NEXT plc (NXT) Earnings Call Transcript & Summary

Good morning to everybody, and welcome to the half year results. Thank you for being here, and a special welcome to anybody who's here for the first time. It's certainly a pleasure for me to kick off the proceedings given that the NEXT results continue to be very positive. The first half has been a strong period for the company with continued sales and profit growth, especially in the International area. Now this performance doesn't come by accident, and it certainly reflects the hard work and courageous decision-making of all our employees worldwide. And I want to thank them. And I'm going to leave all the details of what has happened in the first half to our Chief Executive. So over to Simon.

Right. Thank you, Chairman. Good morning, everybody. Total group sales up 9%. Full price sales up 7.7%. The difference is not the subsidiary companies growing much faster. It's all about markdown. Last year, we had a very small end-of-season sale because we exceeded our targets by a significant amount. This year, sale stock returned to more normal levels. In terms of our expectations, we thought we were going to be up 4%. You can see that we exceeded those expectations in the U.K. and Overseas, Much more so in Overseas. And I'll be talking about both of those in more detail as we go through. Profit up more than sales, up 10.5% Net interest significantly higher than last year. That's all about the fact that last year, we had a lot of cash on deposit because we weren't able to buy back shares. This year, we bought a lot of shares back at the beginning of the year, which means we didn't get the interest income that we had this time last year. Profit margins short by 0.3%. Profit after tax, pretty much the same, no change in tax rate and earnings per share up around 2% more than underlying earnings, and that's as a result of the -- mainly as a result of the buybacks that we did very early on in the current financial year. Interim dividend, we're planning to increase in line with earnings per share, up 12.6% to 98p. Moving on to cash flow. And just to reemphasize the cash flow and balance sheet, I'm going to talk here, unlike the P&L and all the other numbers I'll talk about, this is done on a consolidated basis. So this is done on the basis that we own all of the subsidiaries, most of which we only partly. So good cash flow from operations and profit, GBP 54 million. CapEx, up GBP 44 million on last year as expected. What you can see here, we showed you this graph at the beginning of the year hasn't changed much. The big increase came in warehousing, where we spent much more than last year and less on stores. This time last year, we spent a lot on Thurrock and a new concept store, which we didn't have again this year. In terms of CapEx going forward, we're expecting roughly the same amount of capital expenditure for the next 3 years. And at the end of last year's -- sorry, the presentation 6 months ago, we talked about how that corresponded to capital consumption pretty much in line with our 20-year average. In terms of working capital, working capital up pretty much in line with sales in the first half, which is what we expect because a lot of that is stock. Corporation tax, a bigger increase than you'd expect. It has gone up more than profits. That's all about timing because we pay tax on account. And last year, we were consistently increasing our profit expectations, which means that our tax didn't quite keep up with the result that we delivered at the year-end. Surplus cash down GBP 26 million on last year, still its a GBP 180 million of positive cash flow and the difference in this year and last year, mainly about CapEx. This is where the big change came. Last year, we pretty much locked out of the market, and we thought we'd take a slower steady approach to buying back shares. This year, when we had the opportunity, we bought as many shares as we could. So we've done GBP 355 million of share buybacks in the first half. That means net cash outflow in the first half is GBP 177 million. For the full year, we're expecting that to reduce to around GBP 100 million, which is the amount we plan -- the cash outflow from the business that we're planning for the full year. All of that is funded by a planned increase in debt. And just to explain that in a bit more detail, we started the year with leverage at 0.6x, [ GBP 713 million ]. Now these numbers are -- where we target is fairly arbitrary, but we had targeted 0.63x. The reason it was lower than our target was because we generated a lot more cash in January than we expected as a result of better sales. We aim this year to push that back up, the leverage back up to 0.63x, which will push our year-end debt to GBP 815 million, and that accounts for GBP 102 million net outflow. In terms of how we get there, we think we'll have very strong operating cash flow income, around GBP 996 million cash inflow from operations, GBP 245 million of CapEx GBP 319 million of ordinary dividends. And that leaves around GBP 500 million to distribute, of which we spent GBP 355 million. The GBP 180 million balance, we will either give back a special dividend, buybacks or some other form of capital distribution. In terms of our cash resources, we're very comfortable with cash resources. We started the year with GBP 1.2 billion of cash resources. We have increased our RCF by GBP 200 million. And the lion's share of that increase, about half of it will be used in October to pay off the 2026 bond. And that will leave us with GBP 1.3 billion of resources. And if you compare that to our peak cash requirements, we've got about GBP 300 million of headroom, which we think is comfortable. We have increased our net debt partly to account for making sure that not only does NEXT have sensible of headroom, but so do all the subsidiaries that we lend money to that have headroom sufficient to get them through a blip as well. In terms of [ bad ] debt, it is much more than balanced by the financial assets that is our customer receivables. So we've got GBP 1.36 billion of customer receivables, which is much more than our peak headroom and broadly equal to our cash resources. Moving on to the balance sheet. Balance sheet investments went down in value by GBP 42 million. This is a wonderful piece of accounting where one of the reasons that the balance has gone down is because they've become more valuable. Amortization takes GBP 19 million off. GBP 10 million is because joint ventures have paid a dividend to us. And final thing, the provision for minority acquisition. This is because a lot of the minority shareholders and management teams have options to sell their shares in their business on a fixed multiple of profits back to NEXT. Those options vest over the next sort of 4 or 5 years across various different businesses. Because those businesses are doing much better than expected, we have to provide a higher number, which has the ironic effect of reducing the value of the more valuable businesses that we want to buy. Stock, up 5.8%, pretty much in line with our sales expectations for the second half. Customer receivables, up 2.5%. Now these are sort of the beginnings of an important story, which we're going to elaborate at the end of the year here because what you can see is that unlike the last 10 years, our credit sales are beginning to grow much faster. So we would normally expect that number to be plus or minus 1% or 2%. Our credit sales are now growing at 6.8%. The reason it's growing so strongly is partly because we're offering a new product, which is our Pay in 3 offer. Pay in 3 offer allows customers to -- it's a very similar [indiscernible] type offer, allows customers to buy the goods. And if they pay off in 3 installments and pay all on time, they pay no interest. If they don't pay those installments on time, if they choose not to, they can do that, but they then incur interest on the balance that they haven't paid. Because those Pay in 3 accounts by design, pay down much faster, it means that our credit -- our receivables don't grow by as much as sales. And we expect that to be a sort of continuing trend as we move forward. These balances are n-- incur less overall debt, and we think are likely to be to incur less bad debt as well. So I think going forward, you'll see credit sales rising faster than balances with perhaps some benefit on bad debt. I'd say perhaps, we've yet to prove that. Other debt is up GBP 53 million. Our biggest number here is international aggregators. This is the fact that on websites like Zalando, they sell our product, take a commission and then pass us the balancing proceeds a month later. That month as we've grown so fast, that month that they owe us of net sales is growing. So that's a good thing. Cash in transit, I think I mentioned this last time, one of the very few accounting standard changes over the last 20 years that I can remember that makes a lot of sense. And this is because the -- we are no longer able to count the cash that we've taken on credit cards that we haven't yet received. We're no longer able to count that as our cash. And that change on last year cost us GBP 31 million. And then the subsidiaries have been better at collecting in their debt as well. Credit is moving in the opposite direction to what you'd expect. We'd expect creditors to grow in line with sales. What's happened here is that the 53rd week puts a large payment week into the first half that wasn't there last year. So that's all about the timing of payments rather than any underlying significant change in the creditor base. That leaves net debt, GBP 890 million, all driven by the -- and the increase is all driven pretty much by the timing of buybacks. In fact, the GBP 352 million is only GBP 3 million of what we spent on buybacks. Moving on to the detail of the business. And just to remind you, we don't -- we're treating the U.K. Online business separately from the International business because they have very different moving parts. Starting with U.K. Online. U.K. Online was up 7.4% the full. The total sales up 8%. And what you'll see on all the Online businesses is that the markdown sales grew faster than full price sales, which didn't happen in Retail, and that was a conscious decision for us to push more of the changing balance of markdown into our Online channel rather than Retail. Full price sales accounted for GBP 83 million of growth. And just looking at how that breaks down between the different types of brands we sell on NEXT. So GBP 14 million came from NEXT, which grew at just 2.1%, on wholly owned brands and licenses by a really strong 33.5%. I'll be saying more about that later and continued growth of our third-party Branded business, driven largely by better selection of existing brands rather than new brands. In terms of margin, margin up just 0.1%, but a lot going on underneath the surface here. Bought-in gross margin down 0.2%. We've got two things pulling in opposite directions here. First of all, the underlying bought-in gross margin on NEXT stock went up. And we consciously did that in the U.K. to pay for inflationary wage costs and overseas, the change in prices, which we'll see later on, driving up margins to pay for increased fuel surcharges. But underlying NEXT product, 0.5% up. The impact of mix because we're growing our [ wobble business ] and third-party business much faster than NEXT pushed net margins down because you make lower bought-in gross margins on those. And just to give you a flavor for the different net profitability of those 3 businesses in the U.K., NEXT makes around 21% net margins. That's net margins after accounting for the allocation of all fixed overheads. WOBL lower than that at 18%, still very respectable, but 3% lower. And third party, as you'd expect, because we don't put effort into building the brand, we make lower margins on that at 12%. And it was the growth of WOBL and third party that offset the growth in the underlying margin of NEXT. Markdown adverse movement, we had more stock going into the end of season sale and our clearance rates online, which dropped a little bit against last year. Warehouse and distribution, a significant gain here, but again, lots of different things going on, wage inflation and fuel inflation between them adding 0.8% to costs. We've got big productivity gains, 0.6% on productivity and 0.3% leverage over fixed overheads. And really, you need to take the productivity and fixed overheads together because the fixed overheads include all the depreciation on the money we've been spending on mechanization. And both of those numbers put together a testament to the efficiencies that we're now getting out of the mechanization and investment that we've made in our new Elmsall 3 warehouse. And then returns on average selling prices meant that we handled fewer units to achieve the same sales, and that pushed warehousing costs down by 0.5%. Again, that's partly as a result of mix on the whole third-party and WOBL business is more expensive than NEXT branded stock. But also it's that continued effort to weed out the high-returning low-priced stock, particularly within brands that pulls our profitability down. Getting leverage over technology. And you can see that warehousing and technology contributed 0.7% towards margin, and then we pretty much spent all of that on marketing. I should stress that it's not that we build up a pot of money and then spend it on marketing come whatever. The marketing -- the returns justify us spending that much money. But you can see in terms of the shape of the business, what you're getting to is a business that spends less on facilitating and serving the customer and more on telling the customer about the service. So the shift out of technology and warehousing into marketing. So longer term, I think that is a theme that you'll see continuing in the U.K., but we'll see more strongly overseas. And just to remind you that the return per pound spent on marketing and the way we measure our marketing is we look at each campaign, measure what we think of the incremental sales, which is not an exact science, look at the incremental profit on those incremental sales, depending on the mix of that particular campaign in terms of product mix and returns rates and then compare it to what we spend. We have to spend at least -- we have to make at least GBP 1.50 of incremental profit before fixed overheads, every pound spent on marketing. And basically, as long as we can do that, we'll spend as much as we can. Central costs, a big gain here. This is mainly about last year's exceptional performance leading to an exceptional staff incentive payments at the end of the year. So that normalizes this year. Looking forward, if full price sales were up 5.2% in the U.K. Online, we expect our margins to nudge forward by around 0.2% with a very similar story for the full year that we've had for the half. Moving on to International. This is a little bit more exciting. Full price sales up 24% total sales, including markdown up 26%. That number really only tells half the story because the first quarter was adversely impacted by disruption in the Middle East. So you see that the second quarter was much more exciting. Don't get too excited by that second quarter number because there was definitely -- we could see this in the Middle East. There was definitely pent-up demand in the Middle East that contributed towards that 37%. So don't assume the underlying growth would have been 37%, all things being equal. Full price sales were GBP 133 million Overseas. In terms of how that breaks down by region, lion's share of the growth, nearly GBP 100 million coming from Europe, which grew very strongly at 28%, partly driven by the step change in our sales on Zalando as a result of the [indiscernible] integration. Middle East, 14%. But again, you'd have seen an even bigger swing first quarter to second quarter in the Middle East. We've mentioned the United States for the first time. We've never really talked about the U.S. before because we've really had no traction. We have a very small business in the United States. But we have just managed to find -- this year, we've really managed to find productive ways, profitable ways of marketing the business, and we're seeing very significant growth in the United States. Still small numbers, nothing to get excited about today because the numbers are so small, but it does bode well for the future of the business in the States. I should say actually, Rest of World, if you that 3% growth, that is 2 countries have pulled that back, and we're not sure why. Kazakhstan and Australia. If you know why those 2 businesses underperformed, please let us know because we're racking our brains. In terms of the breakdown between the different brands, what you can see here, much stronger showing from the NEXT brand, up 16% overseas. WOBL, astonishing growth, 82%. Part of that is that we have put more options of our WOBL brands on our overseas sites and customers are beginning to find them and get used to them. And then third party, very respectable as well, but much, much smaller overseas because most third parties will go to a local aggregator rather than to NEXT. Profit margin down 0.4%. And as with all businesses, when you've got a bad number, the good thing is to blame it all on one thing that you hope will go away. So we'll start with that. We think the Middle East conflict in the first half cost us 0.8% and that is the balance of the increased surcharges, which would have cost us 1.3% and the price increases that we put through, which because we wanted to see how the war would pan out before we put prices up, we only managed to put prices up sort of half the season. So that's why we didn't quite cover the cost or didn't cover the cost of those surcharges. As we move into the second half, those things will balance out, and the Middle East will be cost neutral. In terms of conflict, we will be cost neutral because the price increases were put through. [indiscernible] gross margin up 0.4%. The same but different story here. Underlying NEXT margin up 0.5% impact of mix, not nearly as adverse as it was. And the reason for that in simple terms is because we make -- relatively, we make a lot more margin on our WOBL brands overseas, where they appear to have much more pricing power than they do in the U.K. WOBL brands make 20% compared to NEXT at 14%. Now you would -- you're looking at that and thinking, well, if WOBL brands have grown by 82%, that should be pushing margin up, not marginally reducing it. The reason is -- the reason it doesn't is because actually there are 2 competing factors. Yes, Wobble grew by more, and it does make more margin than NEXT. But last year, it was at 24% net margins. And we consciously took the decision to lower prices to become more competitive in our wobble brands last year. So those 2 effects pretty much offset each other. Aggregator commission, we're getting better rates of commission from our partners. Warehousing, technology central costs, I could go through all of those in detail, but I would just be repeating the sorts of changes, the movements that we've had in the U.K. because the story is pretty much the same. And again, all of those gains invested -- more than all of those gains invested in marketing at 1.5%. That leaves sort of overall margin movement of minus 0.4%. And we think a very respectable margin for the business to make full year. We anticipate margins around 15.1%, flat on last year. So that is because the price increases will pay for the surcharge in the second half in a way that they didn't in the first. In terms of customer base, some interesting things going on here. U.K. credit and cash, for years, you'll have looked at this and seen cash growing much faster than credit. The reason that, that's changed is because of the Pay in 3 product I talked about earlier. And some of that -- the reason the cash number isn't up by more is because a lot of the credit customers will trade with us first on cash. And then after they've transacted a few times, we'll convert to Pay in 3. So that net reduces the cash customers. But total U.K. up 7%, pretty much in line with sales. International sales up 29% total excluding aggregator is up 14%. In terms of sales per customer, there's no real story here other than the fact that there isn't a story is important because overseas, given the level of growth we've had in sales, you would expect to see our sales per customer moving backwards because new customers tend to spend less than established ones. If you fill up with a lot of new customers, that will push the sales per customer down. The reason we think it hasn't is because of the increasing choice of product on the website, particularly the WOBL brands, which have, we think pushed sales per customer up. Moving on to Retail. This is our new store in Bluewater. And you might be thinking sort of, oh, here's another one of those Thurrock. And physically, it doesn't look the same, but financially, it's much, much better. Because when you look at -- I know a lot of you are thinking, oh, yes, white elephant. Good joke. It's actually -- the economics of the Bluewater store were much better than they were in Thurrock and they still make a very healthy return on the capital that we invested in. It's already open and delivering sales ahead of our expectations. In terms of space for the full year, we expect space to grow by around 1.3% as a result of opening 8 new stores, 6 of them are open. And if we look at the 6 that we've opened and the forecast that we've got since opening, we think the internal rate of return on the investment will be around 30%. So much healthier than the portfolio we opened last year. Retail sales down 0.4%, full price down 1.7%. We didn't put more stock into the end of season sale in retail, not significantly, but actually, we did have better clearance rates. New space was 1.6% and like-for-likes minus 3.3%. That number, although it's bad, is better than we're expecting. I know that's no constellation, but we're expecting it to be around minus 5%. Margin up by 0.4%, it's a bit of a story here. Bought-in gross margin, this was the planned increase in bought in gross margin to help pay for national insurance and national living wage inflation. Markdown was flat. Payroll at all of the gross margin gains and that -- the increased costs of themselves would have eroded margin by 1%, but lots of productivity measures that we've taken in our stores have contributed to around a 0.4% gain in productivity in shops. Store occupancy -- the new space is more expensive than our existing space. That's partly because on the whole, it's slightly higher rent, but mainly because none of its assets have been depreciated. So that when we open a new store, you've got a full depreciation charge. A lot of our existing stores have low or no depreciation charge. And you can see the expenditure on depreciation on new stores is largely offset by the stores that are now fully depreciated, which gives us a 0.5% gain in the opposite direction. Warehousing and distribution, wage inflation and fuel, but a relatively modest impact because warehouse and distribution is a much smaller percentage of retail sales than it is of online sales. Technology costs, there's again here slightly higher than in online. And I'd love to say that, that was because we've had a big review of all of our tools and communications networks. It's not -- it's really because we've rebalanced the allocation of stock between retail and online to get -- to make that allocation more accurate. So Online gained about 0.1%, Retail 0.2 between the 2 of them, it's about halfway between the 2 of them. [indiscernible] central overhead, same story there about staff incentives. And that's what gives us our sort of net margin movement up 0.4%. If we look forward to the full year, assuming that sales for the full year are down 0.9%, our margins will -- we think will come in at around 10.2%, so just over 10%. If you would look at that 0.9% and think that I would look at it and go, that looks ridiculously optimistic because it's much, much better than the first half -- to give you some comfort on that, if you look at the difference in quarter 1 and quarter 2, you can see that quarter 1 was where we took the big hit, and this is because last year, the summer came early. So you got a big benefit in Q1 last year. Q2, we had the same warm weather as last year, but we also had the competitive disruption last year, which we think benefited the stores. And so we think if we take the Q2 number and flow it forward into H2, that is a sensible guess. But if you were to ask me, what is the one number that has -- that you're most worried about is that number. I think that might be optimistic. Moving on to Total Platform. Total Platform has had a really good -- the subsidiaries had a really good half year. Profits up 40%. You need to discount half of that GBP 8 million growth because it was all about provisions that we took in the first half of last year. But underlying profit growth in the subsidiaries of 18%. And what we're finding is that the businesses that were doing well last year are doing better this year and the businesses that were doing badly last year or were struggling are doing much better and one in particular has gone from loss to profit this year. The services on Total Platform profit is up by 23%. That looks high, but it corresponds to the growth in our partners' Online businesses, which is what we charge them for. In terms of the margin on our services, we make around just under 20% on what we charge our clients, which amounts to around 6% of their online sales. Looking forward to the year-end, we expect another GBP 15 million of additional profit in the full year from our Subsidiary businesses. And we -- if that comes through, then the return on capital on all the investment we've made, both in buying those businesses, funding them and also on the CapEx for Total Platform is around 26%. So it's looking like a very good -- as a portfolio, it's been a very good investment. Moving on to full year guidance. These are the H1 numbers. 3.6% in the U.K. We are anticipating H2 relative to our expectations, we have lowered our expectations for the U.K. And obviously, part of my assumption at this meeting is to depress everyone a little bit. I've seen too many people smiling. So I just want to explain what -- why we are cautious about the U.K. And it's a combination of the fact that fuel inflation, in particular, but other forms of inflation as well, look like they will begin to bite harder in the second half than they have done in the first, and that's going to put pressure on the consumer. And unlike past squeezes where government has been able to intervene, we think that there is really no room for government to move. In fact, worse than that, we think that they may have to -- the problem is going to be for them funding GBP 100 billion deficit that they've got. And if you look at these three graphs, they kind of tell the whole story, spending as a percentage of GDP has not been as high as it is today for the last 65 years other than in the oil crisis, the financial crisis and COVID. So it doesn't look like there's a lot of room to increase spending. In terms of debt to GDP, we haven't seen debt levels in the U.K. this high since we were in the shadow of the second war. And tax as a percentage of GDP is higher than it has been for the whole of the last 65 years. And we think that's important because, again, potentially unlike in the past, we think any attempt to increase any taxes is likely to have some negative knock-on effect on the economy. And we've gone into a little bit of detail about that in the tax. So I think tax increases will be self-defeating if they're used to fund stimulus. And if they used to fund a government deficit, then they will place a further drag on growth. And all those things put together mean that we think it's wise to trim our expectations for the second half. In terms of International, we were at 14%. We've gone to 20.5%. You might look at that and go, oh, well, they're still being a bit cautious because they're up 24% in the first half, and that was with the Middle East war. There is one factor that I just need to remind you of, and that is that this time last year, as we went into the second half of the season, we got a huge boost in our aggregation business from Zalando. So aggregator business in the first half of last year was up 33%. In the second half, that jumped to 61%. As we begin to annualize that number, that growth will begin to reverse out. And you'll see the beginnings of that as we move through the presentation. So that's why we are more optimistic about our International business, but not as optimistic as the numbers that we delivered in the first half. And that gives us 6.7% for the full year. In terms of what that means in terms of profit, the growth in online sales, we think, will deliver GBP 116 million worth of additional margin. We will lose -- assuming we hit our targets, we'll lose GBP 6 million from a slight decline in Retail sales. That gives us GBP 110 million, add GBP 15 million for subsidiaries. Then in terms of cost increases, there's lots going on here. And it's important, I think, to separate them out. So the GBP 44 million increase in marketing, it's not -- this is a willful act of cost increase, and we consider it to be an investment because all of that GBP 44 million has a return attached to it. Not all of it will come in the current year. So obviously, the customers we recruit this year, their second order and a lot of the profit is only made next year. Then true underlying inflation that we can't do anything about is around GBP 70 million of fuel and wages. And then the higher interest costs are really about the capital returns. We're paying higher interest costs because we don't have the interest income from the money that we had on deposit this year -- sorry, last year, but this year, we have given back to shareholders. In terms of cost savings, stroke margin gains, you can see GBP 37 million of margin gains, which go a long way towards paying for the wage inflation, lower employee incentives because last year was so high and the warehouse and distribution efficiencies coming through at around GBP 22 million. And that number is -- that estimate is higher than it was at the beginning of the year. That gives us about GBP 1.255 billion of profit, up 8.4%. In terms of what that means for shareholder returns, post-tax EPS, we're expecting to be in the order of up 10%. And if we add dividend on top of that to look at TSR, which is what we ourselves measure as a measure of the sort of total return to shareholders, we think that will come in at around 12.6%, which we think for old-fashioned sleepy retailer is quite a good number, but particularly exciting given that last year, the equivalent number was more than 20%. So we weren't expecting to deliver as strong returns this year. So that's all I've got to say about numbers and guidance. Moving on to what is the more interesting, but not necessarily interesting part of the presentation. I'm going to talk about two things, a little bit more insight into the numbers, in particular, focus on our Online customers and then three areas of the business where we've got, we think, exciting things going on. I just want to share with you sort of some of the things we're doing. In terms of the insight into our numbers, if you take the total GBP 203 million of growth we've got, there's a brilliant page on Page 6, which tells you pretty much everything you need to know about NEXT. It gives you by product category and territory, the growth of all of our businesses and the percentage of our business that each one of those segments provides. When you look at that, I think there are some things that need to sort of calling out. The first is that in the first half, more than 2/3 of our growth came from non-NEXT brands, WOBL and third party. When you break that down, it's the wobble that has performed best. And that's really important because although they're non-NEXT brands, the WOBL brands are owned by NEXT. We buy the stock in on licenses, although we pay a royalty, we own the stock, we develop it, we buy the stock, we take the stock risk. So when you look at the net margins of the WOBL business, and this is the net margins balance blend of overseas and U.K. and compare the 3 businesses, you can see that NEXT is at 18.3%, WOBL at 18.8% and non-NEXT at 12%. So I guess that whilst the -- it might look worrying from a margin perspective that it's non-NEXT brands that are growing the fastest because the lion's share of that growth is delivered by brands that are owned by NEXT, it's actually good news for margin. Incidentally, we had a very exciting conversation about acronyms because we thought we've got NEXT owned brands, which we thought we could call Novel. But I was banned from doing that. So there we are. It's next owned brands. The WOBL brands have done exceptionally well. And you might expect me to talk a lot about them. I'm not going to because for those of you who had to sit through the last 6 of these presentations, I've done -- talked a great deal about WOBL brands and what we're doing to develop new licenses, new brands that are either starting or buying, creating an environment that is a brilliant place to incubate and build brands. I've talked about the fashion price mix of the different brands and how we're trying to make sure that they don't compete -- that they add something new to the NEXT customer's wardrobe. So I'm not going to talk about that. But what I want to show you is it doesn't mean that it isn't an area of the business that we're working really hard on. We still think there's lots and lots of opportunity to grow our WOBL brands. Focusing the Other number that I think is the number that sort of sticks out is the NEXT brand in the U.K. was down, around GBP 7 million, but it was down. And that number looks worrying because I think the question that it poses is, well, is there something fundamental about the NEXT brand that is on the way. We don't think there is. And in fact, we think the NEXT brand overall is in better shape than it was this time last year, but that number does need some explanation in the U.K. First thing to say is Online, obviously, we were up, but only slightly. That number of the GBP 14 million increase needs to be taken in the context of the GBP 70 million increase in the sales of non-NEXT brands. we work very, very hard to ensure that the brands we have are offering something different. But inevitably, there must be an overlap. So the fact that those -- the fact that we've parked so many powerful competitors on NEXT front lawn means that we think that some of -- that NEXT wouldn't -- would have grown by more had those brands not been there. So we think that 2% is not a fair reflection of how much better or worse the NEXT brand is than last year. And in Retail, the -- this time last year, we think that we got a big gain from competitive disruption in the second quarter. and that obviously reverses out. And if you look at the 2-year number for Retail, it's up around 2.3% for the NEXT brand. So we think taking those two things together, we think that we're not concerned about the NEXT brand, particularly as when you look at the International business, the NEXT brand is still growing very strongly, 16%, delivering the lion's share of growth or more than half the growth overseas. If we just sort of break the overseas growth down into aggregators and NEXT Direct. What you can see here, I think, straight away is I think this is the first time for many years that we've reported the NEXT Direct business growing much faster than aggregator -- or not much faster, growing as fast as the aggregator business. And you can also see that aggregator business at 23% is lower than the 33% and 61% that we reported that we told you about for [indiscernible] in the second half. So you can see that our aggregator business is -- the growth there is beginning to moderate as we begin to annualize some of the gains that we made last year. Focusing on the NEXT Direct business and breaking that down into the sales that were driven by marketing and those that came naturally from underlying growth, 23% of our growth in the half came from marketing. And the surprising thing here is that having grown our spend by 63%, we haven't seen any erosion in the rates of return we're seeing on the advertising. I mean they've nudged forward, but we would have expected those to move back not to below the GBP 150 million, but we expect them to be below last year's number. And if there's one thing that is driving the exceptional growth of our Overseas business, it is the maintenance of these returns because we don't start with a fixed budget for marketing. We spend as much as we can as long as we're getting the returns. The things that we think are driving those returns are improving technology. This is not our technology, improved technology of the media partners we work with the Googles and Metas of this world, partly their better technology, better targeting of customers and partly us learning to use their technology more effectively. Secondly, and I can't understate the importance of this. All the work that we've done to improve the website functionality and delivery services serves to reinforce the marketing. If every customer who comes to our website through an advert has a higher probability of a sale because we've improved the functionality or the payment type or the way the basket works or the delivery service makes them more likely to come back, that marketing pound becomes more effective. So our websites and services have driven marketing growth. E-media costs have come down. We can't take the credit for this. This is all about the de minimis tax and GBP 3 levy in Europe, which has dissuaded some of the companies who import very cheap stock at very low prices under the -- were importing under the tax radar. It has dissuaded them from spending as much on advertising in Europe. And finally, there are lots of countries where we had virtually no advertising last year. So it was sort of virgin territory, and we were able to spend more money in those territories without eroding overall margins. And the final thing, and again, I will stress this much more when I present this to my colleagues back at [indiscernible], but cost control and maintaining the right margins by getting our pricing right is absolutely central to delivering the profitability required to drive the marketing. All that sounds fantastic, but that doesn't mean that 1% number looks a bit anemic. It's not as bad as it looks. And this is all about the war in the Middle East. First quarter, our underlying growth was down 8%. In the second quarter, it was up 11%. The 11% isn't a good number because of the pent-up demand. But if we look at the last 15 weeks of trade, our underlying trade, the trade that isn't driven by marketing was up around 8%. And that begs an interesting question about the nature of our overseas customers versus U.K. And what I'm going to talk about, first of all, on this is the customer spend. What this graph shows is for the U.K., the spend by tenure of customers. So customers who have been with us 3 years, on average spend GBP 203 a year. Those who have been with us for just a year, spend GBP 101. If we look at the shape of that maturity curve overseas, we were surprised to see that it was pretty much identical, a little bit lower, but it's pretty much identical. And that is not what we thought we'd see. We thought we would see far more occasional customers and customers not coming back or extending their portfolio of products overseas nearly as much as they have done in the U.K. And that number, in fact, is a little bit understated from the mathematical number because the Middle East takes so much per customer. So these numbers are basically for all of our international customers, excluding the Middle East. The Middle East customers, this slide just shows what the Middle East customers spend per customer, and it's much, much higher. So if I had included that, it would have given an artificial view of what our International customers are spending. But underlying international customers seem to behaving in a way that isn't dissimilar from our U.K. customers. We were, I said, surprised by that, by all of you because you're so much clever, won't be surprised because, of course, in the U.K., it doesn't mean that the NEXT brand is as attractive overseas as it is in the U.K. because in the U.K., we've got stores, and most of our online customers also spend in stores. So it doesn't mean that we're going to -- the NEXT brand will be -- is as powerful overseas because we don't have the store sales, but it is nonetheless very encouraging for the economics and development of our Online business. In terms of retention rates, again, we were pleasantly surprised by this. We thought these are the U.K. numbers. So if we have 100 customers in who we recruit this year, 35 of them will reorder next year. These numbers are sort of the normal numbers we'd see in an online world. If we look at the overseas numbers, again, remarkably similar. We thought they'd be a lot lower because of we think stores, both acting as collection and returns points and our credit offer would make the U.K. customers far more retentive than overseas customers. The balancing factor is the amount we spend on marketing. So in the U.K., we spend 3.8% on marketing. And in effect, overseas, the part of the 10% we spend on marketing isn't about just gaining new customers. It's about doing the hard work that the stores are doing in the U.K. to retain customers. And all of that filters through into the economics of the 2 businesses. So actually, I should say on that -- sorry, that the -- if you look at the difference, part of that is paid for by NEXT by making lower margins overseas online than we do in the U.K. and part of it is paid for by the customer by us making higher gross margins. If we look at the three areas of focus, starting with product. we've talked -- again, I've talked what many of you will consider to be and certainly my colleagues consider to be [ nauseam ] about newness, quality and choice driving our ranges forward. And I do think that we have made a step change over the last 2 or 3 years in terms of the newness in some of our ranges. What I want to talk about today is just some of the work that we've done on quality and choice because it could sound like this was just an active will that, yes, the board save more newness and everyone rushes off and gets more newness, but it's much, much harder than that in terms of giving customers real choice, because it involves a whole lot of work that if you don't do it, you won't get the choice and quality that you would if you put the hard legwork into inspiration. And I think, again, there's a bit of a myth here that people think, well, in an AI world, all you've got to do is ask ChatGPT what the latest trends are, and it will tell you. But of course, all ChatGPT can do and all our numbers can do is tell us what has been. It can't get those flashes of inspiration that our buyers and designers get when they go on inspiration trips to overseas capitals to new fabric fairs in Shanghai to new mills to new suppliers, exhibitions, art galleries, all of the things that give people that little spark of inspiration that actually you need a human being in our experience, you need a human being to drive. And you then need to spend the time designing it. And again, here, I think there is a trap because when AI first came along, people who -- I did actually sit down with one of our designers at NEXT been to St. Martins college, and she was saying what I've got to do is put these prompts into AI and it generates this wonderful graphic. And what -- that was exciting at first. What we found is those graphics didn't sell. It was pretty much universal. AI-generated graphics didn't sell. And the ones that -- it seems to be going the other way, the graphics and stripes and designs that really work, the color balances are the ones that are down with human hand, human eye and sort of have an emotional response. And we're putting more time into the work we put into painting, drawing, screen printing, wood block design, shape design. And finally, you can get the inspiration and the design, but you then got to do the development to really elevate particularly fabrics and washes and dyes. And that involves not going to the supplier and say, can you give us a fabric that looks roughly like that. It involves going to the mill often long before you've decided what garment is going to go into and develop fabrics with mills. We're not doing that universally. Others do it more than us, but it is something that the more we do it, the better fabrics we get. And it's not just about the base fabric. It's also about the wash techniques and dying and spinning and yarn manufacture that drive or materials that go into home. The more we can do further upstream of the development does two things. First of all, it means we get better quality. And secondly, it means we're exposed to some of the new trends in materials that ultimately drive the trends in fashion. All of that takes a huge amount of time. And so whilst technology and warehousing, we have reduced their cost as a percentage of sales, actually, in product, we've increased product costs as a percentage of sales, partly to do this and partly to seed new WOBL brands. And so our question, is this just a question of us throwing more money at it? Or can we be clever? And we think there is a -- we think there's a big pot of gold here basically. And that is the amount of time that we spend our product teams, this is not an admin team in product. These are the product people themselves. The amount of time they spend on admin is around 25% of their time. And I checked with one of our product because the sort of O&M people came back and said, "Oh, it's 25%. And I print one of our directors, is it really 25%? It can't be. And she actually said to me, no, no, it's more than that. And so the feeling is that it's actually taking more of their time. It's a bit like kids homework, things you really hate doing, things take a lot longer than the things you enjoy doing. But nonetheless, those admin tasks are because there is more and more data that only the designer, the buyer, the merchandiser can put into the system, whether it be information that the fabric mills need, the regulators need, the imports team, export team, the whole management of data for the websites. If you don't put in a feature of a garment, if the buyer doesn't tell you that this is a super soft touch, whatever it is, then when the customer searches for Spersoft, it won't appear in search results. So buyers have to do far more in terms of managing website attributes, admin and store planning, warehouse warehouse management, all of those things require an enormous amount of data and only the product people can do it. And our systems at NEXT, when we have new people come to NEXT, the product people say these are brilliant systems. Your systems do in an integrated way what other companies do with a whole load of spreadsheets that are sort of loosely held together. And that's great. But that integrated system is one that we haven't fundamentally changed since 2003. We just added the amount of data people put into it. And it works like a spreadsheet. It's incredibly arduous and difficult to put in the data and it's slow. So we're introducing three new product systems this year, Production Management System. This is -- production management. This is where we -- our merchandisers need to check that when the mill says they're going to produce the fabric on the 14th of September, they really have done that. And there's an interesting lesson for life here in that if you don't ring them and say, have you done it, it will run late. So you've got to do that. The squeaky door gets the oil and at every step of the production process, if we are not monitoring it and managing it step by step, things run late. So that's -- it's a very arduous task at the moment, the production management system, which to a degree allows our suppliers to link in and tell us when they've done things rather than us have to chase them or save a lot of time, got an online imagery and attribution system and the data entry system. And then the last one of those, of all the systems at NEXT, that is the one that I get the most [indiscernible] complaints about. You could argue that's because product people are naturally more [indiscernible] than those who are prepared to take in other parts of the business, but it is a really important thing. We think those systems altogether can save at least half the time that we're spending on admin. And that is time that we can spend on doing the things that the product people are employed to do, which is produce brilliant product. In terms of productivity, the other area we're getting good gains in productivity is our warehousing. We've talked about the GBP 500 million that we are planning to spend on Elmsall 3. We spent about half of that so far. And the good news is that is beginning to pay dividends in terms of costs. So we can see warehousing costs coming down as a percentage of sales. And I should stress that this includes the cost of the depreciation and rent of the new warehouses. So this is fully costed, still coming down as a percentage of sales. A lot of that growth, but not all of it is driven by people productivity. And what this shows is the pence per unit dispatched per customer, the amount we spend on wages divided by the amount that we send to customers in a given month. That varies month by month naturally. That's what it was in 2024. What you can see is that when we -- when we introduced new mechanization last year, we got a big dividend. And we thought that, that was it. The exciting thing is that partly as a result of fine-tuning the mechanization and getting it working better and partly as a result of a whole raft of things we've done to improve warehouse productivity. We're still seeing gains. And I think there's further to go on that versus last year. I think that will continue. And part of that GBP 7 million upgrade today is the gains that we've got from the sort of unexpected improvement in productivity in our warehouses. In terms of service level, there's a sort of good news, bad news story here. Good news is this is our what we call NDOTIF, not delivered in full and on time. And this is a potential parcels that are not delivered on time and in full. I should stress this looks like a very bad number. This is the 2024 numbers. It looks like a very bad number. It's not as bad as it looks because the vast majority of failures are parcels where we deliver 4 of the 5 items today, but one of them is late and it delivered tomorrow. So this is not -- we're not saying 10% of everything feels late to customers, it doesn't. 2025, when we -- as we got the new mechanization working and new capacity, you can see that improved, still off our target of below 6%. This year, we started really well. We were below our target of 6% for 2 months. So we thought we were on to our winner here. But as volumes begun to increase and the warehouse begun to run hot, -- you can see that still better than last year, but we have seen a significant uptick. And what that is all about is it's about the glitches in the system and the mechanization. When you're not that busy, you can rectify and correct in day. When you're running up to the wire at volume, you can't afford to have those glitches and errors and some of the mechanization failures and system errors that are corrected very quickly, and we would tolerate in quiet times, we cannot tolerate at busy times of year. So we have got -- I think we've got a big job of work to do to go back and say, actually, we kind of -- we need to move towards a sort of zero tolerance approach to glitches. And it's not enough just to say we fixed them -- we need to do much more root cause analysis. We need to do much more volume testing to actually generate those errors in advance of them happening in simulation so that we can correct them before they happen. And that's a big job of work for our systems team. It is a lower job. It won't be as burdensome as it was with the help of AI. And that is my slightly clumsy segue into the next section. Last time, we talked about AI across the board about how all the different departments were developing AI. I just want to focus on what is by far the most exciting part of AI development in the group, and that's how our technology team are developing AI. We started to introduce Assistive AI that I get told off for this, but to me, this looks a bit like a sort of a spell check on a predictive taxes for coders. So this helps you write code, and that's -- we introduced that in 2024, and it did affect both our headcount and our cost as a percentage of sales. So we think that was very positive. What we're looking at now is how we deploy Agentic AI. And where Assistive AI does it says it assists. Agentic AI really does the work. It actually does it. And when you see it, it is is astonishing. And only comparison I can think of is like the difference in a calculator and a spreadsheet. Obviously, you're all far, far too young to remember the first calculate. But I remember the first time age 7, I saw a calculator, and I was owed by it because I could work out whether my parents pivot the right amount of pocket money. I could adjust for inflation, all of those sorts of things. And I thought that was wonderful until the spreadsheet came along and then you realize the calculator is like a toy. And that's what Agentic AI is compared to Assistive AI. And just to sort of put that in context of what we're doing about it. This is our total development life cycle. And a lot of businesses will use the same development life cycle. It's called Agile, starts with ideas and concepts, specification, coding and deployment and support. And those are the stages you have to go through. In terms of the people required to do that, you have lots of different roles. And they don't all do one -- they overlap in terms of those elements of the total development life cycle that they look after. And what we're doing is we are developing agents to sit alongside all of those roles and to do those functions. And the people managing those functions will move from sort of doers to managing the agents doing the task. That is kind of the vision. There is a lot of work to do here. And you can't buy an agent out of the box and stick it up and say, well, you get on with it, do my business specification. You need to put in the time to give the business context, the business rules apply guardrails, very importantly, apply security. It's a bit of a thing at the moment, but security is a big part of designing agents. You need to train the agents. You need to design them in such a way as that they're LLM independent so that they consist -- it doesn't matter whether they're using as a sort of underlying imagine, whether they're using Claude or ChatGPT or another provider of AI because if you don't do that, you end up overpaying for the underlying AI. And you need to ensure that they are cost effective in terms of the way that they use processing power. So where we are up to with this is we've developed and piloted 3 of the agents, and we've [indiscernible] to deploy those 3 agents across some areas in the business. In terms of the total plan, we anticipate that we will have piloted all of these agents by February next year and be well on the way to deploying them by June, July next year. Now to me, this -- when you talk about systems project, normally, it's measured in years, not months. So I was surprised at the speed with which the systems team came to me and said they can do this. But so far, we have managed to pilot and deploy agents much faster than I thought would have been possible. And just to sort of give you a sense of the power of these agents, we -- on some of our websites and some of our apps, we don't have a share function. And we've always looked at it, and it's just been too expensive to develop and considered not worthwhile. And traditionally, in terms of development time, it would have taken 11 days to develop using traditional coding. With Assistive AI, we've have got a good gain, maybe between 10% and 20%, it might have taken 10 days. When we gave the specifications to a coding agent to write, it took them -- it took it 24 minutes and 29 seconds to do what one of our coders with assistive AI would have taken 10 days to deliver. That's sort of 10 working days. It took about half a day, obviously, to manage the agent and then get the coding right and to make sure that to weed out the errors, but still a very dramatic reduction in time and a huge increase in productivity. There's a caveat here because it's a bit like that sort of demonstrated that works brilliantly. But then when you look at the whole thing, we actually -- and this whole project ends when we only saved 17% of the time. And that was partly because we had to put a lot of effort into developing the agent as we went along, and some of that will be reusable and partly because these agents will only really work really effectively when you stitch them all together. And that makes it looks like it was a sort of a rugby line with the ball being passed down actually, because there are lots of functions where different agents have to talk to 3 other agents. So the work to stitch these all together means that the huge gains that we think are possible will take time to deliver, but we're targeting 30% improvement in productivity by February '28. Now you'll know that we spent GBP 200 million on software. So instantly, you'll go back to your spreadsheets and type in GBP 200 million, that GBP 60 million saving. Don't do that, partly because nearly half of the cost of what we spend on systems is infrastructure and software. And obviously, the cost of that with agent will go up. But also don't assume it's GBP 35 million saving on people because if we -- if that were to be the case, it would be a huge failure because the really exciting thing about this in the context of a company turning over GBP 6 billion, the GBP 35 million isn't what matters. What matters is the ability to deliver projects so much faster. There isn't a single project in the business, single new project that doesn't, in one way or another, involve a system change, and it's normally the rate determining step. So our hope is that the speed at which we can write software will accelerate the speed at which we can develop the whole business. I think the other point to make is that it's not just about speeding up projects we would have done. It's about doing projects that in the past just wouldn't have been conceivable. We have a mainframe that sits at the heart of all of our stock and price processing. It's incredibly effective machine doing huge amounts of data in a very short period of time. And -- but the coding for it, I said in the report hundreds of thousands. I checked with the hundreds and thousands of lines of code. But the person who knows about these things said, actually, it was millions, but I didn't want to put that because you think I was exaggerated. It's millions of lines worth of code that is -- have been built over time and desperately needs to be modernized in order to speed up the rate at which we can improve and modernize the rest of our software. When we last looked at this, the cost was GBP 50 million. We still would have done it, but it wasn't the cost that held us back. It was the fact that actually the business with the ground to halt, -- because so many other projects would need mainframe development time that we just couldn't afford the time. We now think and are planning to start modernizing our mainframe coding, and we think the cost will be GBP 10 million, and that we'll be able to do it in a modular way that allows the rest of the business to continue moving forward. That does beg the question like what are people going to do? And there's -- I think as well as producing far more volume and developing ideas we wouldn't have been able to do before. I do think the nature of systems work is going to change. Solving business problems, advising new applications, making the business aware of what this new technology can do, training and managing our agents [indiscernible] agents, building new agents, maintaining the security of the system, integrating with third parties and controlling costs are going to be a big job of work. So my hope is that we don't see a big reduction in people. I think over time, we will see less cost, but the real drive is to generate more productivity and more production. Just as a final note on costs, we spent GBP 200,000 last year on AI, GBP 1.2 million this year. We'll spend at least GBP 3 million next year. And unless we design our agents very, very rigorously to use costs carefully, we could end up spending more on AI than we did on the original people, not quite, but we could end up spending a lot of that productivity gain. So there is still a big job of work to do, albeit a very different type of job. And on that note, we're going to finish and go to questions. I think the summary is we haven't done what we often do is go through every area of the business and say what we're doing. We just -- we picked three areas where we think there are really exciting projects, product, warehousing systems. What I want to assure you is that pretty much every area of the business has exciting projects. The -- this is not the three things we're working on. These are three of the things we're working on. But what I hope they can do is give you a flavor of the sort of depth of thought and energy that is going into different parts of the business to move it forward. And on that almost motivational note, I'll throw to sorry, -- remember to use the microphones, everybody.

It's Anne Critchlow from Berenberg. I wonder if you could talk, please, about input costs, so polyester, cotton freight and any sort of price increases that you might expect to put through for spring/summer on the back of those? And then secondly, could you talk about your thoughts on whether it's more attractive at this moment to create new WOBL brands organically or perhaps look for acquisition opportunities?

Yes. So I think the answer to the first question is that we're not seeing nearly as much of the costs filter through to factory gate prices as we thought. So if I look at what our current -- the contracts we have placed so far for spring/summer, we haven't placed a massive amount, but the amounts we have placed for spring/summer, we're looking at like-for-like price increases around 1% to 2% -- so not much. And I think there are two different things going on here. There is an increase in input prices, but part of that has been papered by an increase in productivity in our supply base. And there all this technology, they're getting smarter as well. And I wouldn't also underestimate the impact of the reduction in duty on goods from India. We get a meaningful percentage of product from India. If all of that -- if -- and I'm not saying we do, but let's say we get 10% from India and duties come by 12%, that reduces our total prices by 1%. But it's not the effect on India that is the most dramatic thing. It's the effect it's had on all the markets competing with India. They have had to sharpen their pencil as a result of that. So 1% to 2% is the short answer. And in terms of WOBL, are we getting more productivity from buying in a brand or from developing a new one? There is no rule of thumb. Some of the brands that we bought in have been amazing and some have been average. And some of the brands we started from scratch have been phenomenal. I mean -- but some of them have failed. So I think what it comes down to ultimately in both the ones we buy in and the ones that we develop is the quality of the idea behind them rather than whether they are bought or devised.

It's Freddie Wild from Jefferies here. A couple, if I may. So first of all, possibly the slightly more boring obvious question on this U.K. slightly downgraded outlook for half 2. Is that anything you're seeing now? Or is this more just you looking ahead at the various input costs people are facing and the extent to which that could carry through to FY '28 as well, please? Second question is around the competitive landscape in Europe. Obviously, big changes to [indiscernible]. You flagged the impact on the media spend there. Are there any broader impacts you're seeing in terms of pricing, competitive behavior, stuff like that? And if I could squeeze one very final one in.

If we had time at the end, we will definitely come back. Otherwise, you get inflation. We've got to fight inflation in all things, especially analyst question. So in terms of your boring obvious question, are we seeing the numbers now? I think the boring -- giving you a boring obvious answer is that I have to tell you what our trade was. And that will be a trading statement, that would be new information that will be coming to the market, which I don't think is appropriate in an environment even as August as this one. I think we have definitely seen some softer weeks in the -- I'm not going to give you an average for the last 7 weeks. We have seen some softer weeks. I think there are signs. But I think what we said in the book is that this is more about our anticipation than what we've seen. And the next question, which is much more exciting is the competitive landscape in Europe, which is such a big subject that I wouldn't want to sort of get into the nitty-gritty of it. I think we're not seeing any significant change that we can discern. But I think the other thing is, to be honest, is we don't spend a huge amount of time looking at it because the most important thing that we look at is what response are we getting to the product that we are selling in the various countries that we sell it. And we're constantly experimenting with prices and services in order to get that balance right. So we're kind of trying to feel our way to what is the right level of price for us and how much we can spend on marketing rather than looking at market data and then trying to tell you what we do to that because actually, much better to be responsive and look at your own numbers and you are to try and impute what's going on and then respond to phantoms that may or may not exist.

Matt Clements from Barclays. First question on -- you mentioned the shifting spend from cost to serve to marketing. I just wonder what the market implications of that are? Is it going to drive an acceleration in market consolidation, do you think? Or is it net neutral? And the second one on international margins. You mentioned what having more pricing power. Any color on that relative to NEXT brand? That would be very helpful.

Yes. So I think just starting with the pricing power. I think, again, this is very much test and trial. It's that's not that we started with the assumption we can get more pricing power. It's that we put it on at a certain price. We've got very good traction. And it was only when we stopped and looking at actually at 24%, we are over profiting and leaving these businesses vulnerable to competition that we lowered it to 20%. So I suppose I think if you say to me -- put the question slightly different, do you think you've got room to increase your prices in Europe without affecting sales? My guess to that would be probably yes. But I wouldn't want to do that because we wouldn't want to risk it. We're making a very healthy profit. So -- and it would be quite a big risk to take. In terms of market consolidation, again, I think that's way above my pay grade. I think that is something that you and your big economics departments will do far better than I can guide you on, Richard?

Richard Chamberlain from RBC. Simon, I think you mentioned that the U.K. retail stores, you've seen some quite significant productivity improvements in the first half. I wonder if you can give a little bit more color on which ones had the most impact on margin? And then second, in the U.S., which you're now breaking out in terms of disclosure or giving a little bit more, I guess, given the dominance of Amazon in that market, particularly in areas like kidswear online and so on, should we assume that the U.S. sort of structurally will be a lower margin market on the sort of medium- to long-term view for NEXT, given Amazon sort of customer focus likely presumably higher take rate and so on. Yes, they are the two questions.

Yes. So U.S., I think it's a good question. I think it will have a lower profitability than our other businesses, mainly because at the moment, we're shipping stock from the U.K. to the U.S. and that is much more expensive than shipping it to Germany or even the Middle East. So I think that structurally, the question do we pay for that to be the customer, well, we're paying for most of it. So the margins are low, but still more than 10% so still comfortable margin in the U.S. In terms of productivity in stores, but just sort of coming back to sort of your question on pricing, we didn't start by looking at the market and saying what prices can we afford to charge. We started with the prices that gives us a respectable margin and then see if we can sell that. In terms of stores, there is no one measure that was introduced. It's not like we suddenly introduced self-serve [indiscernible], although we are trialing that at the moment, but we have not introduced any new technology. What we have done that is making a big difference is we have lots of different productivity measures in stores. So we can see how productive each member of staff is on, say, TIL transactions, the average time and TIL transaction, a store pickup for collection for a directory online customer, replenishment from Stockholm and shop floor. We can measure all of those things. And what we were doing is we were measuring -- in the past, we were measuring those in silos going, well, let's look at our delivery productivity and focus on improving the productivity of the people at West. What we're now doing is looking at it horizontally. So we're saying, for you, Richard, how do you do on all these different areas because that is a much more effective way of finding people who've got the opportunity to be more productive than it is by trying to do it. Your delivery manager doing deliveries and your TL manager doing TILs. So I think that's the biggest change.

Richard Trainor from Bernstein. First question on AI. You contrasted its inapplicability to the inspiration and design process in clothing with its applicability in the technology function. I was wondering if there are other areas of the business where you think AI will be very helpful? And secondly, I was wondering about the impact of peer-to-peer marketplaces and whether you see them having an impact on markets in the U.K. or abroad? And if so, in which types of product?

Peer-to-peer...

Yes, peer-to-peer reselling marketplaces. I'm thinking of [indiscernible].