new head of procurement for the group and appointed our brand-new first ever head of indirect procurement as well to drive out costs in that space. So a lot of activity going on there. We are at the moment in the middle of a search for a third-party to engage with us on bringing down those support costs generally across the group. More on that later. Ocado Logistics, remains a strong, reliable generator of EBITDA and cash flows; revenue growing by 8%. Encouraging performance in logistics. I think the key thing to highlight is that costs growing at 8% have grown less than the volumes that they're processing, a 12% growth in eaches, 11% growth in orders, but costs increasing by only 8%. Key driver of this units per hour, improving at 9%. Units per hour, the higher that is, the lower the labor component, the more productive we are, the more efficient we become for our partners in providing that service to them. And here's a chart that just summarizes that improvement in UPH, 227 for fiscal '24. Worth highlighting, by the way, that Luton CFC with On Grid Robotic Pick now in place for a decent proportion of their volumes exited with a UPH of 269 so a strong performance there. We should expect to get even better. On the right-hand side, DP8, deliveries per 8-hour shift, a bit of a headwind there with the transition from the Hatfield site to the Luton site and slightly longer delivery times as a consequence. Ocado Retail, a terrific year for this business, revenue growth of 14%, remaining the fastest-growing grocery in the U.K. We're seeing that sort of growth in the early part of this year as well. So strong performance, and that's a continuing strong performance, delivering an EBITDA of GBP 45 million, and that's an underlying EBITDA margin of just less than 3% when you add on those Hatfield fees that we're charging. So good performance in retail. I think let me go through some of the core metrics here as well. Average orders per week, up 12.5%. Frequency is driving orders per week as well. There's an improved frequency. The customer is now shopping, I believe, for the first time in less than 14 days for the next order. So the gap between orders is reducing. Frequency is improving quite significantly. Customers now well over 1 million customers, growing by 12% over the course of the year. Basket value stable at about GBP 122. Average eaches per basket, again, stable at 44 items. The average selling price nicely is -- this is not a price-driven performance. Average price is at only 0.4%, significantly below U.K. grocery market inflation of 3.0%. So it's volume-driven, customer-driven, order-driven, frequency driven. Those are just some of the key metrics that I've run through. Worth calling out on the right-hand side, you can see the low levels of capacity utilization we've been at -- sorry, on the right-hand side, we've been at in '22 and '23 and improving now materially to 83%, covering those fixed costs and delivering more margin to the bottom line as revenue grows. An accounting point to call out an item that was part of the original JV agreement back in 2019 that between 5 to 6 years from the JV contract being struck that consolidation of the numbers would pass from Ocado to Marks & Spencer. That is now taking place around the -- well, around early April to coincide with the Marks & Spencer year-end conveniently for them. And that's -- we're working towards that date is an intent. Due to that intent, we are now required from a statutory basis to report our interest in Ocado Retail as a held-for-sale asset. You'll see that we're showing there for the sake of comparison for this year's numbers, a total group perspective to compare '24 with '23 on a fully consolidated basis. Key point to call out, no significant changes to the decision-making rights or governance of the business and no changes to our economic interest in Ocado Retail or indeed any of our intentions around Ocado Retail. It's an accounting entry and more on that later. We're going to be running an accounting seminar for those of you that are interested in early May to share with you some more specific numbers and accounting so you can see how we're looking at this. In short, though, there will be a single line in our profit and loss account in our income statement as a share of profit after tax for our share in Ocado Retail. On the balance sheet, we'll have an investment in JVs and associates, again, another single line, and we're measuring that at fair value at the point of deconsolidation. That's an exercise that's work in progress. The cash flows will reflect any cash flows to and from ORL as if it were an independent business. And that will be either dividends paid to us when Ocado Retail has a healthy cash balance and is routinely generating cash flows or it could be loans to them in the event we're funding CFC capacity build. So the Group, the cash flows for the Group, taking you through from EBITDA through to underlying cash outflow. Here's the GBP 249 million improvement year-over-year, a very strong second half. If you look at the splits there, EBITDA, some growth, as you might expect. Movement in contract liabilities, GBP 76 million versus GBP 21 million. We benefited there from orders from cash receipts from Kroger from Coles. Coles, those 2 CFCs that went live, the final portion of their design fees; Aeon CFC3, the upfront fees from them; Kroger from their reimagined orders. So strong growth in those cash flows in the second half of the year. The GBP 30 million in working capital, that was just a straight reversal of the outflow that we had in the first half. And then capital expenditure; capital expenditure coming down reasonably materially as construction activity is less given the amount of CFCs that we've been constructing and going live over recent years. Reported cash flow of just GBP 113 million outflow. We had the GBP 100 million AutoStore settlement receipt. As a reminder, we received GBP 42 million in fiscal '23, GBP 100 million this year. And bottom right there, we've got GBP 58 million to go in fiscal '25. Here's the refinancing discount. We bought in GBP 700 million of debt for GBP 650 million. That's the amount that we captured there. And then other adjusting items is a mixture of various, for example, disposal of the Dagenham and Coventry spokes, consistent with the message that I shared earlier on the expense side of the adjusting items. Next slide? So here's the outlook for cash flow. This shows the trend that we've been on over the last few years. These are the reported numbers, and we're expecting this trend to continue. I'm going to go through now the key drivers of that delivery of cash flow positive in fiscal '27. So contribution growth, this will be driven by growing number of modules, growing to that 150 live modules by the end of fiscal '27. As a reminder, they generate just short of GBP 4 million per module. And by that year, fiscal '27, we expect to be turning around 75% of that into contribution. Our fixed costs are across technology spend and support costs in fiscal '24, a total addressable cost base of GBP 463 million are expected to come down materially. We're coming out of a tech R&D cycle. I gave you that previous guidance of around 20% of recurring revenues. And I alluded earlier to the opportunity in support costs as well. So there's going to be a significant reduction in those fixed costs over the next 3 years. Similarly, on the right-hand side, Ocado Retail will go to a high mid-single-digit EBITDA margin as it improves the CFC utilization levels. Logistics remains a stable cash generator for us, and we'll start to see those cash flows coming through from OIA as it wins new business, installations go live and generates revenues and profits and cash flows for the Group. R&D, just a little bit more on that, GBP 289 million of total cash spend in fiscal '24. I gave you that earlier guidance that will reduce to GBP 250 million in fiscal '25 and then will decline over the next coming years. Ultimately, the guide for us there will be that 20% of recurring revenues as to where we land on our technology R&D spend. On the right-hand side there just gives you a bit more insight as to how we're spending our CapEx today. You can see CFC Technologies is the bulk of the spend and then e-commerce and then last mile, doing this as natural R&D, but also often at the request of our partners as well. Support costs, I've talked about this, no more than GBP 150 million in fiscal '27 across corporate and partner-facing teams. I think we can do even better than that actually, but let's see where we get to. Strong liquidity at the year-end, GBP 772 million of cash in the bank and the revolver. We've managed our debt maturities in an orderly way over the last 2 or 3 years. We expect to do the same in respect of our debt instruments as they stand at the moment. I thought we carried out a pretty effective refinancing exercise with some good discount capture in the summer of this year, and we'll continue our recent track record to ensure we stay ahead of the game on our maturities and the addressing of them. And here they are. This particular item later this year in December '25, if we wish, we can draw that from existing cash balances. But of course, we may also be considering refinancing some of the other debt that's in the pipeline. Guidance for fiscal '25; Technology Solutions, we've got 10% growth. I call out in here, by the way, one point of detail that is worth highlighting, the McKesson contracts, there is -- we're not expecting this of all contracts to be in this way. McKesson contract due to a software arrangement attached to it will now be recognized over 3 years that revenues and profits rather than a single year. So what we thought would be a full year of revenue and profit recognition in '25 will now be across '25, '26 and '27, which has some impact on that 10% revenue growth for Tech Solutions. Reporting a 20% to 25% EBITDA margin. Let's see the progress that we can make there on that contribution margin, but also on bringing support costs and tech costs down. Logistics, no real change in what we're expecting for our Logistics business. Ocado Retail, we're looking at above 10% growth. We have a strong start to the year and with a circa 4% underlying EBITDA margin for that business. Underlying cash flow, we're reporting a circa GBP 200 million outflow in fiscal '25 and CapEx of GBP 300 million, and that is CapEx across CFC CapEx of around GBP 130 million or so and then technology R&D of around GBP 170 million. That's it from me. And now over to Tim.

Good morning. Thanks, Stephen. Thanks, Adam. So online grocery, just as a reminder, is still the fastest-growing channel in the global grocery market, driving a big share of the organic growth of the market. This slide just illustrates a few of our clients and their competitors and just showing the online growth in blue versus their total growth in green. So moving on. The key to our technology, why are we here is that our technology drives the best customer experiences and does that with positive economics at scale. You can see some of the reactions. And in fact, Coles, they were reporting last night around a doubling of their perfect order rate around record NPSs and a fast early ramp-up in the 2 facilities that we spoke about earlier that they opened in the last year in Sydney and Melbourne. Talking about EBITDA margins there, traditional offline grocery margins illustrated and the reality of online picked orders from store being negative. Today, adding Ocado Retail and Ocado Logistics, most of our clients run their own logistics together, starting to show small positive EBITDA again. But over the medium term or the midterm, that will rally to a point ahead of the stores in the U.K. market is our strong expectation. Just as a reminder, why is that? And that's because the conventional way is running from suppliers to RDCs, from RDCs to stores, from stores to home and our approach allows a significantly shorter supply chain from suppliers into our CFCs and directly to customers' kitchens from there. Our proprietary automation obviously makes this radically more efficient. On top of that, our software solutions are generating a huge level of efficiency with some extraordinary technology deployed, and we wanted to give you a little bit of a flavor of what some of that that is going on behind the scenes. Our web shop is generating more than 10 billion product predictions a day to enhance and personalize the customer experience. Our supply chains are making more than 70 million forecast a day to make sure our clients know with pinpoint accuracy what is needed in their warehouses to drive availability, but at the same time, driving down waste to the lowest levels in the industry. We're using digital twins across the business to optimize and derisk our CFC inventory management. Over the last 12 months, we were able to simulate 270 years worth of CFC operations across our [ DASH ] software, which controls our bots and our inventories. In last mile, we're making over 100 million optimization calculations per second to drive the highest levels of slot availability whilst also driving efficient delivery routing. All of this is underpinned by our IP estate, which is 10x larger in terms of patent filed than the rest of the U.K. grocery industry combined. What does it mean for us and our partners? At a core level, it means time and cost saved. Today, an Ocado CFC is capable of fulfilling a 45- to 50-item order with around 10 minutes of labor. We estimate the equivalent process in the store network would take about 70 minutes. So it's an hour of labor saved for a single order. And this is at a time when labor costs and labor availability are one of the major pressures facing retail supply chains worldwide. Just to reillustrate that point in another way, here is a little bit of history. The average time it's taken for Ocado to fulfill an order in 2012 was 25 minutes of human endeavor, and it is now down to 10 minutes. At the same time, minimum wages across our client markets, and this is just minimum, not actual wages have risen by over 75% with real minimum wages rising around 15% as compared to the price of groceries that has underperformed those indexes. So the true cost has been rising, but in an Ocado warehouse, the true cost has been falling and is now less than 1/7 or around 1/7 of the human cost of fulfilling these orders in a supermarket. In the last few years, we have been through a particularly intensive round of R&D to extend our leadership in this space. And our latest automation that we've spoken about, our reimagined technologies are now being ordered and rolled out by the vast majority of our partners worldwide. We had last updated this slide yesterday, and I've been told this morning, it is outdated. We need to put one more of those blue ticks on On Grid Robotic Pick [indiscernible] that got signed last night. You can see here a significant rollout going on through 2025. I'm going to take a moment to look at some of those innovations. In particular, On Grid Robotic Pick. We're today using On Grid Robotic Pick in Luton and it is doing just around 40% of our ambient pick and about 37% of our chill pick and that is a number that is growing week-on-week. The first thing to say about this is it's genuinely boundary pushing for AI and robotics. It uses a type of AI called Behavioral Cloning and is pre-trained and leveraging reinforcement learning. This is pixels to action with no code. Secondly, it's doing this in a chaotic physical environment. There are circa 30,000 relevant items to be picked, innumerable potential variables to be adjusted for in the process of picking and packing, all within a highly active industrial robotics environment. Thirdly, reliability and accuracy are key with zero margin for error versus human picking. If the error rate exceeded human picking by even a fraction, the economics of doing this by robots would fall down. The same is true for packing density. If we generated more boxes as a result of Robotic Picking, we would spend the savings in incremental delivery costs. But if you can get it right as we have, the productivity gains are huge. And if we can exceed the accuracy of humans and exceed the density of their packing, there are further gains to come. And obviously, they're repeatable across all our clients. Finally, enabling technologies like these to operate seamlessly together in the smallest possible footprint is a huge feat of engineering. We've massively reduced the space and cost required to operate these arms with innovative engineering solutions. In 2024, we picked over -- sorry, over 30 million items, ending at about 1 million items a week and saw huge productivity gains with just a small number of arms installed. Over the next year, we expect to scale rapidly and should end the year picking 1 million items a day. Our experience operating this technology at scale in these environments will be another huge differentiator, enabling us to look to increasingly generalizable systems. And even in our existing CFC operations, we continue to make new efficiency gains. By the end of 2025, we expected to have reduced the number of bots required to manage $100 million of sales by a further 15%, having seen significant improvement in the last year already. These bots are already market-leading in their work rate. Ocado's fleet of around 17,000 bots covered 153 million kilometers in 2024. One of our large and well-known competitors recently announced their own statistic, and it is clear that we traveled a slightly greater distance using 1/4 as many bots. Therefore, the average one of our bots is doing 4x the work rate of our nearest competitor in this space. I want to move on and talk about our global CFC operations. We continue to learn many valuable lessons. Internally, we think about these in 3 categories. The largest category is the growing well. They include sites that are growing but with some improvements needed. And finally, a small number of sites -- sorry, the largest categories are growing well and the next category is the growing but with improvements needed. And the third category is a small number that require a different strategic approach to generate their full value. Moving on to the growing well category, which is the largest category comprising a number of CFCs operating at or approaching profitability. These sites are located in good locations with good customer densities in target demographics with robust catchment strategies. Generally, the partners have got good logistics operations in place with more efficient labor planning with high volumes processed across the day and with cost-effective upstream supply chains and consistent high van fills. We're generally seeing good customer acquisition at these sites, and they're trending towards being full. Of course, there remains room for continuous improvement and optimizations, but these sites are tracking well. There's an example here on the right of one that's been open for a few years that's now at 74% max utilization and bumping right up against its drawn-down capacity. There's an example of one that's been opened in the last year that has already grown up to being nearly 40% full only in a matter of a few months, is trending towards filling up its drawn-down capacity. And if it can continue at this rate, we'll obviously not be that long before it continues to draw down more capacity and reach its max capacities. Moving on to the next category, the second largest category are sites that are growing but need more improvements. They're a key focus of our Partner Success teams. They may be a mixture of early-stage sites requiring support as they scale quickly or they may be more mature CFCs that require some support in specific areas, and they vary not only between clients, but also within clients' own portfolios. Remember that for some partners, operating online business at scale means developing new muscle. That's why we've embedded growth experts and experienced operators. The chart on the right is to illustrate some of the ways that our teams are helping them to work. This is a client who opened a site and by month kind of 10 to 12, we're seeing declining sales volumes. That's the point at which our customer growth experts went over to help that client and got a 2.5x growth in that facility over the following 6 months. These are examples of some of the work that we're doing with our global clients. Lastly, we have a small number of sites that we believe require a different strategic approach. Whilst our technology in them is operating at a consistently high level, there are some reasons why those partners are struggling to scale the volumes as we and they would like. We're opening up an exciting range of new use cases for those partners to help tackle these challenges. For instance, we're widening the ways in which partners can shop for CFC fulfilled orders, not just directly from a retailer's website, but opening up integrations with marketplace platforms. We're also enabling the CFCs to serve a wider range of customers or a wider range of use cases for those clients, including B2B as customers or acting as upstream sites for store replenishment for small stores where single pick is a big advantage or for the tail of range in large stores. So while we're sitting with these sites, we're confident of generating good long-term value for them, for us and our partners. Lastly, I want to look at some of the things that we're focusing on in 2025, which is a small number of very-targeted innovations that are going to deliver outsized and long-term benefits for our partners. We want to open up more growth opportunities and flexibility for the shoppers. We want to automate more of the logistics planning to help our partners run more efficient operations, and we want to deliver a step change in capital efficiency for future CFCs. So let's have a look at one of those ideas. We are now starting to roll out OSP Swift Router, following some successful early trials of the first part of the technology in 2024. In essence, the product allows CFC to radically increase the number of orders it can serve in short lead time and same day without sacrificing the significant economic benefits of large automated fulfillment and large van route deliveries. You can see the chart here on the left, how weak this might impact the general shape of the day at a typical site. So this is in terms of what we're producing. You can see that some of the production is for same day. Some of the production is for the short lead time 1- to 6-hour slots and some of it is for next day. But you can see the balance of it is more same-day and ultra short -- short lead time rather than next day that it has historically been. This will help new shoppers to convert to our clients' platforms. This will help increase the total number of orders. It will help existing shoppers take advantage of more options and short lead time orders deliver the next day with better economics. We will also be later in the year adding the ability for existing picked customer orders that have not yet been shipped to have an add to option, a last-minute add to as well, which will help drive basket size and client profitability as well as shopper satisfaction. This is in rollout in a couple of CFCs right now as we speak. Another thing that we're doing is making our systems easier for our clients to operate and to operate better. A couple of the things that we're doing here. So we're removing the complexity from the planning and execution of the logistics processes to reduce efficiencies and as I say, reduce the amount of people required to run these. A couple of examples, automatic route release. This essentially allows our retailers to set a dial between how much demand they want to achieve versus how much efficiency they want to demand. And then we use AI to automatically release routes into the schedule, a process that's been done for the last 24 years manually by humans when they were available to do it and without the accuracy that the AI is possible to achieve. Another one that we will release this year is dynamic finalization, which essentially allows us to pull work into the CFCs in the last minute in a way that doesn't interrupt the efficiency or the wider operations of the CFC, but does hold open slots for last-minute orders and to be able to serve them with the economic benefit of a next-day order for far longer than we could manually. This is an extraordinary accomplishment that we have made over the last 24 months which is that we're delivering a step change in the capital efficiency of future sites to the extent that if we were to rebuild sites that we have in the last 2 years, we would today build them on just over half the footprint we had, we can halve the size of the warehouses. This obviously generates a huge amount of benefits to us and our partners, including lower rent, lower CapEx. It enables more direct delivery because the sites can be closer to dense urban populations because smaller sites are more readily available and obviously offers a path to a faster breakeven. And that comes with an additional 45% uplift in productivity from our reimagined technologies. And that number actually is growing since we put this deck together based on some new products that OIA will be launching in the coming months. So to conclude, Ocado continues to deliver standout solutions to the labor productivity challenge in global supply chains. We're seeing growth across the vast majority of our global CFCs with Partner Success working to accelerate volumes in some sites and widening use cases in others. We're delivering exciting new innovations in 2025 to continue building volumes, enhancing CFC operations and driving a step change in capital efficiency for future sites. By the way, on those sites, it's not just that we can now build them half the size that they were before for the same throughput, but we can go back into the historic sites and gain 10%, 20%, 30% or more percent incremental throughput than they were originally designed for. And in fact, I was in Luton yesterday, and it will exceed its original design size by something like 5% this week already. We are delivering exciting new innovations in 2025 to continue building those volumes. We're delivering a strong -- we have delivered a strong set of '24 results in line with guidance, revenues, profits and cash flows all improving significantly. Okay. We're going to take some Q&A.

Difficult questions to Stephen, easy questions to me, please.

Hi. Tintin Stormont from Deutsche Numis. Two questions for me, and hopefully, it will come back again. In terms of OIA, I appreciate you talk about the building pipeline. What's the holdup there for winning new contracts? Do you think it's just the market? We've seen obviously, competitors that have reported revenue declines in this area because the market is tough at the moment or do you -- where do you think your pricing and relationship with partners, which tend to be important in this ecosystem, where are they at? And then secondly, you didn't talk about it, so I'm assuming it's not important or it's not material, but tariffs -- how have you thought about tariffs going into the U.S.? How should we think about it? Any mitigations there? And then just in case -- just in case you're bored and you want to answer it, cash flow without ORL, the GBP 200 million guidance, but without the ORL, once it reconsolidate, what does that number look like?

So OIA pipeline, there's a lot going on. Would we like to have signed some work that we could have told you about? Yes, let's just watch this space. There were some -- you are rightly alluding to the fact that when interest rates rose the way they have and some of the things going on in the global economy, there were some projects that we were told we were going to be awarded and clients have just put them on hold for 12 or 24 months. We still expect to do them. But at some later point, obviously, that's disappointing because at this early stage, it would have been nice to have signed them and done them. But they've got a big pipeline. There's a lot of interesting projects being worked on. And I would just say watch this space and let's see what they announce in the coming months. Tariffs is an excellent question. And it's a really difficult one to answer because you don't -- we don't quite know what the stable situation on tariffs is going to be because one minute there is a tariff from somewhere, one minute there isn't a tariff and somewhere. It seems to be moving around. We are scaling some sites in the U.S. for our client. We are finishing the bills as we spoke about of [ 9 and 11 ]. And because we switched from a third-party freezer solution to our own freezer solution that is what caused the delay in contracting. It got signed in January. They will open around January next year. I think almost all of the equipment that we need to do both for scaling next year and to build those automated freezers and get those sites live are already in the U.S. So it's not -- there isn't an immediate question for us in terms of bringing things into the U.S. that we are aware of. And tariff is a very complicated area in terms of what it does apply to and what it doesn't apply to based on where it's assembled and what the -- where the parts are from and stuff like that. We did do a lot of work as a plan C when we were involved in IP wars, and we were challenged at the International Trade Commission in terms of trying to block our importation rights into the U.S. So we do know that we could build in the U.S. if we -- if that was more cost effective than building in Mexico and shipping into the U.S. And it's something that we're obviously, like every company, watching to try and understand and hopeful that we will be able to manage and mitigate our way around it. I don't know if you want to answer the third question.

We can do that one as well. We're not actually assuming anything material from Ocado Retail in respect of cash flows either in fiscal '25 or in fact fiscal '27 outlook. We are expecting Ocado Retail to recycle any cash flows that it generates into adding capacity where it sees the growth so no heroic assumptions on retail in our numbers.

James Lockyer from Peel Hunt. Two questions from me, please. So just on the U.S. warehouses there. It looks like you said it's been delayed at sort of a few months there. So logically, I imagine it has -- does nothing really to the lifetime value of those warehouses given it's a couple of months. But with the addition of the Auto Freezer sort of part of that, does that increase the take rate potential and therefore, actually increase the lifetime value of that -- those warehouses?

James, thank you. We built one third-party freezer in Luton. We're now looking at being able to do double-digit increases because of the efficiency that we're driving from our bot fleet because of the efficiency we're driving with On Grid Robotic Pick. We can already talk -- we are already talking to our U.K. client, Ocado Retail, about how much we can ramp up the capacities in Andover and Purfleet. We have got more work to do in Luton to be able to match that because it is significantly harder to scale up the third-party freezer to the volumes that the rest of the site can grow to. So there is definitely an advantage of putting in our own freezer that will scale as the rest of the build -- as we work out to drive bot efficiency and bot algorithms more effectively, we'll get the same uplift in the freezer that we would. And therefore, yes, if we can scale those sites from whatever number of modules they were originally designed to 30%, 40% more, it's going to be easier to do with the new freezer that we delayed -- the new Ocado Auto Freezer that we're delayed and having that would have been if we put in a third-party automated freezer.

And then secondly, just your philosophy around take rates. So I think -- so it looks like it was about 4.8% in the year, just gone up slightly. But given the timing of rollouts that you see, do you foresee '25 being at least 5%? And then historically, you spoke about a 5.5% aspiration based on reimagined. But obviously, you've been working with clients closely, what the processes they go through to actually run these solutions. Are there other services that you've sort of spotted or think you could be able to do for them, which could drive your take rate higher? For example, simply by given your scale, could you negotiate cheaper payment processing fees, for example? What's your aspirational around sort of long-term midterm take rate?

I think our long-term midterm is still to try and drive from -- to 5% and then from 5% and above. There's a mix of clients in there and some people have got some earlier kind of more advantageous rates. And then the -- as we add in, as people draw down Robotic Pick, as people draw down Automated Frameloading, as people draw down more of the things that we produce and they're paying extra for them and gaining significant benefit though from the labor that's coming out. On something like AFL, I think yesterday was the first day that we did 100% of Luton volumes, for example, across Automated Frameloading. So it can be a bit of a binary. We started installing them only a number of weeks ago. And we've -- while during live production, installed a whole lot and migrated it all over one machine at a time. On the On Grid Robotic Pick, it's -- the revenue is going to grow over time as we can successfully pick more and more of the percentage of the volume of the warehouse. So the take rate of what we're earning from Ocado Retail in Luton will double as we move from the average penetration of high 30s to an average penetration over 70%. So yes, we would still look to try and get towards the 5.5% long-term target.

Sarah Roberts from Barclays. So just 2 from me. Firstly, I think the news of the Kroger site slipping into '26 is a little bit of a surprise this morning. Obviously, it's positive that they're investing into the new technology. But just want to understand your thoughts on could there be further downside risk to the pipeline between now and '27, particularly if there are sites who maybe -- might start investing into some of the newer projects before they kind of want to take the plunge and go live. And then secondly, I just want to touch on the refinancing. I know you briefly mentioned that you were kind of thinking about.

I just want to pause for a minute and just try to do the technical one first. The sites going live between now and '27 have most of these products built in as go-live products. So they have -- if they're having Auto Freezers, they've got Auto Freezers in the design from the get-go. The difference in Phoenix and Charlotte is when we first did the designs for those, it was when we were doing the designs for Luton they had a similar freezer solution going in there. It's a third-party freezer solution. It was disappointing in its cost and its throughput. And when we turned the site on, it wasn't as good at ramping as our own technology. So we went -- when we -- while we were developing our own technology for the future, other issues delaying those sites meant we could go back and kind of change that. So that's those sites. In terms of do I think that there is a downside to the number of sites going live by the end of '27? Unlikely. We kind of know exactly what those sites are, and they are all in some form of build. There is an upside that if we sign something else with a client between now and the end of '25, there is a possibility that -- whatever that new site would be would also go live. So there's no expectation in there of signings of sites during the course of the year. But we will continue to bring new products. Hopefully, they will not cause delays. They hopefully will not need things like permitting and stuff like that, that can cause delays on sites, but I'm not expecting that.

Yes. So just on the refinancing, I know you mentioned that you have the optionality to pay down the [ 25 ] in cash, but would be curious to hear your updated thoughts on how you're thinking about the rest of the '26 and '27, if you have any updated thinking on timelines of when that could be communicated?

Yes. Well look, I'm not going to share any specifics. But look, we're highly alert to the topic, as you can probably see from sort of the way we would typically address our maturities well in advance of them going current. So you should expect more of the same of that. And I guess I would point to the fact that how our high-yield bond that we issued in the summer last year, how it's trading today, it's trading well. The market is clearly pretty active at the moment. And well, I've probably said enough to give you some indications of how we might be thinking about it. But no, we'll always address these well in advance of maturity.

Hi, William Woods from Bernstein. I think, Tim, last time -- last year, we were talking about kind of inflection into 2026, kind of I think you were talking about J-curve inflections. But we've obviously seen a slower build of modules and CFCs into this year. I suppose do you still see an inflection coming in? Is that more like FY '28 now when you think about it? And then linked to that, just the second question is Kroger, obviously, there have been delays there. I suppose do you see any thoring of the relationship there? Is it going well? Do you see them ordering more sites?

Great. Look, we do expect an inflection at some point. We just don't know when it is. So we're being very cautious in our guidance. We're talking about very clearly what we can absolutely see that we're building and contracted to go live with, plus where we can see sites that if there's no improvement in the way that they perform at all, will result in that level of drawdown. And so as I say, we're very cautious in what we're saying. We would like to see that inflection, and we are in the very early stages of starting to roll out the stuff like Swift Router that we spoke about that could cause a material growth in demand and some of those sites accelerating further. Does that happen in '25? Does it happen in '26? Does it happen in '27? Does it happen in '28? I'm not 100% sure, but there's a lot of improvement coming down the line that is both improvement for the shoppers as well as improvement for the client retailers. So we'll wait and see. As I say, we're being cautious. Thoring, there isn't a bad relationship with Kroger to Thor. What is definitely true is that there is an accelerated level of engagement that is at the highest level we have ever seen with them at the moment. There seems to be an extraordinary amount of focus. Obviously, there were some other distractions that they had before that have fallen away. And we and they are working very closely on these sites. We still believe the U.S. is a huge opportunity. We would like to address that huge opportunity with Kroger. We have to address that huge opportunity somehow, but we would like to do it with Kroger, and we're working very closely with them to try and do that.

It's Giles Thorne from Jefferies. The first question was on the normalization of the technology spend. And I wanted to dig a bit more into the logic there. And I suppose the 2 bookends are on the one end, reducing costs to address your cost of capital. On the other, there's a genuine tapering of the innovation curve. So yes, between those 2 bookends, where is your logic sitting?

Yes. No, thank you for the question. Look, we have done an extraordinary amount of work over the last 4 years, in particular, in a couple of areas. One has been a huge replatforming of everything that we historically developed. So a lot of it was not adding clever functionality. It was just getting it into secure, scalable, sub-2rated micro services in the cloud. And we are 99-point-something, I would say, percent through that. So that is a reduction of things that we need to do that really doesn't have a bearing on the level of innovation and improvement that will come out. The second thing I would say is that we launched Reimagine with quite a lot of very hardware-specific projects that we're also coming to the end of and hardware-specific projects are much more expensive to develop, test, prototype and deploy. And we've got them to a point where we can do most of the enhancements now in software. We don't need a fourth-generation robot, our third-generation robot will last us for quite a long time. The benefits over the second and the third are material, and we don't need another step change to achieve what we need to do. We will continue doing work on them to drive up -- to drive efficiency and reliability and cost down, but we don't need a complete rework of it, right, which is what we've done 3 times. Bear in mind that a large competitor we spoke about earlier is still selling a 30-year-old design. So we've gone through very rapidly 3 major steps in design. The last thing that I would say is we are already seeing some double-digit type improvements, low at the moment, but improvements from things like AI tooling that the engineers can use to drive productivity. So we do expect to see productivity of engineering improve, focus more on meaningful functionality, needle driving kind of change in terms of the focus of what they're actually working on and less physical, which means that we can do an awful lot with less.

Just add to that as well. For me, this is -- a lot of this is about cost discipline as well. We're getting to a point where we have a guide for what our largely discretionary R&D spend ought to be, that 20% of recurring revenues and now moving into having a period of significant investment moving into delivering cash returns from that investment.

And then second question and it comes back to the categorization of the CFCs, and thank you for that color. To shed a bit more color on how we can expect CFCs to transition between the categories. Can you talk about how you're going to be scaling up or scaling down or leaving alone your Partner Success program over the next 2, 3 years?

And I think what you could see in Steve's numbers is that almost every team in our business is shrinking other than our Partner Success team. So we are continuing to deploy resource in that area. We're continuing to embed ourselves with our clients. As I say, even within clients, there are a range of CFCs. But you can also see how the more recent launches, like the Coles comments they made overnight have learned from the historical launches and trading to be able to staff in a better position, okay? But we have no -- there's nothing that we need to do to hit the financial targets that we said that would involve cutting the size of our partner success teams.

A couple of technical questions from me as well. The first one being logistics business. As you're separating ORL and you're clearly moving in the direction of just having a technology business, what happens to logistics business going forward?

At the moment, nothing happens with it. We own it. It's a logistics business. It generates stable cash flow. It obviously does a great job serving its 2 U.K. clients. Its bigger U.K. client has a right to take over some of its operations and pay a cost -- pay a buyout kind of price to do so. That is something that it may or may not choose to do or we may or may not have conversations with it and our other partners around whether those businesses should be consolidated and a retail and logistics business and a tech business. But at the moment, we're very happy to own it. It generates stable cash flow. It helps our partners significantly. It understands the technology, and it's a good talent pool as well for people that work in Partner Success, helping our global partners.

But ultimately, as you evolve in technology business and you want to be asset-light, labor-light, just focused on R&D and delivering your solutions to your clients, logistics is a bit of a -- is a heavy business.

[ Alex ], it may make sense at some point in the future, but it's not -- it's a valuable business. So if Ocado Retail and its 2 joint venture shareholders want to buy it, we would want to be paid the right price for it, otherwise we're very happy to continue earning the type of money that you see us earning. But philosophically, are we averse to them acquiring it? Would we try and stop it? Probably not because I can see the logic that you're alluding to, which is that it would make sense to have a retail business with 10,000 to 15,000 staff doing GBP 3-and-something billion of sales. It would be more understandable as a retail business. The retail business would stand on its own 2 feet more. It would be more understandable. And then the technology business wouldn't have a large kind of 3PL logistics firm attached to it. But we are where we are today, and we'll see where we move to in the future.

The balance sheet point that you highlight is a compelling one. Our Technology Solutions business is a capital-light business. Basically, we have a few office buildings, and that's it, whereas if you look at our logistics business, you've got all those CFCs and vans and kit and so on. So there is an attraction to that as that part of it, as you rightly alluded.

And my second question is on McKesson. I had anticipated that you would recognize the revenue from McKesson contract this year in 2025. It sounds like it's now going to be spread over 3 years. I want to understand the logic behind that. And separately, how do you account for the expenses? Have the expenses been booked in '24, '25?

So the cash spend clearly doesn't change through to the accounting.

Meaning that 100% of it has been spent and therefore, it's now on the accounts.

Yes, the cash has already been spent in respect of the build funded by the partner for the CFC -- for the site for McKesson that's going live this financial year in fiscal '25. The revenues, while there will be receipts of the revenues in line with the contracts and so on. The fact that there is a software provision around the contract allows us to only record it over a 3-year horizon. We can't do it in a single year.

And going forward, do you anticipate similar setup?

No, we don't. And it's something that I think it will be important to build into future contracts to address this particular point so we don't run into that same outcome that I just described.

It's Marcus Diebel from JPMorgan. Two questions. The first one is on your last presentation, Tim you mentioned that some of the sort of like good mature CFCs are turning to 75% utilization rates. Are they getting also then profitable for your clients? So how do they think about it? Is it that it's necessary to get to 80%, 85% and then you have scale, then you turn profitable and then you buy another CFC -- kind of like what is the current thinking there or will they turn profitable really when you have proper scale and have like 4 or 5 CFCs depending obviously on the country at 80% utilization rate? So that obviously drives whether you sell more or not. That will be the first question. Maybe take that one first.

Sure. Look, Marcus, A, I can't speak about the clients' profitability. It's not for me to do so. But I think when we do modeling of scenarios for clients and potential clients, what you can see is that if you have healthy margins in your business and you're able to translate those healthy store base kind of gross margins across into the warehouse and you've got efficient inbound supply chains and you run -- if you run our supply chain and you run ATP and stuff like that to drive down waste and you make sufficient margin at the top line, you should be able to drive your warehouses profitable well under that kind of utilization. But of course, that does rely on you having good labor planning and management. It relies on you having good margins. It relies on you controlling your waste. If you don't do those things, you could get to 100% utilization and you won't make any money. So there's kind of a combination of factors that means it depends who and where and on what day, obviously, the delivery cost and the densities and all that kind of stuff come into play. It should be possible to turn a warehouse profitable at 50% utilization if you built it for the right place in the right place and you run it well. Some of our clients are still learning the muscles of what I would describe as running it well, and they are not as efficient as the U.K. operations are. If you opened a new site in the U.K., in the North, for example, and got it to 50% utilization, I would expect it to be generating positive contributions for doing that.

Perfect. And then the second question is on the efficiency of new CFCs. You mentioned half the size. How much of a game changer is that really? Because obviously, it will allow you and you mentioned it to go closer into the cities. Is it a game changer or is it just really about land and those things in the first place?

And by the way, it's half the building size, but we've also done all the design work to be able to utilize half the physical land footprint as well on the outside of the building despite doing the same volume from effectively the same asset of a van. It is very important, okay? It's very important because it's not just the rent that is obviously half or close to half. It's also largely the client fit-out beyond the developer's spec building is largely a fit-out cost based on a cubic meter. And so it's almost half. It's probably about 60% of the fit-out, i.e., a reduction of 40% of the fit-out cost, which is the largest part of the upfront that it costs a potential client or a potential client to build one of these warehouses. So that's material as well. And then you speak about building it closer in, that can have a very positive impact both on final mile delivery costs by reducing the stem times, but also then on the ability to do more kind of clever things to do with multiple routes in a shift and stuff like that. And so we're kind of at a point where you're talking about mid-100 -- 100-and-something thousand square foot to have a kind of GBP 500 million a year site. It's really quite positive. And I know the team at ORL were quite excited when we explained to them that they didn't need to look for GBP 250,000, they could go and look for GBP 150,000 or sub GBP 150,000. It's meaningful.

Can I check the cash burn of GBP 200 million, does that exclude the Autotore? So it's GBP 140 million.

That's underlying.

And then the contract liabilities, what do you assume there? Like if prepayments are GBP 50 million better, does that cash burn become less than GBP 100 million, GBP 50 million worse?

We expect to see an inflow from upfront fees of around GBP 40 million.

GBP 40 million of prepayment Okay. And could you do better than that? I mean GBP 97 million last [Technical Difficulty].

Its helpful. That's just come through last night.

It depends on what new business we signed during the course of the year.

It's Luke Holbrook from Morgan Stanley. My first question is just on the point of the refinancing, the equity raise, which I think, Tim, you mentioned on the media call this morning was not a route that you intended to go down and implied by Stephen's comments earlier. I'm just wondering why you wouldn't want to pursue more of an equity raise just in the context of the cuts that you're making to your engineering base and your intangible CapEx this year could largely be offset by the additional coupon costs that would result in going down the bond route? And then the second question that I have is more from Kroger. So I think at the last set of results, you implied that Monroe had 5.5 modules. Project Endeavour was going really well. And it seems like one of the unnamed customers on the presentation today, you more than doubled the cost of demand when you had advised the customer. Is there a point here that you could consider running the facilities effectively on behalf of your customers to try and actually drive additional demand? Because it just comes across you've got the technology, but actually the way the customers are struggling to operate it at this point.

So Luke, sorry, I'm slightly confused because I don't think I discussed anything to do with equity raises on any call. So I'm not quite sure.

Sorry, I saw a headline on Reuters saying there was a media call.

It definitely wasn't a media call with me because I definitely [Technical Difficulty]. Just in terms of the tech part, it's not -- yes, it is important to reach the cash flow breakeven, but it is also important to understand, we will be developing smart functionality probably at a faster rate than we have in the last 4 years with the reduced teams because of the amount of heavy ground work that we've done in terms of delivering Reimagined, delivering the whole tech replatforming, delivering the physical products that I said are more expensive. And so this is not we've raised debt rather than equity. We've got a coupon to deal with. We can't afford any technology. We've got a slash and burn kind of thing. We're not doing that. We're going to continue to innovate on this platform at a very fast rate. And we are seeing improvements, as I mentioned before, in the productivity of engineering. And we are not banking on the types of levels of productivity enhancements, the likes of Sam Altman are privately telling people, I was fortunate enough to sit a thing with him very recently. We're not banking on the kind of enhancements that he's expecting. We will get through a lot of technology. In terms of the -- there's 2 sites, we could run -- as we do in the U.K. on behalf of Ocado Retail and on behalf of Morrisons, we could run some warehouses for clients internationally. I have offered to do it. The clients tend to believe that this is a long-term play where they're going to have a lot of these warehouses and they want to learn. They, at the moment, would rather incur more upfront operating losses to learn that muscle themselves, so they don't find themselves locked into a long-term contract across a huge volume of business where we're earning a fee for running it because they never built that muscle up themselves. So that's currently where I found the clients generally have been. We could run them very quickly at better operating levels. In terms of the kind of customer acquisition and growth part, almost all retailers want to control that themselves. They all want [Technical Difficulty]. What is true is that they are more appreciating today in some of them the difference of how you acquire and build up a customer base and focus on the nursery journey and the frequency and the loyalty and the basket sizes and stuff online, which does work somewhat differently to their existing strong kind of muscle sense in terms of how to trade a bricks-and-mortar retail business. And so those are kind of examples of where we have gone in and helped someone to start to grow their business at attractive CPAs with good retention because it's really easy to go and acquire a lot of customers. You go and give the wrong people, big discounts and you can show great customer growth. What we've been focusing on for years and what Hannah and the team in U.K. are doing and some of the skills that we're trying to transfer to our clients is to really understand how to direct the right level of incentivization and messaging to the right types of people to get them in and kind of hold their hands through those first 3 to 5 shops and to create that loyal shopping pattern that we've been successfully doing. It is literally completely new to some people, just a completely different way of thinking about marketing and customers and lifetime value and stuff to what they're used to doing. And some people are further up that curve and knowledge point already, but there's plenty to go after.

Sreedhar Mahamkali from UBS. A couple of questions please. I think maybe just going back to the classification, Tim you've shared with us, how long have you been thinking along these lines? And I'm particularly curious about the third set of sort of sites that require a new strategic approach. What happens if some of these sites don't actually improve? And like what is the path ahead for those? Can you move them out of the scope or what needs to change? That's the first question.

So one hopes to help the clients to fill them by increasing the use cases that they can use in them. And frankly, some people might use those increased use cases in even sites that are growing faster and just fill them up even quicker. So that is kind of plan A, right? Plan B, obviously, for a partner, if ultimately they don't work is that ultimately, we have to come together and work out what to do with those sites. We own the equipment in those sites. And yes, a lot of it is valuable equipment that could be moved like robots into other sites. We don't, as you recall, sell the robots to people. So we sell capacity to people, and therefore, they could be moved. But it is a small number, and we are hopeful that we will be able to usefully deploy those sites given the capital that has been sunk into them by both parties in a useful way in their organizations.

Have some of them actually moved up the curve since?

Some of them have seen quite spectacular 30%, 40% year-on-year growth. It's just from a small number to a less small number. And it will require quite a number of years, even a 30% growth to fill them up. That's the challenge. You don't really want to look at a site that's taking the length of 5, 6, 7, 8 years to fill up. And so it's how do you accelerate it. So some as we put them -- they're stuck up on a graph, I remember seeing in the office where someone calling kind of no growth. And you go, excuse me, but didn't that one grow 38% last year? Oh yes. So they can be growing. It's just that they're not growing from a big enough base -- at the level they are, you want them to be growing 100% year-on-year. It's that type of thing.

Second one, just to Stephen, on the cash flow guidance, if you could just share a little bit more in terms of what are the other moving parts we should be thinking about? Clearly, you've talked about movement in contract liabilities. '25 because I think what we're seeing is GBP 200 million outflow, but CapEx is down by GBP 100 million. We're still seeing EBITDA grow. There must be other elements there. If you could help us if there's anything to add.

I did expect this question, so I can take you through. I'll try and keep it headline rather than get into the detail on this one. So EBITDA, when you flow through all the guidance that we've given will be about GBP 240 million or so. Upfront fees, I've mentioned the GBP 40 million. We've got visibility around that from the new deals that we're talking about, but also the design fees that are due from Aeon and Lotte. We're going to start to get some cash flows coming through from OIA. There's a small number there. I think you should be thinking sort of GBP 10 million, GBP 15 million or so. Technology CapEx, I've highlighted GBP 250 million in total across tech. So CapEx will be about GBP 170 million, GBP 180 million of that number. You then got a little bit of CapEx across retail and logistics, you're thinking about sort of GBP 20 million or so in that -- across those 2 businesses. Small working capital inflow of around GBP 20 million. Net interest costs are going to be around GBP 50 million. That's the net of the coupons on the bonds that we have less the cash that we earn from the balances that we keep on the balance sheet of around -- we typically have around GBP 500 million, GBP 600 million of cash on the balance sheet, generating around 4% or so from that. So there are net interest cash costs of GBP 50 million. And if you put all of that together, you should get to an underlying cash flow pre-growth CapEx of GBP 70 million. And then you take off the GBP 130 million of MHE CapEx that I referenced for fiscal '25, and that gets you to the GBP 200 million number. Those are the key moving parts.

[indiscernible]. Just a question on Kroger. Have Kroger added any modules in '24? And if so, was that just Monroe or our CFCs as well? And then second question also on Kroger. You point to a number of different products in Reimagine that they're ordering. Do you expect a step change in the performance of their CFCs this year, including also Swift Router, I guess or are there other issues that are holding things back there?

So we've done a lot of work with Kroger on improving their utilization of the drawn down capacity. So we've done a lot of work in the last 6 months or so on PM growth versus AM growth, for example, which doesn't require a drawdown of more capacity. So they've been kind of limiting the slots in the AM that were at or above the capacity that they're paying for and driving growth in the PM. Swift Router will dramatically accelerate the availability of doing that in terms of showing those slots to customers during the course of the AM rather than needing to drive customers from the previous evening to take PM slots over AM slots. So it's a massive enhancer to that. So I'm offhand, I can't remember the exact whether they've drawn down a little bit more in Monroe or not. They're doing more volume in there -- materially more volume in there, but that's because they've seen a material increase in PM, which has been what a lot of the work has been around and filling other sites closer to their drawn capacities. They've got a little bit further to go. So there is some drawdowns expected in Monroe that will take it up to the 7. We would then expect to take at least 2 of their sites beyond their original design criteria during the course of the year, where we will offer them more capacity, particularly in 2 of the fuller sites. And yes, Swift Router is -- should accelerate this significantly.

One more question on cash, which people are clearly asking about. So you have GBP 772 million of cash on your balance sheet today, and it looks like underlying cash outflow, GBP 200 million in 2025, plus the payment from out of store. So you're looking at cash outflow, roughly speaking, GBP 140 million, GBP 145 million. So GBP 770 million minus GBP 145 million, you get with GBP 630 million, GBP 635 million of cash at this point. Then you go into 2026. And clearly, in 2026, you're expecting to become free cash flow positive. This is what you're saying.

Turn positive in '26.

In '26. So again, using that trajectory, what is the lowest cash balance, therefore, that you will have between now and the point that you turn free cash flow positive? Because it doesn't seem to me that it would go below GBP 600 million.

No, it wouldn't. It's not -- we don't plan it to.

And if it's -- well, I don't know if you can answer the question, what is your planned lowest point of free cash flow?

Plan -- I mean one of the things that I am also alluded to is how the rating agencies look at us and what the metrics are important for them. Liquidity, our liquidity is very -- I mean where we are rated right now, liquidity keeps us at the level that we're at right now, our liquidity levels given our cash burn. So I think the more we improve that cash burn, we'll demonstrate to the rating agencies and they will lower their sort of liquidity targets or thresholds that they have for us. So I wouldn't want to take our cash balances materially below GBP 500 million.

So therefore, the [ stuff ] that you owe could theoretically come from that balance. You certainly don't need to issue any equity.

Yes. I mean I'm not going to get too much into the details of our refinancing plans. But I think you've got to be a sort of -- what the markets are doing, how our bonds are pricing and how attractive those markets are and the benefit of the certainty that goes with that versus sort of taking a punch on what they might look like in 12 months' time. So that's just another consideration for us, [ Alex ], on this one.

So again, above GBP 600 million in cash on the balance sheet at the lowest point. And theoretically, given your today's situation, which hopefully will improve, your rating agencies do not want you to go too far below GBP 500 million.

Correct.