Ocado Group plc (OTCPK:OCDGF) Q4 2021 Earnings Conference Call February 8, 2022 4:30 AM ET
Company Participants
Rick Haythornthwaite – Chairman
Timothy Steiner – Chief Executive Officer
Stephen Daintith – Chief Financial Officer
Conference Call Participants
William Woods – Bernstein
Fabienne Caron – Kepler Cheuvreux
Victoria Petrova – Credit Suisse
Andrew Gwynn – BNP Paribas Exane
Nick Coulter – Citi
Simon Bowler – Numis
Andrew Porteous – HSBC
Sreedhar Mahamkali – UBS
Rob Joyce – Goldman Sachs
Xavier Le Mené – Bank of America
Tom Davies – Berenberg
Operator
Welcome to the Ocado Full Year Results Presentation for Analysts and Investors. You will hear a presentation from management and then there will be an opportunity for questions afterwards. But first, let’s hear from Chair, Rick Haythornthwaite.
Rick Haythornthwaite
Having been on the Ocado Group board for just over a year now and chaired since May 2021, I have been able to take a good look at the business and make some important observations. The first is that Ocado is a cauldron of creativity. Innovation really seems to be part of the DNA of the business. Throughout the organization, colleagues are motivated to take on some of the biggest technology and engineering challenges of the age and then create practical solutions, which through our close and collaborative partnerships with a growing number of the world’s most forward-thinking grocers make a material difference to the lives of millions of consumers. This commitment to discovery, innovation and improvement is encapsulated in Ocado Re: Imagined, the next leap of game-changing technology and innovation, which we unveiled during a virtual product launch held on January 26 this year. Many of you would have been at that event. And I would encourage those of you who weren’t able to join us to go to the on-demand replay on ocadogroup.com following our results presentation. In any case, we will be taking the opportunity today to reprise some of the messages from that event. And Tim and Stephen will be digging deeper on what Ocado Re: Imagined means from a financial and strategic perspective.
Secondly, I have been consistently impressed by the energy, creativity, focus and vision of the people of Ocado. A healthy culture creates the conditions for success and Ocado has been able to retain many of the characteristics of a startup: pragmatism, teamwork, constant striving to find ways of doing things better, with the power and organization that a big business brings to a collective effort. This has never been in greater focus than during the COVID period, where challenging conditions called on the great reservoirs of resilience and teamwork within the company. I would like to take this opportunity to thank all my colleagues for the incredible work they are doing and the unflagging commitment they have shown to clients and colleagues alike. On behalf of the Board, we are proud of you all. The grocery industry worldwide is at an inflection point, and in many respects, so is Ocado Group. Our challenge is to focus our creativity and harness our enthusiasm to take advantage of our many strengths and to decisively set the bar as we change the way the world shops for good. I can assure you that this challenge is well in hand.
Timothy Steiner
Thanks, Rick and good morning everyone. The past year has further reinforced that demand for online grocery is here to stay. In the majority of mature markets, the fastest-growing channel is online, and to truly win here, food retailers need to deliver the best offer with the best economics across all customer missions. For innovation that is powering the development of the unique and proprietary Ocado Smart Platform is focused on providing an unequaled customer experience through groundbreaking technology, which also leads to an unrivaled low cost operation. The new generation of Ocado technology, which we have called Ocado Re: Imagined represents a transformational leap forward, allowing our partners to comprehensively outcompete peers online.
Partners ordering CFCs today will be able to go live quicker at lower cost and achieve higher margins and returns on capital. For Ocado Group, this means a bigger addressable market, the opportunity to win new partners more quickly and fresh opportunities to accelerate growth with existing partners. Over the last 20 years, Ocado Group has been a pioneer in the development of online grocery retailing. With the innovations to the Ocado Smart Platform we announced this year, we have again reset the bar, demonstrating decisively that an online grocery service powered by OSP is able to offer what the customer wants with the economics that retailer needs. We are going to begin this morning with an overview of the FY ‘21 results from Stephen Daintith, our CFO. I will return for a quick resume of what Ocado Re: Imagined means in practice for Ocado Group and its partners, and then Stephen and I will discuss together what this all means for us strategically and financially.
Stephen, over to you.
Stephen Daintith
Thanks, Tim and good morning everybody. So it’s great to be back again with you all for the Ocado full year results. During the half year results, I talked to you about my first impressions and the priorities I have defined in my role as CFO to support the continued strong growth of Ocado Group. Today, I wanted to start by revisiting these topics. My first impressions are even more strongly held. Our unique culture of creative problem-solving and self-disruption continues to unlock a growing opportunity set for Ocado Group. Ocado Re: Imagined reinforced this, a couple of weeks ago.
Now in terms of priorities, we’ve made good broad-based progress since July. I am committed to a strong communication around our progress as we deliver on an accelerating rollout of OSP globally and conversations with many of our shareholders confirm that this would start with incremental clarity on Ocado Smart Platform economics. Today, I am pleased to say we will be announcing some group solutions KPIs that will help you track our progress on the rollout of OSP, including towards the target operating cost we previously outlined.
Now with respect to effective capital allocation, our recent Re: Imagined event highlighted seven key innovations that we expect to materially enhance the value of OSP, both for our retail partners and for Ocado Group. At almost £1.5 billion, we maintain a strong liquidity position to support our bold growth ambitions. Our successful £500 million bond issuance in the second half was also an important milestone, reflecting the growing maturity of the group’s capital structure and business both as our first unsecured note and priced at a lower cost than the secured note that we can currently refinanced. We’ve strengthened our finance team, too, with some key new hires across group finance, commercial finance, tax, financial planning and analysis, and treasury functions. And we’ve gone live with cloud-based systems for both treasury and accounting that will better empower us to deal with the complexities of scale and globalization. Similar work is also underway now for the supply chain. So it’s been a strong start, and I’m excited to build further on this progress in the coming year. At a critical turning point for the industry, we are well positioned to take transformation of the landscape for grocery retailing worldwide to the next level.
Now let’s have a closer look at our financial performance in fiscal ‘21. The headline numbers primarily reflect three things: number one, strong underlying performance in Ocado Retail with second half performance impacted by the temporary disruption associated with the fire at Erith and the cost of the mitigation measures taken to address industry-wide challenges in the UK labor market; secondly, our successful rollout of significant capacity for OSP partners in the UK and internationally; and thirdly, investments in the development of the Ocado Smart Platform. Reflecting relative scale, retail performance drives group revenue growth of 7%. Growth in fee revenue from partners was strong in both our UK and International Solutions segments as we successfully brought live another further 5 customer fulfillment centers for our partners. A £12 million decline in EBITDA compared with last year reflects good revenue and gross margin performance of Ocado Retail and productivity improvements in UK Solutions and Logistics offset by the impacts of operational disruption at Erith and significant investment in platform development.
The increasing rollout of Ocado Smart Platform naturally brings high depreciation and amortization costs. These were only partially offset by net exceptional income and reduced net finance costs, resulting in an increased loss before tax. As previously highlighted, we maintain a healthy liquidity position of almost £1.5 billion to support our significant growth ambitions. In 2021, Ocado Retail continued to build on the exceptional performance in 2020. Underlying performance throughout the year was strong, but the second half was also impacted by temporary challenges associated with Erith disruption and UK wide labor market challenges. As a result of these two competing forces, strong underlying trends and temporary disruption, revenue was up 5% in 2021, 6% if we adjust for the disposal of Fetch. Similarly, EBITDA was broadly flat at £150 million, reflecting the same temporary impacts in addition to strategic investments in marketing to support long-term growth. We’ll look at each of these more closely shortly.
Importantly, performance in 2021, even including the impact of the temporary challenges continues to demonstrate an operating step change compared with prepandemic. 2021 revenue was up 41% versus 2019, whilst EBITDA has almost quadrupled versus 2019. Performance on key KPIs emphasized the strong demand for online grocery that underpins this continuing step change. The business grew active customers by 22% in the year to 832,000 and now has its eyes firmly set on the 1 million mark. This strong new customer growth drove orders per week growth of 12% in the year, allowing capacity previously deployed to larger baskets to be reallocated to new customers as shopping behaviors began to trend back towards pre-COVID levels. Ocado Retail also successfully maintained a strong gross margin performance experienced in the first half of the year, including investments in retail prices.
Now, let’s take a look at all of this in more detail. Here, you will see that the retail business achieved a 6.6% EBITDA margin in 2021, slightly below the 6.8% achieved in 2020. It’s worth exploring here some of the differences in margin profile between these 2 periods. As previously mentioned, gross margin performance was consistently strong in the year, a result of a combination of higher order volumes, improved product mix and the commercial sourcing benefits that strengthened buying team has been able to achieve following the Waitrose transition. Importantly, this meant the business was able to invest in retail prices and still achieve the strong margin performance.
Turning to distribution costs, trunking and delivery costs increased slightly as savings from non-repeating COVID costs were more than offset by the impact of lower drops per van per week and investments in labor incentives in the second half. The increase in CFC costs reflects the typical cost inefficiencies you would expect with the go-live of 3 new CFCs in the year as well as impacts associated with the challenges of the second half, disruption following the Erith fire and investments made in temporary incentives in a tight labor market. Importantly though, underlying efficiency continues to improve units per hour at our mature sites rose to 170, and that’s inclusive of the period of disruption at Erith. Excluding this impact, it will be 172. And all 3 CFCs that were launched in the year, that’s Bristol, Andover and Purfleet have already achieved this level of performance, and that’s less than a year from go live.
Now looking at marketing costs, we can see that the retail business is investing for long-term growth to drive increased brand awareness and strong new customer acquisition for years to come. Of course, 2020 also provides a particularly low comparable for marketing spend given lower levels of customer acquisition during the pandemic. Increased capacity fees mirror the significant growth in available capacity during the year. As previously guided, robotic capacity also brings a relatively higher fee rate than our heritage sites, reflecting the improved levels of efficiency that this capacity brings. The rise in admin costs reflects investments in key teams such as the buying team to support future business growth. This was partially offset by a reduction in the accounting charge associated with the senior management incentive scheme compared with last year. During the third and fourth quarter, retail trading statements, we talked about the temporary but acute challenges we’ve been experiencing in the labor market, their associated impact and our plans to resolve them. It’s worth a short update on this today.
The main takeaway is that these are temporary issues, which we are close to resolving. They haven’t distracted either the Ocado Retail business or indeed Ocado Group from delivering on its long-term growth potential. You can see how the business has invested to accelerate growth in the left-hand chart. In 2021, Ocado Retail went live with 3 CFC sites, materially increasing available capacity for growth by around 170,000 orders per week in response to strong demand for online grocery. Actually the impacts of the Erith fire compounded by labor market challenges in the second half constrained the business’ ability to grow as rapidly into that capacity as originally planned in the near term. This impact is already reflected in our EBITDA margin guidance for fiscal ‘22. Since the fourth quarter, we have continued to make good progress addressing labor shortages across delivery and CFC roles. We have now more than half our vacancy rate since the October 2021 peak. And on this trajectory, we expect to put this constraint behind us during the second quarter and for growth to build strongly throughout the year from there.
We continue to expect Ocado Retail to deliver strong mid-teens growth in fiscal ‘22. This strong growth combined with the inherent operating leverage in the model underpins our ambition for EBITDA margins to rebuild towards fiscal ‘21 levels following a year of significant investments in fiscal year ‘22. Of course, Ocado Retail is investing to support strong long-term growth and the underlying dynamics we see in the market continue to support this decision. By the middle of this decade, online grocery is expected to account for almost 20% of the UK grocery market, and that’s up from 12% today. With a leading customer proposition, Ocado Retail is well placed to seize this huge growth opportunity as evidenced by the strong new customer acquisition achieved this year. Seizing this structural opportunity means investing in both capacity and capability to successfully support a much larger business in the years ahead.
With this in mind, Ocado Retail continues to invest in three key areas: firstly, capacity. CFCs announced to date will bring ultimate orders per week capacity potential to around 700,000 in 2023. Today, we’re announcing two further CFCs with one for the Northwest and one for the Southeast planned for 2024, in total adding around a further 200,000 additional orders per week. Altogether, this will mean total potential CFC capacity for Ocado Retail of around 900,000 orders per week, more than double the amount of capacity that was available to the business at fiscal ‘20, at around a 50% increase versus financial year ‘21. Second, transformation, investing in talent and systems that will empower the business to make the business the most of the opportunities to drive both growth and value into the future. Third, marketing, as we previously discussed, a new multi-channel brand-driven approach will underpin improving awareness and continued strong customer acquisition and has already shown great early results. As you can see, the business is ready to drive strong growth. Our OSP technology means that growth will also bring significant improvements in efficiency as more robotic capacity comes online. And that is before the impact of Ocado Re: Imagined, which we expect to further transform operating economics for our partners.
Now, turning to UK Solutions and Logistics, you will see similar drivers in the performance of this segment to those that we’ve just discussed in the retail overview. Strong fee growth of 28% reflects the increased capacity rollout for Ocado Retail and Morrisons return to Erith. In total, the 3 new CFCs that we launched with Ocado Retail this year represented around a 20% increase in live sales capacity, meaning capacity that from a technology readiness perspective is available to grow into. When fully ramped, our existing UK CFC base will now have a total potential capacity of over 750,000 orders per week. And this is before the additional 300,000 orders per week capacity from the sites expected to launch in the next 2 years, including the two new sites announced today that are each expected to launch in 2024.
Cost recharges grew slightly ahead of total volume throughput, reflecting the inefficiencies associated with immature capacity as well as the impacts of investments made in labor incentives in the second half. This was partially offset by improving efficiencies in mature sites and on the road, which we share with UK partners through reduced cost recoveries. Aside from cost recharges, distribution costs also include engineering costs for which our partners pay a fee. These costs naturally run higher in immature sites as a minimum level of spend is required for a live site regardless of volume, but underlying progress continues to be very encouraging.
Erith achieved a 36% reduction in engineering costs on a cost per each basis. That’s despite the impact of the fire. And Bristol, our mini CFC that opened in March is already operating at a cost per each level that’s similar to Erith. That’s less than a year after opening. Admin costs increased ahead of revenue growth as a result of investments made in additional head count and technology resources in the year. Overall, strong fee growth, partially offset by these increased costs underpin the £24 million increase in EBITDA compared to 2020.
Now, turning to International Solutions, we are starting to see revenue build strongly as we bring live more CFCs. We recognized £49 million of fee revenue from OSP partners in 2021, quadruple what we achieved in 2020, and that’s reflecting a full year of fees from sites in France and Canada and the go-live of two more CFCs for our partner, Kroger in the U.S. The remainder of revenue is split between Kindred Systems and equipment sales to partners, the latter which has no impact on EBITDA. Revenue from Kindred has taken a little more time to come through due to delays in contract signings, but we remain very excited about the long-term opportunity for their robotic picking solution, particularly in the general merchandise and logistics sectors. Of course, we’ve already gained significant value from the Kindred team and accelerating our progress towards delivery of a robotic picking solution for our grocery partners. The first installations of this solution will be delivered this year, and we expect to be able to take over 50% of the range by volume by the end of 2023.
Turning back to the OSP rollout, as mentioned earlier, we allocate a minimum level of engineering support to each new CFC at go-live, and in the early stages of ramp to ensure our partners the best place to grow into their new capacity. EBITDA decreased by £36 million, reflecting the support relative to the early ramp stage of our operational CFC sites as well as the higher allocation of investments made in technology talent to develop the OSP platform. Though fee growth of 15% came in below our initial ambitions as travel restrictions remained constraint in most markets for majority of the year, performance still highlights our strong pipeline of CFC and in-store fulfillment commitments, including those from our 10th partner, Alcampo, who we welcomed to the club in July. Indeed, we’ve already made a strong start to 2022 bringing 3 of the 8 international CFCs planned for the year live since the year-end.
As guided, 2021 brought increased capital investment in the accelerating rollout of OSP, both internationally and in the UK. Around 60% of our CapEx investment in the year was on CFCs deployed to ramp those sites already live or across the 21 CFCs we had in build going into fiscal ‘22. Development of the OSP platform continues to be our second greatest area of investment. Around half of this spend is allocated to delivering transformational innovation such as the 7 key innovations we explored in our January product launch Ocado Re: Imagined. We expect these innovations to reset the bar in online grocery fulfillment, once again with respect to both cost efficiency and flexibility. This will enable our partners to grow faster and Ocado Group to reach a larger share of the sizable grocery market opportunity and faster. Digging a little deeper into our investment in CFCs, you can see that over 60% of CFC CapEx was related to the ongoing rollout of OSP internationally in 2021, up from around half in 2020. This reflects the increasing scale of the international rollout.
At the beginning of 2022, we have 21 CFCs in varying stage of build. This year, we will deliver a total of 9 CFC sites, almost doubling our live CFC count once again. Happily, we have a healthy liquidity position with almost £1.5 billion in cash and cash equivalents, and have to meet existing commitments and deliver future growth in the near term. Operating cash flow was marginally negative in fiscal ‘21, which primarily reflects a large working capital outflow as the group continues to scale. This was partially offset by strong retail trading performance and growth in invoice fees from International Solutions partners. That’s shown in contract liabilities.
Looking at cash flows related to financing and investment, inflows primarily related to a drawdown of treasury deposits not previously included within the cash definition and the net proceeds of the bond refinancing and issuance we completed in the second half. These funds were used to fund the investments made to support future growth, capital investments and the acquisitions of Kindred and Haddington. We ended the period with gross debt of around £1.8 billion and net debt of around £350 million.
In 2022, the rollout of OSP globally is set to scale dramatically. We are ready to bring a leading online service to hundreds of thousands more customers for our partners. The online channel is becoming an increasingly important part of grocery sales in most markets around the world. This trend is only expected to increase. Those that invest now for the long term will take the lion’s share of the huge opportunity that this represents. These 2 things inform our outlook for fiscal ‘22, the increasing rollout of capacity for partners and investments to both drive a bigger opportunity and seize it sooner.
We already provided guidance for Ocado Retail at the fourth quarter trading statement, and this is reaffirmed. With respect to our Solutions segments, expectations for strong fee growth in UK and Solutions and Logistics and especially International Solutions mirror the significant capacity we expect to deliver for our partners. We expect EBITDA to increase strongly in UK Solutions and Logistics, reflecting the increased fees we will receive from bringing on more capacity for our clients as well as a reduction in engineering costs relative to this live capacity.
Guidance for EBITDA stable on 2021 International Solutions reflects the investments we are making in our technology teams to deliver a new age of OSP technology for partners as well as the minimum level of engineering support we put down for partners to best support them in the early stages of CFC ramp-up. This increased technology investment can also be seen in our guidance on central costs. There, we are making important investments, group operation costs are expected to grow below group revenue. And in future years, we expect central P&L costs to grow significantly below group revenue growth, reflecting inherent operating leverage as the business scales. Finally, we expect CapEx of around £800 million this year as we bring more sites live and continue to ramp others. And surprisingly, the largest share of this will be dedicated to the rollout of international sites, although with a meaningful proportion remaining in the UK with the 2 new CFCs that we have announced today.
Timothy Steiner
Let’s talk now a little bit about the Ocado Re: Imagined event we held a couple of weeks ago. Ocado Re: Imagined is really groundbreaking new technology that is an absolute game changer to Ocado OSP and the online industry. We announced 7 innovations, where the sum of the parts doesn’t reflect the total impact of these innovations. Let’s go through them one at a time. Firstly, the new 600 Series bot. Amazingly coming so fast on the heels of our amazing 500 Series bot, of which we’re already so proud and is achieving so many of the metrics that we needed to, to hit our long-term plans. But we went out there to see if we could do something absolutely radical. And that’s what we’ve achieved with a 600 bot.
Using ultramodern topology optimization software, we’ve been able to design a machine, built using additive manufacturing with over 300 3D printed parts, and we’ve removed 80% of the weight of the bot for this new 600 Series bot. It’s materially cheaper to manufacture and to maintain than its predecessor, but the weight has enormous impact as we go through the grids, the floors, the buildings, et cetera. So a massive improvement and absolutely the cutting edge of innovation and manufacturing. The 600 grid, taking advantage of the low weight of the 600 bot, really means taking advantage of the lower forces that it generates. Because of that, we’ve been able to introduce a new lightweight 600 grid that is also much faster to put up and is capable of being erected in parallel rather than in sequence. So we’re able to build it quicker with a lot less material and at meaningfully less expense. We are also optimizing the design of our grids and we’re now able to install them in many existing buildings, which clearly can have a material impact on the time between a client deciding to build in a new city and getting that facility live and ramping it up.
Automatic frame load replacing one of the most physically demanding jobs in our warehouses, literally doing the work for the human of lifting the customer bins and placing them in the frames. A process that we optimized many years ago with advanced algorithms to ensure we put the right totes in the right order in the frames, to minimize the strain on the humans that are unloading them to make sure the bands are equally balanced, to make sure we’re always taking the front tote out before the last one on our routes. But now not having to use a human to do it, doing it using a machine, and literally, as I say, taking out the toughest physical job in the warehouse.
The next and phenomenal innovation is On-Grid Robotic Pick. You would have seen our robotic pick arms before working in our existing warehouses. We’ve had some live in Erith for a while. We’ve moved them on top of the grid. It’s hugely significant in terms of the throughput it means that we can achieve, in terms of getting more throughput from the same warehouses, in terms of the cost to install the robots and therefore, the savings that can be split between ourselves and our clients and the virtuous cycle that, that creates. The On-Grid Robotic Pick arms are only able to do this now because of the lighter weight arms that we’re able to use and these very, very complex coordination and collaboration between those arms and the robots that are moving around on the grid, where the robots are dropping bins in front of the arms that the arms are subsequently dipping into and picking from. And remember that robotic pick is complicated actually not so much because of the pick, but because of the pack because it’s absolutely important that the robot packs with at least the same density that a human could, so we don’t generate more totes to deliver to customers. So, lots of advanced technologies, advanced vision and a lot of coordination of the robots, going live at the end of this year in our first sites, initially covering around 50% of the range, growing to 60%, 70%, 80% in the next few years.
Really excited about Ocado Orbit, the world’s first virtual distribution center, allowing our clients to build smaller warehouses, but still maintaining the inbound efficiency and the direct delivery capability of big warehouses, which are worth hundreds of basis points in their cost structures, but doing that in smaller warehouses without the need for a large physical regional distribution center that supplies them by basically taking that regional distribution center and building it in software and each of the small warehouses act both as a distribution center, as a picking center and as a delivery depot. And the second part to Ocado Orbit is the ability for any one of those centers to serve as a primary distribution center for new micro facilities, allowing our clients to compete in Q-commerce with ranges that are completely unmatched in that sector. So instead of 1,000 to 2,000 products that the dozens of entrants into that sector have in their little micro facilities, enabling 10,000, 12,000, 15,000 products and at a very small incrementally higher handling costs than in the big facilities, meaning you can get a big supermarket offering out at convenience store pricing, but in a Q-commerce type format that will just blow away that industry, so super exciting.
And Ocado Swift Router, enabling very short lead time deliveries from large warehouses on optimized large routes. So now instead of having to choose between a crowd-sourced store picker, who can deliver point-to-point back to a customer’s home for say, a 2 to 3-hour delivery from a store’s range but at 20% or 30% premium in pricing, being able to operate at full automation, scale, efficiencies, but being able to deliver to a customer’s home in 1, 2 and 3 hours for maybe 30% plus of the output of that facility and being able to do same day for up to say, 60% or 65%, taking advantage of the unique nature of our high automation that allows us to store completed customer orders at the end of the process without having to ship them immediately in order to keep the warehouse functioning.
So we can pick overnight. We can then open the website during the day and take orders that we process immediately, combine them with some overnight orders and put them in a van route, where the van spends the first, say, 2 hours of a 4-hour route, delivering orders that were only placed in the previous hour, bringing goods to customers 50,000 range of supermarket prices in 1 to 3 hours and using the return leg of the journey to deliver the today for tomorrow orders that had been placed, phenomenally clever and only available with our high technology.
Ocado Flex. Ocado Flex is what is known in the industry as headless e-commerce. One of the issues that some of our potential partners have had was their own desire to control more of their front end or to integrate some of their own apps or capabilities into our front end, maybe for their store-based experience. Ocado Flex contains two sets of APIs, our core APIs and our smart APIs and allows our clients to build their own front ends at very low cost, integrating in any other APIs, any other bits of functionality or data they want from their own systems or from third parties, maintaining that level of control, but leveraging our core and ultimately, whichever they want of our smart APIs, to make sure they can still achieve the best-in-class experience for their e-commerce customer, whilst for example, integrating a scan and go functionality in store or making sure that something that is on offering the scan and go app in store can be moved instead into a basket online, so giving more flexibility to our partners and something that they’ve really been asking for over the last few years.
Now I am joined by Stephen Daintith, our CFO, and we’re going to talk about what we see as the scale of our opportunity.
Stephen Daintith
That’s right. So Ocado Re: Imagined, Tim. 2 weeks ago now, a very exciting event. You talked a lot at the event about the virtuous circle of Ocado Re: Imagined the way Ocado works. Can you just bring that to level a little bit more? What do you mean by that virtuous circle?
Timothy Steiner
I think the first thing is we’ve often taken what we made in our retail business historically, but basically, we’re very big innovators with big investors, and we have, over the last few years increased the investment in our platform to not just keep ahead of the competition, but to just massively drive it forward. And that’s what Ocado Re: Imagined was about. And what does it mean from an Ocado Group business perspective is really if we can enhance the platform in the way that we have, so we can make it lower cost for us to produce, passing on some of those benefits to our clients, so lower cost for our clients to deploy then that’s obviously a good thing. If at the same time, we can make it not only cheaper, but actually much cleverer, so that what it can do for our clients and what they can then offer their customers is better, i.e., they can offer their customers 60% plus same-day deliveries out of the bigger warehouses and maybe one-third or something going out in the first 1 to 3 hours, matching kind of the best short lead time services coming out of supermarkets, but with all the economics that you could achieve from the big warehouses and from planned routes, right? So if we can do that, we create an opportunity for faster partner growth. They will grow faster because they can serve more customers from the same buildings for more missions. And they will grow faster because the capital they need to deploy to do it, talk simpler buildings and things like that is less. And also to the extent that the simpler buildings means existing buildings and the time frames from starting to deploy or make those decisions to have a building to getting it live to ramping it up also shorten. Things like On-Grid Robotic Pick meaning, you need to hire and train less people in the building when you open it, to go from start to full capacity. So overall, there really is this kind of this virtuous circle – virtuous cycle, where you – our increased investment allows our customers to invest less to actually be able to do more and they will grow faster and therefore, they will deploy more. And therefore, ultimately, there’s more for us to justify the investment and carry on with an increased investment. And so with 10 partners already on the platform and hopefully more to come, we really are benefiting from the scale, creating the investment opportunity, creating a product that’s unmatched in the market. And we just think we’re creating that virtuous position that nobody else can match.
Stephen Daintith
And actually describe that, Tim it’s a package that must be extremely attractive to our partners, our clients, any responses from our clients so far in the last couple of weeks?
Timothy Steiner
Sure. I mean, everybody got to see our largest client because we give them a sneak preview. So anyone who didn’t get to see it in Ocado Re: Imagined, to go and watch it. But obviously, we saw Rodney, the CEO and Chairman at Kroger, say groundbreaking, game changer. These kind of emotions were the type of responses that he was giving about what this meant and moving on to the next phase of our partnership and accelerated growth and stuff like that.
But I’ve had reactions from our other clients that were also reinspired, really exciting, need to rethink about what this means and how we invest and how we grow together. And during the pandemic, new services came up, new offers to customers. And I think some of the things that is clear – much clearer after we’ve announced Re: Imagined that our clients can do to be the best in class in every mission, really positive reactions. Lot of people saying, no, we want more information. We want to understand how much cheaper, how much gets passed on to us. How do we – what does this mean? How quickly can we get that software live? How quickly can we be using a big shed to do 2-hour deliveries? How many 2-hour deliveries can we do from a big shed in a 1-hour radius? So really busy time for our teams because everybody – it’s really opened people’s imagination as to what’s possible. And it answered so many long-term questions. Even talking to prospects over the last 6 months or a year, people have said, where is this all moving to in short lead time in Q-commerce and range sizes and today for tomorrow, and I think this just really helps people to understand how do I build the lowest operating cost model, but deliver the best service in not just accuracy and pricing, but in accuracy and pricing and lead time. And so big excitement.
Stephen Daintith
So Tim, let’s talk about our operating cost model. And the good news is that we’re well on track with all the goals that we set ourselves. What does Ocado Re: Imagined? We’ve heard a lot about the improvements that Ocado Re: Imagined innovations are going to bring. What does this mean for us? What does it mean for our clients more importantly?
Timothy Steiner
Look, it means there are significant benefits. So capital costs coming down, robots capital costs coming down in grids as well as new things as well. We’re going to share those benefits with our clients. In fact, we’re going to be super generous in that sharing. But the way I think about it is the way that my broadband provider has treated me at home for the last 10 years or so. Every couple of years, they charge me a little bit more. That’s like the inflationary increase in price, but they keep upping the ante on what I’m getting for it.
So we’re going to take some of those savings from those cheaper robots and those cheaper grids that would be on our side and we could just take to profit and actually reinvest those in giving our clients things like robotic pick that drive significant cost enhancements for them, that we know they will pass on to their end customers. And therefore, we will grow the demand, grow their need to build sheds and amount of sheds that they have and the speed with which they scale them up, and we create a virtuous cycle for their customers, for our clients and for ourselves, and that’s how we’re looking at it. And we’re talking about site productivity up over 50% for our clients which means that they get more than a 30% saving in the labor cost they put into those sites. These are quite big numbers, and this is only the beginning, of course.
Stephen Daintith
Great. Thanks, Tim. So we’ve heard a lot from our shareholders, from the market generally. They’d like to hear more about our key performance indicators for our solutions business and to be able to judge our performance. We’ve shared two or three of these today for the first time. Do you want to circle a bit more about how we’re measuring our own performance?
Timothy Steiner
Yes, look, we want to make sure that you can understand what we’re doing. And so we put some indicators out there to explain what’s – what we’ve managed to get live, the kind of the scale of what’s going on in the business and also the direction of travel in the really important cost lines as well. So people understand how we’re – what the progress is that we’re making.
Stephen Daintith
Brilliant.
Timothy Steiner
Thanks, Stephen. And we could talk about this all day, but we’ve only got a limited amount of time.
Stephen Daintith
Thanks, Tim. Really enjoyed it. Cheers.
Timothy Steiner
That brings us to the end of this presentation. Here are our main takeaways. The grocery market is at an inflection point. A huge market opportunity exists online for grocery retailers who can deliver the best customer proposition with the best economics across every customer mission. The game-changing innovation driving the development of the Ocado Smart Platform allows our partners to fully take advantage of this opportunity. Partners ordering CFCs today will be able to go-live quicker at lower cost and achieve higher margins and higher returns on capital. For Ocado Group, this means a bigger addressable market, the opportunity to win new partners more quickly and fresh opportunities for growth with existing partners. We have consistently set the bar in online grocery retailing over the last 20 years. Our deep culture of innovation is enabling us, once again, to reset the bar decisively for the benefit of our partners, their customers, our shareholders and the communities we serve. Exciting times are ahead.
Question-and-Answer Session
Operator
[Operator Instructions] We will now take our first question from William Woods from Bernstein. Please go ahead.
William Woods
Good morning, Tim. Good morning, Stephen. So, a couple of questions from me. The first one is just on the guidance on flat International Solutions EBITDA despite doubling the revenue. Could you provide some more detail on the drivers of this additional cost? And I suppose what I’m trying to understand is, is it the structural changes to the cash profile of the CFC or have you got a mix effect from ramping or headwinds from inflation? The second question is on the Series 600 bots. Are you able to give us a view of kind of what percentage of bots will be live by the end of the year? Over how long will you expect to kind of replace them over time? And then a final question, just on the nonfood opportunity. Have you seen any further uptick from potential partners on nonfood?
Stephen Daintith
Okay. Well, I’ll do the first one, Tim, on the International Solutions. So two key drivers here. Number one, the investment we’re making in our technology platforms for the Ocado Smart Platform, we highlighted today, another £30 million also there. That’s one key driver. And then secondly, as I think the engineering costs, now every time we open a new CFC, there is a minimum level of engineering support for that particular CFC. And so there is a natural inefficiency in the earlier CFCs internationally and indeed globally. So those are the key drivers really. Those two key drivers get us to the guidance that we’re giving today.
Timothy Steiner
William, taking up to the 600s, we will start rolling off the production line about 12 months from now. And about 24 months from now, there will be exclusive bots that we roll off. So we will have probably about a year in the middle where we will be manufacturing both 500s and 600s. I think I’ve explained already that you can use 600s on any of the grids that we built to date. So any sites that have gone live can end up with a combination of 500s and 600s. As we go forward and build grids that are lighter weight and taking advantage of the lower forces that the 600s create, we will not be able to run them with a full fleet of 500s. We may be able to run some 500s on them, but not the density of 500s that would then create too much forces and ultimately end up with metal fatigue on those grids. And then your – the third part of your question was nonfood. There are significant opportunities in nonfood going forward, that the 600 and 600 grids open more than were historically available with the 500s and their grids. Because the 500s and their grids because of the extremely high performance levels that they are able to achieve, which we match with the 600s, meant that the grids were thicker and more expensive than some others out there in the market that are not able to achieve those throughputs. But with the extraordinary reductions in weight in the 600 series and therefore, the lighter grids that they need, you can gain the benefit of the low cost and the extreme throughput capabilities. But it means even on low throughput sites, they would still be the cheapest machine out there of its type to manufacture.
William Woods
Thank you.
Operator
We will now take our next question from Fabienne Caron from Kepler Cheuvreux. Please go ahead.
Fabienne Caron
Yes. Good morning, everyone. Three questions from my side, please. Regarding the metrics for Ocado Re: Imagined, in the past, you used to help us saying, if we take Andover-like CFC, cost for you would be £50 million? We have a 2-year preparation, 1-year then go live and 5% capacity. I’m just wondering how these metrics really change as we move to the new grids and the new bot. It will be the first question. The second question will be, if you move into nonfood, will you work as well with exclusivity and is exclusivity key as well for your new countries? I’m thinking here because you can go smaller, maybe – there may be more food retailer in the country that could be interested now to use your solution. And the last question would be financially, where do you expect your net debt to be at the end of ‘22, please?
Timothy Steiner
So I’m going to take the first couple of questions. The model that we work with our clients is going to remain very similar to you outlined. Obviously, Re: Imagined means that the bots will be cheaper for us to manufacture, the grids will be cheaper and quicker for us to install. The go-live time from the client will be shorter than it was before because of the robotics. We will then reinvest some of that savings into robotic pick and into automated frame loading. So overall, we’d expect the client fees to be there or thereabouts. We’d expect the margins to be slightly higher than they were before. We would expect the clients today to go live faster. And also because of the less – the lower labor requirements to operate the facility, we’d expect the clients today to ramp them faster. And because of the increased opportunity to cover more missions in terms of shorter lead time deliveries as well as the existing kind of services, we expect the client demand to grow faster as well. So hopefully, that gives you a flavor on what Re: Imagined means for the modeling. In terms of the non-food and exclusivity, I don’t envisage just giving exclusivity to anybody in nonfood. And none of our existing exclusivity arrangements cover the nonfood area.
Stephen Daintith
And then the final question around where do we expect net debt to be. Well, I think the key driver here is the £800 million of CapEx that we’ve guided to for 2022. We’re starting the year with £350 million or so of net debt. So I think around £1.1 billion is probably a good guide for net debt at the end of the year.
Fabienne Caron
Okay. Just to come back on Tim’s comments. It’s fair to assume that the CapEx from your side should be lower than with an Andover-like previous CFC?
Timothy Steiner
Yes, the CapEx going forward will be lower. But as William said before, that’s not going to affect 2022, but that will start to hit in 2023 and hit more materially in 2024.
Fabienne Caron
Thank you. Very clear.
Timothy Steiner
CapEx for the capacity, I mean obviously, what – we’d like to build more capacity. But yes, CapEx per billion of capacity will come down, yes.
Fabienne Caron
Okay, thank you.
Timothy Steiner
Thank you.
Operator
We will now take our next question from Victoria Petrova from Credit Suisse. Please go ahead.
Victoria Petrova
Good morning, thank you very much. My first question is on total addressable market. What are the new customer characteristics which could or would not have signed before and might sign now after Re: Imagined OSP has been introduced? So what is the game changer in the customer profile expanding your TAM? My second question is sort of a follow-up from Fabienne. Should we assume that new partnerships might not have exclusivity in food, obviously, in countries outside of your existing exclusive partnership countries? And my third question, could you remind if you have a target capital structure in mind? Thank you very much.
Timothy Steiner
So Victoria, on the TAM, there is two separate things that Re: Imagined is doing for us. One is that because the robots and the grids in particular are materially cheaper than the existing ones, they open up markets where the cost of labor is cheaper than it is in the existing markets that we operate in and that we previously targeted because, obviously, as well as delivering a better service and range capabilities that you can’t really do manually, our system obviously has already taken out significant labor in exchange for capital. And obviously, as you lower the price of the capital that you have to deploy to do that, it opens up new markets that you can do that in. So it just means that we will be targeting markets that have – that are lower income than the ones that we’ve just historically targeted. The second part is that within any market that we’re operating in, Ocado Re: Imagined is opening up more missions. So if customers now want to see kind of 2 to 3-hour type services, things equivalent to what the likes of Instacart would tackle in the U.S. Those kind of markets, those kind of missions or opportunities or market share is now available to our customers out of our fully automated warehouses on optimized delivery routes. And with the Ocado Re: Imagined infrastructure installed in small micro sites, obviously, the whole Q-commerce market is available, but with significantly larger basket sizes and market share available to our clients because of the ability of Ocado Re: Imagined to carry 5 to 10x the range that you’d expect to get in an existing Q-commerce operator. So the combination of countries that were previously considered to have to lower labor costs to warrant automation; and secondly, more market share in existing and those markets. On the exclusivity question, I think that exclusivity is still something that we look out on a deal-by-deal basis, on a market-by-market basis and depends on the scale of the client, their desire for it and how aggressively they want to deploy capital and their confidence and our confidence in them achieving significant market share. We have no obligation to either do it or not do it in a market. But I see nobody in the general merchandise area, if you were to move into that that would warrant having exclusivity.
Stephen Daintith
And on the…
Victoria Petrova
I’m asking about food.
Timothy Steiner
So I mean as I say on food, it really depends on who the partner is. So if somebody wants to come in and is clearly aiming for a number one market share position and very clearly wants to have the exclusive availability of our proposition because they see such an enormous advantage, then clearly we have that discussion, and we will see market by market as we roll out to new clients, where we end up on that one.
Stephen Daintith
And then the final question on do we have a target capital structure in mind, well, first of all, I’d make a point that the most important thing for us is to have healthy liquidity for our investment plans. We think now is exactly the right time to invest, to scale and catch the growth opportunity set that we see ahead of us. I think Ocado Re: Imagined a couple of weeks ago only reinforced that. So having liquidity is incredibly important to us. We have £1.5 billion of liquidity today, which is clearly [Technical Difficulty] 2022 given our CapEx plans. We would rather be a pure debt-funded company. And I think a key thing to get us there would be generating positive cash flow, which we are increasingly moving towards as we add, in particular, to our portfolio of automated warehouses. We finished this year with 10. We’re going to finish 2022 with around 19 or so and then we will be adding to that in ‘23 as well. And then we will be reaching that point over the next few years when we start to turn from burning cash to generating positive cash flow. So that’s how we think about capital structure. I think the key point is having healthy liquidity, which is what we have today.
Victoria Petrova
Thank you very much.
Operator
We will now take our next question from Andrew Gwynn from BNP Paribas Exane. Please go ahead.
Andrew Gwynn
Hi, very good morning. Two, if I can. So first off, just on the CFC build-out. I think we used to think about a figure of around about £50 million for standard-sized CFCs. I’m mindful that you’re not building too many standard-sized CFCs. But where do you anticipate that getting to? Obviously, the 600 robots and so forth, more capital efficient, but obviously we’re seeing quite significant inflation in many of the products that make up the robot in the grid. Second question and kind of bigger picture question. But if you stand back and think about last couple of years, I think it’s fair to say the market has been a little bit underwhelmed by the number of partners that have signed to the platform. What do you think is really holding them back? And can we expect a step change anytime soon? Thank you.
Timothy Steiner
Hi, Andrew, so the first one I’d say is, look, the savings that we’re making on grids and robots as we deploy 600s and the new grids to significantly outweigh any inflationary pressures that we’re seeing on the cost of those sites. And where you do see inflation, if that passes on into food, then obviously, as a percent of – CapEx as a percentage of the sales capacity and CapEx as a percentage of fees would actually be staying flat. So it’s – ignoring inflation, these are two very significant reductions in costs, some of which we will redeploy into increased automation in terms of On-Grid Robotic Pick and Automated Frameload. But we would expect to see our CapEx as a percentage of sales capacity come down as we’re moving forward, even though we’re putting in the incremental infrastructure. On your…
Andrew Gwynn
So are you able to give that – sorry, Tim, are you able to give that percentage today because I think I’ve certainly lost track of what it costs to build the CFC? I think there have been so many moving parts.
Timothy Steiner
Well, I think we don’t put out a forecast exactly. I think a lot of people have got a fairly good idea where they think those come in. I think you mentioned the number before. And I think it’s not wildly off. And as you say, it depends slightly on the size of the facility. Obviously, smaller facilities, pro rata cost less, very slight increase in the percentage cost as they get smaller. There is some efficiency and scale, but a very small amount. Moving on to your next question about – Andrew, we’re sitting here, it’s only, what is it, 4 years now since people thought we weren’t going to sign one, where it was us and Morrisons. We’re now sitting with 10 customers or clients on the platform. We expect that number to grow. I can’t say exactly when or by how many. We’ve covered some of the – in terms of the developed or kind of higher income countries, we’ve covered off some of the largest ones already in the U.S., in Japan, for example. In terms of the high penetration markets like the UK, we’ve got two clients. So I do expect us to see more. But what’s also critical is the growth in those existing markets, the incredibly high NPS scores that people are achieving as they are rolling out those facilities. We’ve gone from no international facilities to, I think, seven international facilities at the moment. And we’ve got nine new facilities going live next year. So there is a lot of activity. And I think what’s really happened with Ocado Re: Imagined is there is a lot of conversation right now going on with our existing clients and with potential clients, huge excitement about what that brings to the offer, to their economics, to their ability to deploy faster and stuff like that. So lots and lots of activity, I don’t think we’ve ever been busier.
Andrew Gwynn
Okay, thanks very much, Tim.
Timothy Steiner
Thanks.
Operator
We will then take our next question from Nick Coulter from Citi. Please go ahead.
Nick Coulter
Hi, good morning. I have three, if I may, please. Just one by one, please. Firstly, I have a question on the scale of the reduction in capital costs for the 600 robot and grid, please. I guess the frame of reference was the 13% of GMV that you talked about for Erith. But are we talking 10%, 20%, 30% reduction? Obviously, it’s useful to understand the unit economics, please.
Timothy Steiner
Well, we’re talking about an 80% reduction in weight. We’re not talking about an 80% reduction in costs, but we are talking about probably more than the numbers that you were – at the higher end of the numbers that you were talking about on a robot on grids. Obviously, robots and grids are not the entire in-store. You still got the peripherals as well. We’re talking there about swapping out 50% to 80% of the human pick stations for robotic pick stations, and then we’re talking about adding the frame-loading machines, where they have phenomenally attractive returns. So overall, we’re talking about quite significant double-digit reductions in cost of those facilities. Equally…
Nick Coulter
Including everything that you just mentioned?
Timothy Steiner
Yes. Equally, we’re talking about the redesign of the facilities, meaning that they are also for the same amount of throughput, a double-digit percentage smaller, which pretty closely equates to a reduction in the capital investment by the client in their building because they tend to be on a per square foot basis as well as a reduction in the requirements of the remaining space in terms of the things like the thickness of the slabs or the quality of the slabs or the amount of electricity required or things like that. So we are talking about material savings across the board. And then in terms of labor, as we said before, we’re talking about 30%, 40% reduction in the amount of labor required in the building, which obviously is super meaningful in terms of the clients’ economics and the overall efficiency of this platform versus anything else. It’s huge.
Nick Coulter
Okay, great. So what you’re actually saying is that even after you put the picking arms in and remove the human pick stations, you put the frame loading in, you’re still going to end up with a capital percentage of GMV that is lower than you started? I think you said double-digit down. Is that correct? But for no extra fee, I guess, is the question mark on that.
Timothy Steiner
There is some fees – because there is some extra services as well. So things like on robotic pick, you’re running the compute for robotic pick. You’re running the tele operations for the robotic pick for example. So we’ve got two teleoperation centers, one in North America and one in Asia. But currently the robotic pick for the Kindred Systems and are now running the same thing on the – we will run it on the first installations of our On-Grid Robotic Pick. So there are incremental costs. And obviously, there is engineering costs to continue to manage the Automated Frameload and stuff. So fees may go up very slightly, but very slightly relative – i.e., the majority of the savings that the clients have in taking out the pickers and the frameloaders, the majority of those savings are going to the clients.
Nick Coulter
Okay. But will you see lower engineering costs as well?
Timothy Steiner
Nick, we obviously do need to recoup some money for all of this R&D through that you will kind of expand [indiscernible]. We – the engineering costs on the 600 Series bot, the bots represent the majority of the engineering costs at a site. We would expect the long-term engineering cost on the 600 Series to be materially lower than they are on 500. The 500s are getting very close to our long-term targets that we set out kind of a while ago for this business to achieve. The 600s we would expect to be significantly more – significantly better, both because the design kind of manufacture process allows a faster turnaround. If I mean there is a part that you see an issue in, you can do a redesign, a prototype, a test and deploy very rapidly and also because the actual are significantly cheaper. And therefore, if you do replace a motor or you do replace physical component, the parts are significantly reduced.
Nick Coulter
Okay. So that 2% cost that you have below the fee, presumably that will come down by, I don’t know, 25%, 30% or something like that. And that’s the margin accretion that you’re referring to.
Timothy Steiner
We’re not getting into very specific numbers. Yes, we would expect to see some margin accretion. And – but we are, as you know, investing that margin accretion over a number of sites into the R&D that the entire, the clients and their customers, benefit from. So it’s to allow this continued and accelerated pace of innovation.
Nick Coulter
Got it. That’s very helpful. One last one, if I may. Can I ask about the frictional inbound and outbound costs that you get from taking inventory around the Ocado Orbit solution. I guess, you’re obviously enabling a lot of flexibility, but are there any additional costs?
Timothy Steiner
So the way that you think about it at the moment is there is two alternatives, right? One is that you run one big warehouse and you run two spokes. And if you run one big warehouse and you say 40% of the goods direct in small vans and 60% of the goods go out of the spoke site, you actually moved 60% of the goods in a one-way transportation from one warehouse to a spoke site, right? And then you take the trailer back empty. If you now run an Orbit network, you move, say, 66% of the goods rather than 60% of the goods, but you move them where the van is the – the trailer is always full when it’s moving. And therefore, you actually end up transporting less goods, but because you receive the same quantity that you would have done in a big warehouse, one of the three warehouses for each supplier, your inbound efficiency is the same. You transport 10% more goods, but you transport them on less routes because you don’t have the empty legs. You maintain very high levels of availability because if you need to reposition stock after you’ve moved it once, you can do that because you’ve got constant movement going on. And so you maintain high availability, a large range, low waste, and you can actually transport with less transportation routes.
Nick Coulter
Very clear.
Operator
We will now take our next question from Simon Bowler from Numis. Please go ahead.
Simon Bowler
Hi, good morning. First one, can you just talk a little bit around invoice fee growth. I know it came in a bit weak than you’re expecting for the year just gone. Perhaps give a bit more color around that. And I don’t know if that’s a number you can kind of talk or guide to at all for the year ahead. And then secondly, and this is maybe trying to ask the same question that others have been getting asked in a slightly different format. But can you talk at all about kind of the different return on capital profile that you are expecting to see, i.e., kind of capturing your comments around CapEx and margins under the Ocado Re: Imagined saw Series 600 versus what we have seen thus far.
Stephen Daintith
Thank you. I’ll cover question number one, invoice fee growth in International Solutions. This is an item that is driven, of course, by new deals that we announced and the revenues that we get in from that. And the fees we’re getting from that, I should say. I think – and they can be quite binary. So, giving guidance is extraordinarily tricky because it’s down to one or two contracts, whether you get them over the line or not. I think it’s fair to say that we had hoped for more of that happening during 2021. But part of that thinking was aligned around COVID not being part of the year and everyone’s experience. And so that, once again, has restricted travel. Now that travel is opening up, we are in a number of live conversations. So we hope to get those to a successful conclusion shortly. So that’s where we are on that particular topic. And Tim, do you want to cover retail?
Timothy Steiner
So let’s, Simon, I think you know how this works as well I do. If we can get the capital costs of what we install upfront down, and we can do it quicker than also our kind of management costs of doing that are lower. If we can take the same upfront fees from the client, then the net investment is net lower. If we can charge slightly more fees because we’re doing considerably more services for the clients, but we can do it without having higher overall operating costs then the annual return is high. So you’ve got a higher return on a lower investment, and obviously that drives an improvement in the percentage return. The exact amount will need some time to quantify and are something we don’t give guidance on anyway. But it’s hard to know exactly what the cost of maintaining a fleet of 1,000 500 Series robots are because we haven’t printed that many yet. And so, we haven’t operated that many at full scale. But we have a pretty good idea because we know what are the key components? We know what are the key components that we replace on the 500 – on the 400 and 500 series robots, I mean we’ve made some really, really clever moves to make some of the key ware components, either last materially longer, be materially quicker to replace or repair and be materially cheaper in the ware part. So we’re making just phenomenal progress. And if you look to other things like even in the software, where you look at the number of escalations in the last quarter, in the warehouse software, for example, and I think we have something like double the number of warehouses operating and half the number of escalations, you see that the escalations per warehouse is coming down dramatically as we’re improving the quality of the software. [Technical Difficulty] Still there?
Simon Bowler
Yes, I don’t know if it’s my line or yours, but you kind of cut in and out a couple of times during that. I think I picked up the majority of points that you were making. So thanks a lot.
Timothy Steiner
Thank you.
Operator
We will now take our next question from Andrew Porteous from HSBC. Please go ahead.
Andrew Porteous
Hi, guys. I think, three, if I may. I’m just trying to get my head firstly around this idea of being labor markets at the same time, sort of maintaining the economics of existing contracts. I mean is that the right way to think about things that you’re sort of…
Timothy Steiner
Sorry, you have cut a little bit. Sorry, Andrew, could you repeat the question? You cut out a little bit.
Andrew Porteous
Yes. I’m just trying to get my head around this idea of sort of demanding in new lower labor markets and your existing contracts as well. I mean, should we think about you basically make the whole solution cheaper, but you – so you can get new contracts, but you’re effectively giving away some new services to sort of maintain the fee structure of existing contracts. And if that is the case, is there a risk of deflation in your existing contracts? Or are you contractually guarded against that if sort of existing customers don’t want to take the new services?
Timothy Steiner
So I think that’s a fairly good way of describing it. So yes and in terms of – obviously, we – the existing contracts are on fixed prices, where we add in new services, we will add them in at attractive rates for our clients, because we want to give them attractive returns, that mean that they will overall grow and grow their use of the platform need and use of the platform. So, in – some of the products are – I mean you can’t gain the benefits of a 600 series robot being cheaper if you have already deployed a full fleet of 500 series robots, but the 500 robots are already very impressive. It’s just that the 600 can do something even cheaper. But you can go back into that facility and add On-grid Robotic Pick and add Automated Frameload. Overall, between us and our clients phenomenal returns on the labor savings that you will generate and we can split those in a way that’s attractive for us, but extremely attractive for our clients.
Andrew Porteous
That’s very helpful. And then you talked a lot about e-commerce and the opportunity there. Just trying to understand, are you thinking about that as being a new opportunity with new customers like the Getir, Gorillas, etcetera, of the world and therefore, more opportunity in markets that might previously been covered by exclusivity, or are you thinking about helping your existing clients competes in that space?
Timothy Steiner
I mean, in most of our existing markets, more of the latter. Obviously, in a new market, we would consider any client that wanted to approach us and we wanted to talk to, but largely, the latter. The latter have the buying power to scale and the customer knowledge. What they – and what they could do with our facilities is bring an unparalleled offer, so bring an offer that is what you would call it, taking a UK kind of terms to be convenient store pricing, sustainable – and a sustainable profitable business at convenience store pricing, not some hugely discounting thing today, but ultimately 20% or 30% premium. And with a range that would not – would be a multiple size of a convenience store, more like a high street supermarket in a distressed purchasing timeline. So, whereas at the moment for a distressed purchase, you have to pay a significant premium and shop from a really small range. This would be 10,000 to 15,000 range at convenience store pricing and with very attractive economics for the retailer. So, it’s a bit of an unparalleled offer out there that I would expect to see a number of our clients deploy. I think one of our international clients has already stated that they are building their first one of these facilities in Florida, and I expect we will see more.
Andrew Porteous
Absolutely. And then a last one, just on the Ocado Retail side of things. It feels to me like a lot of the confidence in building back margins over the next couple of years is reliant on gross margin sort of staying 200 basis points to 300 basis points above where they were pre-pandemic. Can you just help us with the confidence around that happening and what the sort of building blocks are that are seeing gross margins get to where they are?
Timothy Steiner
I would just remind you that a significant portion of that growth is just the – what we – where we historically paid fees to Waitrose that we don’t pay any more. So, well over 100 basis points is purely that. And then also, we have done some other moves where the team have, for example, taken some of the product data that we collect in-house and are generating more probably by selling that directly ourselves from our own systems to the CPG community rather than through a third-party. So, it’s not kind of increased prices that’s driving that. It’s efficiency and low waste and the long tail and the media income and the investments in the platform that help drive more opportunities to raise that as well as the no longer working with Waitrose. We overall improved our price competitiveness last year. So, when we benchmark ourselves to other retailers, we overall came down in pricing, but we can earn attractive margins. That’s part of what having a long tail of selling 50,000 SKUs in an industry that thinks 15,000 is a lot allowed you to do.
Andrew Porteous
Very clear. Thank you very much.
Timothy Steiner
Thank you.
Operator
We will now take our next question from Sreedhar Mahamkali from UBS. Please go ahead.
Sreedhar Mahamkali
Yes. Hi, good morning. A couple of quick questions, please. I think maybe just going back to some of the debate earlier on about liquidity. I think Stephen, you made the point, which is very well noted about strong liquidity this year. But are you able to talk about CapEx requirements beyond this or maybe paint a picture on a kind of 2-year to 3-year view, because I think in the release, you did talk about you needing further requirement to further funding in time. So, how do you think about it on a 2-year, 3-year view in terms of CapEx requirements versus the CapEx that you have seen this year? That would be very helpful. And secondly, I guess just going back to Andrew’s question on capacity, think most of us still think about capacity you are adding via lens of 65,000 orders per week type CFC. But with all the innovation, I guess the question is, is the overall capacity still the same that we were thinking about, or is it greater? Is it less relative to the contracts that you have signed? That will be helpful just to get the answer on the magnitude of these changes. Thank you.
Timothy Steiner
Can I – I will just go first for a minute, and I can let Stephen talk about specific numbers. But obviously, as we roll out these innovations that we announced from Ocado Re:Imagined, the CapEx per site or per 1 billion of capacity will come down. So, our CapEx will be lower for the same amount of capacity. Obviously, if we build capacity at a faster rate then that will offset that. But capacity required per site will go down. And there is two halves of that CapEx that we see. We see that the side as the solutions provider, which is where we see it in international. In the UK, bear in mind that the moment we consolidate the whole of Ocado Retail, you see both the part where we are the supplier of the Ocado solution, but you will also see the buildings part, and there we are making huge strides, as I spoke about before, in reducing the size required for building of a certain throughput and reducing the costs involved in doing that. So, I would expect to see the ORL CapEx kind of billion of capacity also come down in a very useful way. In terms of your question on site sizes, then yes, we do see more of a mix of sizes, and we expect to see more mix of sizes. And we could see sites from kind of anything from 20,000, 25,000 orders up to kind of 120,000, 150,000 orders being kind of range of sites. And hence, that’s why that Stephen persuaded me that we would go with guidance in terms of modules, so that you could kind of work out how much capacity it was, and no one could pull your wool over the eyes by building kind of 21 module sites and pretending that was equivalent to 2020 module sites. So, we are talking modules with you there. So hopefully, that helps. On the last part, on CapEx, I just want to talk about is on the – is the investment we make in the platform, where we have announced plans to increase our headcounts by almost 500 people during the course of the next year or so and increasingly a heavy investment. You are starting to see what some of that heavy investment is generating, not just in terms of being capable of turning on 10 clients onto a platform and making it work in all of their markets, but also in terms of enormous amounts of innovation. And I would just urge you to remember that we do not need to maintain this level of investment if we are not innovating and creating amazing new opportunities for our business. So, if we wanted to just continue doing what you already understand, we could materially reduce that annual CapEx. We currently have no intention of doing that. But do understand that a large part of that innovation – a large part of that spend is to drive innovation that we are not talking about and that we have not explained and that nobody in your community has yet imagined.
Stephen Daintith
And when it comes to the specifics of CapEx, yes, we have guided to £800 million in 2022. I think it’s probably fair to regard that if we look over the sort of ‘21 to ‘23 period is probably being the peak year of that 3-year period of CapEx spend. I think key drivers, again, when we think about this is really the number of sites that are close to opening or under construction. And if we think about sort of 2020, 2022, we are going to open nine sites in ‘22, ‘23 currently we have around eight or so sites set to open. So, that’s the driver of that £800 million. The CapEx going forward probably will be lower in 2023. But one shouldn’t regard reducing CapEx as a good thing. To Tim’s point just now, CapEx growth is an indication of the order pipeline that we have ahead for automated warehouses. And the more automated warehouses we add to the portfolio of the 10 that we have today the closer we then get to becoming in a business that can generate cash flow and at the same time, invest in growing a portfolio of automated warehouses. So hopefully, that answers your question.
Sreedhar Mahamkali
No, it does. If you we can maybe perhaps just add on a little bit in terms of clearly, we are going from net cash to a greater level of net debt. I guess a couple of questions. How do you think about funding beyond this year? And the second one is relative to the level of net debt, are there any sort of covenants that we should be aware of and things like that, please? Thank you.
Stephen Daintith
Well, first of all, no, there are no covenants that you need to be concerned about in respect of the debt profile that we have today. Thinking about funding, there are a number of routes we could choose to take. The debt markets have been supportive, and we expect them to remain supportive. Similarly, with the equity markets as well. I think we have a good profile of maturities at the moment with several years before the next maturities come up, but no specific plans right now. We will keep an eye on it as the year progresses and monitor this year accordingly and choose the right time and the right way of approaching it.
Sreedhar Mahamkali
Thank you very much.
Operator
We will now take our next question from Rob Joyce from Goldman Sachs. Please go ahead.
Rob Joyce
Hi. Thanks very much for taking the questions. I have got a couple. Apologies for flogging this dead horse, but maybe from another angle, I guess in terms of the Ocado Re:Imagined side of things, clearly the major sort of effort here is to pitch this to new prospects, as you say, and existing companies to really drive the volume of demand there. Can you give us some context around how you are describing the improvements Ocado Re:Imagined gives to your existing customers? So, maybe some metrics that you have shown them or explaining it to them, so maybe we can understand what is exciting from that perspective? And then the second one, just on that following on from Sreedhar’s comments, on an earlier comment from Stephen on the cash flow moving into positive territory. Do you have any – given the outlook you have today and the commitments, when do you think free cash flow moves into positive territory? And maybe, Tim, if you could give us an idea of what you think that steady-state CapEx number could be one you said before? Thank you very much.
Timothy Steiner
So Rob, I think that the key with the existing and new clients really is around two main changes. So, one is when they model something out, they used to model it out with kind of circa 200, is the UPH kind of target, i.e., roughly 4.4 orders per labor hour. People are modeling that out now with six, seven kind of plus orders per labor hour. So obviously, that’s phenomenally attractive, particularly as wages in many countries are under very positive pressure. So, we were already dramatically more efficient than any other means of doing this because that was an end-to-end productivity. That’s those people unloading the suppliers trucks and loading the trucks on the other side. So, I have seen many people comparing it to fractions of the work in the warehouse, driving it up into the 300-plus level, between 300 and 400 is a very dramatic improvement as well as the buildings requirements being somewhat simpler and somewhat smaller means that the incremental CapEx that the retailer needs to deploy is smaller. And so I think that’s what’s driving the economic part of the model. But I think there is an equal or possibly almost larger excitement from some of the clients around the capabilities for using those warehouses and optimized routes by managing to do short lead time deliveries, is driving a huge amount of excitement from our existing clients and some from prospects, and the ability to build smaller warehouses close to more clients and still maintain the efficiencies that has been the kind of the elusive holy grail of this automated grocery kind of business is also generating excitement. So, it’s just a really busy time for us. And we are trying to put as many numbers as we can out for our clients at the moment to just try and help them understand the implications as best as possible, but it is really throwing up a lot of conversations and a lot of excitement and a lot of interest. And of course, you can change your financial models by just saying, I am going to build quick and I am going to ramp it quicker. And that in itself has a benefit on both sides. And so there is a lot to digest, but it’s all positive. And then on the…
Rob Joyce
And just very quickly, would you mind – just on the labor cost per hour. Is that broadly saying if we assume a $20 per labor hour, it would have cost $4, $5 per basket before – or sorry per order before and now it’s looking closer to $3 on those numbers you just gave about. Is that a way of thinking about it?
Timothy Steiner
That’s the right way of thinking about it, yes.
Rob Joyce
Thank you.
Stephen Daintith
And then on the cash flow positive year, I am wary of giving a specific here, but let me sort of try and talk you through a way of thinking about it. So, let’s assume we are building 10, a dozen or so of CFCs every year, and that’s a £500 million or so cash outflow. And then we are investing, let’s say, £200 million in technology, £100 million of group ops costs. So, you put that together, you have got about £800 million annual outflow and let’s sort of disregard working capital and so on in the numbers. I think the way to think about it is the number of warehouses around the world that are generating when they are operating at full maturity and capacity, then generate sufficient cash flows to offset that £800 million or so. And I think the number to think about there is when we hit about sort of 55, 60 or so warehouses around the world. And at the end of ‘23, we are going to be at around in the high-20s, so about halfway through. So hopefully, that gives you a guide then. And you will see, as we announced for the committed sites and when they are going to go live but you ought to be thinking about it, a period sort of 4 years or so years out from now – 4 years or 5 years from now, reaching that pivot point, where look, we have got a sufficient portfolio of warehouses to continue to invest and cover our cost base. That’s the – probably the best way of thinking about the whole exercise.
Timothy Steiner
Stephen, it would be fair, wouldn’t it, to say that kind of 20 to 30 would mean that you were breakeven on the level of innovation that we are trying to drive, which is huge at the moment.
Stephen Daintith
Yes.
Timothy Steiner
And then the next, as you spoke about the next 20 to 25 or something like that or 30, allows you to grow 10 a year on a self-financed basis. And then that model just keeps growing on that basis.
Stephen Daintith
Well, exactly right, yes.
Rob Joyce
Very, very helpful. And then very last one maybe that steady state CapEx spend.
Timothy Steiner
Look, we are not on a firm number. But whatever number you thought it was, imagine that we can deliver at least at no more than that number and hopefully less to deliver the efficiency that you just spoke about before in terms of the reduction of costs on the client side. And then we also hope that there is a reduction in the largest part of the engineering cost, which is the maintenance of the robots that we spoke about that – as we have talked about today in the numbers of the KPI is coming down towards our long-term target with the relatively new 500 series, an order of magnitude in terms of kind of – about 10 – as long as it took the 400s to get to these kind of levels. So, that’s doing phenomenally well, but we would expect the 600s to significantly outperform the 500s on that basis.
Rob Joyce
Thank you.
Operator
We will now take our next question from Xavier Le Mené from Bank of America. Please go ahead.
Xavier Le Mené
Thank you, gentlemen for taking my questions. Two, if I may. The first one, just understanding the 600 series and what it can change it is in signing new partnership, because it sounds like potentially, you are waiting for that series to be live. So, does it mean that potentially new partnerships could be back-end loaded, i.e., we need to see the 600 series running before you can convince new partners to potentially buy your technology? So, I just want to understand the kind of timeframe here. The second one is about non-food. You sounded a bit more vocal about the non-food opportunity. Can you confirm that you already have discussion about that or that’s more the prospects of starting discussion about non-food going forward?
Timothy Steiner
Look, the 600, 500s, we need to have availability of robots. If we sign a client up, we make a decision on the kind of pricing and how much incremental actually we are going to put in the building to drive their efficiency at what cost, at what charge effectively. And we are making an assumption on the amount of 500s and the amount of 600s we will deploy in that building or if we will only deploy 600s. And so yes, if somebody wants a building that’s going to go live in the next six months, it’s a 500 series building. If they want one that’s going live in 18 months, it’s probably a 600 series building. And if they want to go live in 2 years, it’s definitively a 600 series building. And so we just have to take that into account. It’s not the client doesn’t need to take a risk on that. We take a risk on that, because we are building one series and we will migrate to building another series. On your non-food question, we have had inquiry from non-food for many years. We have kind of batted it back, because we were aware that whilst our product had a multiple of the throughput of any competing product per robot, the robot itself was more expensive. So, the throughput curve kind of cost of robot may have been similar or in our favor. But what was different about ours was that you could get massively more throughput from a density. And that was – that is critical for grocery, but not critical for many general merchandise or storage or kind of other use cases. But then the machine that you house it on was similarly kind of over specked, but you didn’t get any benefit if you don’t want that throughput. By driving down the weight of the 600 series by 80%, we drive down the weight to the lower – the lightest bot in the industry, but even with the faster acceleration that we generate with it to do more throughput, still we generate less forces, and we can build the cheapest grids in the industry. So, the opportunity now to do anything else is present. It’s like the point where – at any of these innovation examples. But once your digital sensor, it’s higher resolution 35 mm film then suddenly, you have opened up another – that wasn’t why you designed originally, but you have opened up another opportunity. This is the fastest throughput, lightest bot that can sit on the lowest grade effectively grid and therefore, is the best storage and retrieval system available.
Xavier Le Mené
Okay. Thank you.
Operator
We will now take our next question from Tom Davies from Berenberg. Please go ahead.
Tom Davies
Good morning guys. And just three questions for me. So firstly, for like a standard CFC, can you give that CapEx split between like long life CapEx on the grid and then like the shorter life CapEx on the robots and the on-grid robotics? And then secondly, does the Series 600 robot improve the throughput per robot. So, it means that less robots are required from Ocado’s perspective. And then finally, for the innovations from a solutions perspective, you are obviously facing a trade-off between revenue and margin. So, how do you make sure that your partners keep on building that additional capacity from the lower operating costs?
Timothy Steiner
So, I think there is three questions there. Let me just try and make sure I have got them. The first one, there was the split between bots, grids and other stuff and the kind of life of those assets. Is that about right?
Tom Davies
Yes. That’s right.
Timothy Steiner
I think it’s roughly rule of thumb would be to say it’s something like a third grid and bins. It’s something like a third robot, and it’s about a third the other peripherals in the building. The grid and bots, obviously, the grid’s long life, the peripherals are probably – to be honest, the type of equipment and the peripherals are generally, we are still running the original – a lot of the original versions in Hatfield 20 years in, but they have an average life of about 10 years. The bots are probably a bit smaller than that. But again, the overall average is about 10 years. The thing with all of that stuff, though, is like with the bot, for example, there are ware parts on them, but we run the ware parts, that’s the engineering cost that we look at. That’s changing the tires or swapping out the batteries or that type of thing. Ultimately, you run these like an airframe of an airplane, and they should last for much longer than we currently assume and that we depreciate them too. Your third – or your second part just remind me was…
Tom Davies
Do the robots improve the throughput?
Timothy Steiner
So, the 600 series robot is designed to have the same physical characteristics as the existing robot. So, it’s not designed to do more work per se, although we do expect that if it – if when it becomes more reliable than the existing robots, then it will do slightly more work because you have less robots that are coming off the grid and asking for a bit of maintenance. And if when they do want maintenance, it’s quicker, etcetera. There is a benefit there. And because the robot is smaller and lighter, it will need less energy to move itself around. So, it will spend a little bit less time charging. So, there is some benefits there. The main benefit, though, is that the On-grid Robotic Pick allows us to lay out the sites differently and to remove some of the critical congestion on the sites that allows us actually to put more robots in the same space and generate more throughput overall. The On-grid Robotic Pick may also be more efficient in the use of the robots. But that’s something that I think we are working on now to understand. And we are constantly improving all the software that runs our systems and driving robot efficiency up by a few points here and a few points there anyway. But this was – it’s more – the 600 series is more about the reduction in weight, the reduction in costs, the reduction in forces and what that means to crash barriers, slabs, grids, and things like that, that ends up in quite a material reduction in other costs, plus the reduction in their own costs.
Tom Davies
And then just the final one from – about the trade-off between revenue and margins and how do you let ensure the partners – make sure that you keep deploying that additional capacity?
Timothy Steiner
Look, I think the point with the partners is the more attractive we can make the economic model for them and the more missions we can help them to serve, and the better we can help them to serve those missions better than any competition that they have in terms of allowing them to sell fresher food, bigger ranges, better accuracy, better UI, better interfaces and to cover all the different missions from ultra-short lead time on e-commerce through short lead time, two-hour, three-hour, four-hour deliveries to later today, to today for tomorrow, etcetera. The more we can help them to do that at the lowest cost in their market, the more they will drive market share. There isn’t a grocery retailer in the world that I don’t think that doesn’t try and drive growth in their market share and take advantage of competitive advantages to do so. So, it’s a question of innovating, creating competitive advantage, sharing the competitive advantage with our clients, they go out and use it in the market, and it creates the virtuous cycle of growth.
Tom Davies
Thank you.
Operator
We will now take a follow-up question from Simon Bowler from Numis. Please go ahead.
Simon Bowler
Hi. Apologies the call has been going a little bit, but it’s a slightly off-peak question as well. I was just wondering, with some of the innovation has come through with Series 600 and so on, have you in any way, been in conversations with your growth is about using a CFC to top up or fulfill any of their physical networks? The entire focus has been very much on driving their online business, all of certain aspects of the range they want to hold in a physical store, we would have tend to be a more efficient route to fulfilling the store via CFC, via their existing manual distribution centers.
Timothy Steiner
So Simon, it is something that comes up from time-to-time, and it is something that several of the clients have asked me in the last week or so, because of the robotic pick in particular, as well as just more facilities. Because you still build facilities to a peak day, you obviously have effectively downtime in those facilities, particularly with robots. We have got robotic pick as opposed to human pick. The incremental cost of doing some sortation and some organization in there effectively is incredibly low. And so where – these things are not necessarily the right beast to send to a large hypermarket, the fast-moving part of the range, but they are very interesting in terms of either running larger ranges in small shops or big shops, frankly, and doing that more efficiently than you can do that anywhere else. So, they are extremely interesting to do store replenishment of singles, of things that you don’t want cases. So, the existing store replenishment network is mainly focused on cases, which in some places, limits what they can put in a store, and therefore, you could drive greater sales in your small store formats by allowing yourself to replenish in singles and effectively have more lines in there, or in some instances, we have clients who use like the wholesalers to try and – to add in the last parts of the range, and they could do that much more efficiently themselves and drive some margin improvements on that side. So, it is an area of interest. There are a lot of areas of interest of how to utilize this stuff. And some of it’s an area of prioritization for us and our clients so that’s the most important thing for them to achieve first. Would I be surprised if our facilities will contribute to that in the next 3 years to 5 years in some people’s store networks, I would not be surprised. No.
Simon Bowler
Okay. Thank you.
Operator
There appears to be no further questions at this time. I would like to turn the conference back to the host for any additional or closing remarks.
Timothy Steiner
I would just like to thank you all for participating today and look forward to speaking soon.
Be the first to comment