HelloFresh SE (HFG) Earnings Call Transcript

Good morning, ladies and gentlemen, and welcome to the HelloFresh H1 2026 Results Call. The conference will be recorded. [Operator Instructions] Let me now turn the floor over to your host, Dominik Richter.

Good morning, ladies and gentlemen. Thank you for joining our Q2 2026 earnings call. I'm joined today by our CFO, Fabien Simon, and I'm going to spend my time on the strategic picture, why we're making the choices we're making and why we remain convinced they're right for this business over the long run. Our CFO will then walk you through the numbers in detail, and we'll open the floor for questions after that. We've also published our H1 2026 shareholder letter this morning alongside this presentation. It covers a lot of what I'm about to say in more depth, and I encourage you to read it alongside today's remarks. I want to start where we always start because it's the frame for everything else that we're talking about today. Our mission is to change the way people eat forever. Our vision is to become the world's leading digital native CPG company. Every decision I'll describe today, including the ones that show up as pressure in this quarter's numbers is in service of that longer-term vision, not in spite of it. What you're looking at right now is a phase to rebuild, not a single quarter story. Phase 1 was all about fixing our cost base. Phase 2, the one we're in right now, is reinvesting those savings into a genuinely better product. Phase 3 will be turning that better product back into growth on our terms once the numbers tells us it's the right thing to lean in. We're firmly in Phase 2, better products, better customer metrics and a top line that's still catching up to both. I want to cover 3 things today. First, why our structural position in this category gives us the room to be patient rather than reactive. Second, what our efficiency program has actually funded on the product side in H1. And third, what the resulting customer data tells us about whether any of this is working. Let me start with our structural moats because it's easy to lose sight of them during the quarter-to-quarter reporting cadence. HelloFresh has built real advantages against both CPG manufacturers and food retailers in the overall food ecosystem. We run a constant feedback loop with our customers, millions of ratings, swaps, reorders and cancellations every week, which lets us get ahead of what people will want before they can articulate it themselves. That feedback loop sits inside a deeply rooted culture of data and experimentation, which derisks every product bet before we scale it company-wide. Our customer base skews towards the top 40% of the income distribution, which gives us pricing power and margin resilience. We occupy the right corner of the food market. Convenience, health, personalization and sustainability are the megatrends that everyone is seeing and traditional food companies struggle to meet. We've built a well-oiled D2C marketing flywheel that plans around where the shopping experience is heading, not where it's been. And our purpose-built vertically integrated operations from our own fulfillment centers, our own last-mile logistics to our own procurement allows us to push innovation and iterate at a rate that is impossible for CPG or food retailers. Individually, none of these show up in a month or half-year P&L. Collectively, that's the reason we can make deliberate patient decisions instead of reactive ones, and thereby, we have strong conviction in the strategy we pursue. The patience is exactly what our efficiency program has brought us. We're now roughly 85% through our initially communicated EUR 300 million program as of the end of H1, with the remaining approximately 15% planned for completion in H2. These are permanent structural changes to our cost base, not one-off savings we will need to refine next year, and they are the direct funding source for the entire product agenda I'm about to walk you through. While H1 has seen a number of adverse external factors such as heavy winter storms, spiking fuel prices, we have nonetheless seen good underlying progress in the factors we control. Meal kits has already seen a convergence to our target low to mid-teen adjusted EBITDA margins over the course of the last 18 months. Ready-to-eat is where we are focusing a lot this year on the bottom line to also bring it to full-year adjusted EBITDA profitability. We remain on track for that, and margins are already up a full point year-over-year in H1, even after absorbing material ramp-up costs of our new Factor Europe facility and the disruption of the aforementioned winter storms, increasingly driven by a larger share of revenues coming from a more tenured customer base. Getting the rest of the way there in H2 comes down to 3 things for RTE: improving our contribution margins again after the initial ramp-up phase; second, continuing to calibrate our marketing spend to where returns are strongest; and thirdly, scaling our emerging RTE geographies into profitable volume growth. We're doing all of that in that sequence for a reason. Margin discipline first, upgrading the customer experience second and only then leaning back into growth. We are now about 12 months into what we call the Refresh, our stage-gated product investment program. We test in a handful of markets, measure rigorously against preset metrics and only roll out globally once we see the evidence. Early results out of the Nordics and the U.S. gave us the confidence to keep widening it in this half. The logic behind it is simple. With a limited number of meals and a narrow choice of cuisines, there are only so many long-term customers a business like ours can reach. So every dollar we put into broadening that selection is a dollar spent growing our addressable market, not just serving the customers we already have. Let me give you some very concrete examples. In factor, we've broadened our collection of GLP-1 friendly recipes for customers managing their nutrition around that medication, an area where demand has been real and where almost nobody else in the market has a credible answer yet. In meal kits, we've introduced low-lift dinners for the nights when a customer wants HelloFresh quality without the full cook time with precut vegetables or premarinated proteins allowing for a significant reduction of cooking time. We've also expanded protein variety, introducing new SKUs such as barramundi, Chilean sea bass or deer to our menus regularly. All 3 add more SKUs, more sourcing relationships and more variability in our kitchens. All 3 cost a bit more before we scale the volume behind them. We have accepted that ramp-up cost deliberately because it's temporary and the retention benefit behind it isn't. Each of these is a small change on its own. Together, they're what more plate -- more value per plate looks like in practice, and they're exactly the kind of investments that widens the pool of customers this business can serve over time. We are pairing the investment in food with digital features that make the product smarter around the edges, which is also how we think about becoming a genuinely digital native CPG company rather than just a meal kit company with an app attached to it. Our upgraded personalization engine uses onboarding quizzes and each customer's ordering history to build a food profile that recommends the best fitting meals from a now much larger menu. We've broadly introduced other product personalization features as well, such as a wide range of swaps in the weekly menu, which now lets customers change an ingredient, for example, tofu for chicken with one tap chef-tested, so the swap never cost them on taste. And we've built out personalized menu merchandising capabilities, for example, options to include the fussy eaters of the family, kids into the weekly dinner routine with an exciting collection of meals built around that use case. We've also launched our HelloFresh CookBook this year, which lets customers save any recipe they spot on social media or anywhere else online into a HelloFresh style recipe card with full step-by-step instructions, free, no subscription required. Customers have already saved more than 3 million recipes since launch, which tells us there's a real appetite for us to be useful to customers well beyond the meals we ship them. It also gives us a live read on trends and preferences. With these features, we increasingly embed ourselves into customers' lives and make our app useful for active, lapsed and prospective customers. So where does all of this actually show up? In meal kits specifically, constant currency average order value is up 5.7% year-over-year in H1, and order rate is up 4.1%. Customers who are with us today are ordering more often and spending more per order than they were a year ago. What you won't see yet is that translating into top line growth. And I want to be direct about why, because I know it's going to be a question on this call. We pulled back marketing spend significantly again in H1 on top of an already large step down a year earlier. As a deeply data-driven business, we hold every dollar of marketing to a strict return threshold. And that threshold only works when we can forecast our cost base with great confidence. In H1, we had less of that confidence than usual. Given the heaviest winter storms in 75 years, Europe had its own run of severe weather, and the war in Iran pushed fuel surcharges and fertilizer prices up across our supply chain. Spending aggressively into that kind of cost uncertainty means buying customers at a price we can't properly underwrite. And unlike in past times, we've chosen not to do that, even though it costs us in this half's top line. That's a deliberate trade-off to build sustainably for the long term. Where we don't yet have the long-run data to justify, we're staying patient rather than assuming it will work out. We're going to reassess our investment levels as we now move into back-to-school season when we typically get a much clearer read on demand elasticity and our cost base for the second half of the year. That discipline is the whole point of a phased approach, where right now, we're focusing predominantly on making the product significantly better for consumers. That brings me to how we're thinking about the rest of the year. Our focus for H2 remains on efficiency with select opportunities for investment where the data actually supports it. Specifically, we're focused on 4 things. We're recalibrating product investment levels based on what H1 actually told us works rather than scaling every initiative uniformly and globally. We're keeping marketing spend disciplined, and we'll let early results from back-to-school, one of the most important demand signals in the year, set the pace of any incremental investments from there. And after a period of initial inefficiency while driving a lot of innovation in H1, we're now targeting contribution margin expansion for the group again because that margin is what keeps funding everything else on the list. And finally, we're scaling factor outside the U.S. with our EU center ready for growth and into new channels. For example, after a successful trial in H1, we're broadening our access to retail shelves for our RTE product, which gives us a second channel and diversifies our revenue line. I want to close the strategic part of this call with one of the charts I find most convincing because it speaks to revenue quality rather than revenue quantity, and quality is what this whole strategy and Phase 3 build has been optimizing for. Look at meal kits first on the top of the page. Revenue from customers who've been with us more than 4 years now make up 34% of H1 net revenue, up from just 7% in H1 2023. That tenured revenue has kept growing, which tells you this growth is coming from customers staying longer and spending more once they're in, not from throwing more money at acquisition. Ready-to-eat on the bottom chart is earlier in that curve. Tenured revenue there is growing significantly, too. Revenue from customers that have been with us for over 2 years stand at 20% today, having grown considerably over the past 3 years, but the category is younger. So revenue is still a lot more sensitive to the volume of new customers we acquire in any given period. That's simply where RTE sits in its maturity relative to meal kits and tells you about the story of revenue decline in RTE in H1. To recap before I hand over, H1 has seen us continue to execute the multiyear strategy we've been deliberately pursuing. Strong innovation, tight ROI thresholds in an uncertain environment and good progress towards a return to positive full-year RTE bottom line. These trends are compounding in the right direction. It's a multiyear trend, not a quarter blip, and we're not managing this business quarter-to-quarter. The product is improving strongly and the customer reception proves it every single week. Growth, when we lean back into it, will be a choice we make with conviction backed by data we trust, not a target we chase into a cost environment we can't see clearly yet. With that, let me hand it over to our CFO, who will take you through the quarter's financials.

Thank you, Dominik. Good morning, everyone. Thank you for joining us for our Q2 earnings presentation. Before I walk through the quarter, I want to flag a change in how we are presenting our results. At the full-year results and through the various interactions with investors, I made a commitment to enhance the transparency of our disclosure and to update our reporting framework. Beginning this quarter, we are staying true to our words, and we are disclosing full P&L in reported and constant currency by product category, meal kit, ready-to-eat and others separately as a deliberate step towards more comprehensive reporting because each category has distinct economics and a distinct trajectory. We believe investors are better served by seeing them clearly and separately rather than through a consolidated view. So all the details can be found in our H1 report, including 2025 comparison. Additionally, what you would have seen in Dominik's sections and we will continue to see in the coming pages is also further transparency on cohort revenues. Again, this is something that we said we would do, as we believe this will give the market a better view of a more solid foundation and dynamic that we have in our business than sometimes understood. And this come on top of a greater focus on cash flow and the updated metrics of free cash flow plus lease repayment that I commented already during the year-end results. With that, let me now start by walking you through the key themes of the quarter before we look at each line item in detail. In Q2 2026, we delivered what we said we would deliver, as we continue to execute against our strategic priorities. Revenue of EUR 1.5 billion reflects a decline of 7.8% in constant currency for the group. Both the meal kit and ready-to-eat segment came broadly in line with our internal expectation, which I had communicated to the market last quarter. The other segment, that is a new venture in specialty meat and in pet food is up 36% in constant currency for the quarter. I want to draw your attention to 2 dynamics that I believe tell you more about the underlying quality of the business. The first is our tenured customer cohort. Those customers who have built HelloFresh in their weekly routine, continue to show stable, and in fact, improving behaviors in Q2 with order rates increasing year-on-year. This is true for the group and individually for each of the product categories. The second dynamic is average order value, which increased 6.5% in constant currency to EUR 71, driven by higher add-on contribution, a greater share of premium recipes and some price increases. On the operational side, 85% of our planned EUR 300 million efficiency measures were implemented before the end of June. Progress on the cost side was, however, largely offset by the deliberate front-loaded product investment that was communicated 12 to 18 months ago. This is reflected in the group contribution margin of 25.2%, down 2.1 percentage points year-on-year, and I will address that in detail on the contribution margin slide. But note that Dominik already mentioned that efficiency remains our priority for the rest of the year, and we will be disciplined with marketing and product investment. Q2 group adjusted EBITDA was EUR 100.6 million (sic) [ EUR 120.6 million ] with meal kit at 15.1% adjusted EBITDA margin, broadly the same as last year in percentage terms. Ready-to-eat margin was 3% in Q2 and better in percentage terms in H1 versus H1 last year. We continue to work on our goal of turning ready-to-eat adjusted EBITDA margin positive this year. And in fact, for H1, the U.S. market, ready-to-eat there was almost reaching breakeven at the adjusted EBITDA level, and we expect positive margin for H2. Finally, it is worth sharing that in July, we successfully completed the pricing and issuance of our inaugural EUR 350 million bond with a coupon of 5.5% maturing in 2031, which strengthened our capital structure and extends our maturity profile. The adjusted EBITDA outlook is reconfirmed for the full year. For our constant currency revenue growth, we see it trending towards the bottom end of the full-year range, and we will gain further visibility following the back-to-school campaign. Moving to the slide on revenue. So group revenue of EUR 1.5 billion in Q2 declined 7.8% in constant currency. And as I mentioned, this was in line with our expectations. Meal kit declined 8.9% in constant currency. Sequentially, this compares to a decline of 8.5% in Q1, including the effect of Spain and Italy, which has now ceased to operate, so practically flat sequentially. The performance is increasingly supported by 2 structural pillars. The resilience of our tenured subscriber base whose order rates are growing year-on-year and a continued improvement in net average order value, which reached EUR 64.5 in Q2, up 7.1% in constant currency for meal kit. What we need to see in order to witness an inflection in meal kit revenue growth is a step-up in conversions at our tighter ROI levels and in our ability to keep those conversions. We did not see that in Q1, an important acquisition quarter, which is having an effect on the rest of the year. The next natural point in time to assess our conversion momentum will be the back-to-school campaign, which starts very shortly. On ready-to-eat, the category declined 8.4% in constant currency in Q2, which represents a sequential deceleration compared to Q1. Ready-to-eat has a greater structural reliance on new customer conversion than meal kit. It is a younger category with a less deep-tenured subscriber base. Our disciplined application of marketing ROI threshold, therefore, has a proportionately larger near-term impact on ready-to-eat top line than the meal kit. The structural positive within the segment is that tenured ready-to-eat customers continue to expand their net revenue double digits year-on-year in Q2, which speaks to the quality momentum of the retained base. Q2 AOV was up 1.1%. Given how important conversions are for ready-to-eat and the fact that we communicated already that we entered the year with a low level of active customer base, we should not expect ready-to-eat to recover in Q3 and are instead looking at the level of exit rate for the year. I would also highlight our other segment, which includes our new ventures, growing at 36.1% in constant currency year-on-year. It's a continuation of the strong momentum that we have described in prior quarters. On Slide 13, I want to put the meal kit revenue trajectory in the context of H1 as a whole. So while Q2 showed a modest sequential decelerations, negligible on a like-for-like basis when you account for the discontinuation of Italy and Spain, the H1 pictures continued to support the direction of travel. The rate of decline has been narrowing consistently, and importantly, that improvement is happening as we tighten our marketing ROI standards. Turning to Slide 14. This is another very clear example of us providing the transparency around cohorts that you have been asking for. In this slide, I present in more detail the behavior of our customers across the different tenure pools by comparing the revenue growth evolution in constant currency from H1 2025 with H1 2026. So let me walk you through the chart from left to right. On the very left, you see the growth rate of the revenue in H1 generated by customers that were acquired within the preceding 12 months. The growth rate for this cohort remains negative in H1 2026 year-on-year, which is expected. Performance marketing, the primary driver of new customer acquisition also declined. But what matters is the rate of improvement. The year-on-year decline in new customer revenue growth in H1 2026 is about 1/3 of what we recorded in H1 2025. The trajectory is therefore improving. A similar dynamic, and in fact, a more pronounced one is visible for reactivation. The revenue growth rate from reactivated customers has improved meaningfully with approximately half of the decline recorded in H1 2025 having been erased by H1 2026. And this is consistent with the early signals we are seeing from lapsed customers responding to the improved product proposition. Finally, tenured customers. The revenue growth rate in H1 2026 is similar to H1 2025, which is, I want to note, a meaningful result in itself. Every month, we are bringing in fewer new customers who have the potential to become tenured over time, and yet, this cohort continues to grow at a stable rate. These customers are showing increasing order rate across the market and lower churn in the market where our product investment cycle is most advanced. This is the customer base we are building for. And it is a customer base that give us structural confidence in the trajectory of the business, even as the overall top line continues to reflect the managed transition toward a more loyal, higher-quality subscriber composition. Q2 2026 contribution margin came in at 25.2%, a decrease of 2.1 percentage points year-on-year when you exclude the impact of share-based compensation and impairments. And I want to address this directly because it has a lot of moving parts. The decline in contribution margin is not a reversal of our underlying efficiency progress. It is a deliberate consequence of the front-loaded product investment strategy we communicated at the start of the year. We said we would invest in H1 in the product quality, in the menu, in the experience ahead of the back-to-school period, and that is precisely what the margin reflects. Let me give you now the segmental breakdown. Meal kit contribution margin was 28.2%, a decrease of only 0.2 percentage points year-on-year, which I think is a meaningful signal in itself. The efficiency reset is largely holding the meal kit margin in place even as we absorb product investment on that line as well as volume-led operational deleverage. Ready-to-eat contribution margin was 21.6%, a decrease of 5.2 percentage points year-on-year. The larger impact in Ready-to-eat reflects both the product investment, which we are also being rolled out in the ready-to-eat category, and the mix effect from expansions into new international markets, which are not operating yet at the efficiency level of our core geographies. On the cost line breakdown, procurement and cooking costs increased by 2.4 percentage points year-on-year. In meal kit, the figure is 1.8 points, which will moderate in H2 as our procurement and fulfillment teams get better at navigating the initial phase of inefficiency that you typically see when launching a meaningful product innovation cycle, for example, around operational waste. We are on track to meet our previously mentioned target of 150 basis points. In Ready-to-eat, the cost increased by 3.6 percentage points, half of which is product investment, which will also moderate towards 150 basis point target in H2 and half of which is a ramp-up of our European kitchen, which in ready-to-eat shows up in this line. Fulfillment costs decreased by 0.3 percentage points with efficiency program savings partially offsetting the impact of the volume-led operational deleverage and fuel cost inflations. So the direction of travel and the underlying cost structure is intact. The back-to-school campaign will give us the first evidence on whether the consumer response to the improved product justifies the level of investment. Turning to the marketing spending. Q2 2023 (sic) [ 2026 ] group marketing declined 16.3% year-on-year in absolute terms, now representing 14.9% of group revenue, a reduction of 1.4 percentage points compared to Q2 2025. Marketing declined faster than revenue in both segments, which is the intended dynamic. The continued reduction in marketing is the ongoing execution of the structurally differentiated approach to customer acquisition, which is anchored on stringent return on investment threshold rather than volume. In meal kit, marketing spend reached 11.6% of revenue in Q2 2026, down EUR 23.4 million in absolute terms year-on-year. But, in H1 2026, marketing to revenue for meal kit was broadly stable at 14.8% compared to H1 2025, which reflects consistent and disciplined application of our ROI criteria and some incremental marketing spend in countries where we see product cycle is more advanced, such as in the Nordics. In ready-to-eat, marketing spend reached 19.2% of revenue in Q2 2026, down 23% year-on-year. The higher relative level of ready-to-eat marketing compared to meal kit reflects the fact that it is a younger category with a greater reliance on new customer conversion. At the same time, the year-on-year decline in ready-to-eat marketing in percentage terms was actually steeper than in meal kit, which is a direct consequence of the disciplined focus on higher ROI customers and channels that we have applied with equal rigor to both segments as well as a reduction in the very high brand marketing spend incurred in H1 2025. I would flag that marketing efficiency during the back-to-school period will be an important input, both to our assessment of the conversion trajectory and to the marketing spend profile we will carry into the remainder of the year. Turning to Slide 17 on the adjusted EBITDA summary for Q2 and for H1. Q2 2026 group adjusted EBITDA was EUR 120.6 million, representing a margin of 7.8%. So let me walk through the segments. Meal kit delivered EUR 167 million in Q2 at an adjusted EBITDA margin of 15.1%. This compared to EUR 184.8 million and 15.2% in Q2 2025, a margin that is broadly stable year-on-year despite the product investment being made in the category. The resilience of the margin is a direct function of the structural improvement in the meal kit cost base even as there is some volume-led operational deleverage. For H1 2026, meal kit was EUR 271 million adjusted EBITDA at 11.7% margin compared to EUR 334.6 million and 12.8% in H1 2025. That H1 comparison reflects the front-loaded investment profile, which is weighted to the first half of the year and the winter storm impact we witnessed in Q1 of this year. Ready-to-eat delivered EUR 13.1 million in adjusted EBITDA in Q2 at a 3% margin compared to EUR 17 million and 3.5% in Q2 2025. For H1, Ready-to-eat adjusted EBITDA improved to negative EUR 13.6 million from negative EUR 26.4 million in H1 2025. As mentioned earlier, the U.S. ready-to-eat business was not too far off from breaking even in H1 2026. The year-on-year H1 improvement is meaningful, and it is directionally consistent with our goal of modest full year profitability in this segment. Holding costs at EUR 53.8 million in Q2 were higher than Q2 of last year, reflecting primarily some IT service cost inflation, including those related to AI and some salary adjustments that had to be made as the share-based compensation program is reduced, but not, of course, on a one-for-one basis. On cash flow, H1 2026 free cash flow was positive at EUR 49.4 million. This represents a decrease of EUR 107 million compared to H1 of last year. But H1 of last year included a one-off tax refund, which I communicated at the beginning of the year. Excluding that nonrecurring item, the year-on-year free cash flow decline narrows to about EUR 69 million. So the remaining decline is principally driven by the adjusted EBITDA development, which, as I have described, reflects the front-loaded product investment and winter storm impact in this half. CapEx in H1 2026 was EUR 76.7 million, up from EUR 66 million in H1 2025. This step-up has a specific composition. So investment in Factor EU infrastructure as we scale the European ready-to-eat business and CapEx directly related to automation in support of our menu expansion and product investment program. But it is important to note that at the midpoint of our adjusted EBITDA guidance range for 2026, we expect free cash flow to remain positive for the full year, with H2 expected to be materially more favorable than H2 2025. That expected improvement is supported by a stronger H2 2022 -- 2026, sorry, adjusted EBITDA profile by normalizing lease liabilities relative to the one-off termination payment made in 2025 and by a working capital profile that is expected to be more favorable than last year. So before we go into the last slide on our capital structure and then open up the call to your questions, let me bring together our outlook. As mentioned at the beginning, the adjusted EBITDA outlook is reconfirmed in full. This means that at present and with the visibility we have, the range remains within reach. On constant currency revenue, our view is trending towards the bottom end of the full range, and we will gain further visibility following the back-to-school campaign. So I want to be clear about why these 2 statements can coexist. The strength of the efficiency program and the improving unit economics of our customer base give us the confidence to reaffirm adjusted EBITDA even as the revenue picture trend to the lower end of the range. And again, the back-to-school campaign is the next meaningful read on conversion and marketing efficiency, and we will share those findings with you as soon as we have them. On Slide 19, I want to address the bond transactions we completed in July and specifically the rationale behind it. So in July, HelloFresh SE priced its inaugural EUR 350 million bond with a coupon of 5.5% and a maturity in 2031. The notes carry a BB+ rating from S&P, which is consistent with our issuer rating and stable outlook. This transaction is primarily about extending our debt maturity profile and diversifying our sources of funding. The proceeds are being applied to general corporate purposes, including the refinancing of existing term loan debt, which reduces near-term maturity concentrations and gives the business a more durable and balanced financial foundations. And I want to be equally clear that this transaction does not change our capital allocation policy, which prioritize investment in organic growth opportunities and maintaining our adequate level of leverage as communicated before. With that, I open it up for questions.

[Operator Instructions] So let's start with the first question. This one is from Andrew Ross from Barclays.

I guess, inevitably, it's to come back on meal kit declines in Q2 and when you're thinking into H2. And the first thing I want to understand is exactly why the meal kit declines didn't improve in Q2 versus Q1. You obviously spoke about that on the Q1 call, it played out. But back at Q1, it sounded as though that you had a specific market where maybe customer acquisition had been softer in April. Now to me, it sounds like the tone is that it's a broader-based factor across markets with kind of lower ROI you're getting on your marketing spend. So I'd love to understand kind of what you've seen in terms of the cohort behavior in more detail in Q2 in meal kits. And I guess that feeds into a follow-up, when we kind of think about the implied improvement in H2 for meal kit to get to the low end of your full-year guidance range. What gives you confidence that you can do that? Do you have kind of any visibility into that back-to-school period that the ROI versus CAC math could have improved? Anything you can give to give us some comfort that growth is going to get better would be helpful.

Thanks for your question. In general, I think it's important just to kind of like think back to the phased rebuilds that we're going through, right? So really, our focus for H1 was making sure that we genuinely improve the product and broaden the TAM that we're going after before we then kind of like try to lean back into growth. In Q2, I think overall, like very much according to our expectations, very much according to the expectations that we communicated in Q1. So I think this was actually nothing like a super special that happened between what we communicated on the Q1 call and now in Q2, very much in line with our expectations. And I think on the product side, we've also like really done a lot of what we wanted to bring to customers and generally like have seen strong reception, which is reflected in the customer metrics. I think the more interesting question is the second one, the implied improvements. Obviously, on the cost side, it's very much that we're going to drive contribution margin expansion again. We talked about some of the known inefficiencies that we took on board during the first half of the year, which is generally normal when you try to do a lot of new things. I think with the lessons learned and applying those, we'll see some contribution margin expansion into the second half of the year. Efficiency program, more things will come through in the second half of the year. And then on top line, it's really a function of, number one, the improving order rates that we've seen among -- and AOVs among our customer base, which obviously also play out for all the customers that we've acquired in H1, which compared to 1 year earlier, will deliver more revenue per customer in the second half of the year. And secondly, then it's a function of how much we're leaning into growth and marketing in the second half of the year, which is, on the one hand side, something that obviously we control, but where we also are very much looking at the overall cost environment. It's definitely been a situation in the first half of the year where it was not fully clear to us how input costs would develop over the next 6 months. And that generally then, I think, as a rational operator, leads to applying higher ROI thresholds to your future investments. So it's really those 2 things.

The question is from Joseph Barnet-Lamb from UBS.

I was going to build off the sort of second part of Andrew's question and your response, Dominik. I mean, at the midpoint of your adjusted EBITDA guide, you need to do, I think, plus 24% year-on-year for adjusted EBITDA in H2, having just done minus 33% in 1H. Now, I think you sort of alluded to some of this. You said efficiency savings -- you have done more efficiency savings in H2. And also, you said expanding contribution margin. Can you quantify either of them to help us? Because it just seems like an absolutely enormous gulf that you're expecting to deliver in H2 year-on-year versus what you did in H1. So if you can quantify that or flesh out a bit more, that would be really helpful.

This is Fabien. I will try to help you on that one. So indeed, there is a meaningful improvement that we are expecting in H2, but first, I want to remind that the decline we are seeing in H1 is -- I mean, close to half of it is coming from a one-off event we have had with winter storm. So if you exclude that, the step-up looks less dramatic overall. And indeed, for us to deliver on our adjusted EBITDA ranges, we need to have, I would say, at a minimum EUR 24 million adjusted EBITDA improvement year-on-year in H2, which is like EUR 10 million to EUR 12 million minimum per quarter. And why we are confident about it? If we look at what we expect for Q3, I always like to give how we see the coming quarter coming. On top line, we can't because it's going to rely a lot on the back-to-school season and the conversion. But on bottom line, we have a pretty good visibility. We see this, I would say, EUR 10 million to EUR 20 million improvement coming already in Q3. And we believe that in Q4, there is really the possibility with a good campaign coming to achieve the same level. So we feel confident that both quality of conversions, continued discipline on marketing ROI, which will continue, but as well a lower impact of product investment, which we started that in H2 of last year, and we continued in H1 of this year is going to have a lower year-on-year effect than what we have seen for the first part of the year.

And the next question is from Giles Thorne from Jefferies.

If you get into the back-to-school window and see a similar cost environment that you saw in Q2 and you decide to keep the marketing ROI thresholds at the same level as Q2, where do you think you'll land on the full-year overall group revenue decline? Is it the bottom end of the range? Or is the bottom end of the range requiring a change in the marketing threshold, so you would actually land below the bottom end of the range?

I think it's a bit of a difficult question because I think ROI on marketing is not something which is linear throughout the year. It's very different month on month and by season. And of course, the mix always change. What we would say is maybe to bring some perspective on what we have done in the past. If we look at the progress we have been making last year between our H2 and our H1 and assuming we will deliver the similar improvement between H2 and H1 of this year, we would be at the bottom end of the range. But what we are keen to commit is as soon as we have visibility on the back-to-school, which may be well before our Q3 earnings, we will communicate to the market what we have been seeing and if there is a need for readjustment or not. But so far, we are planning for a success for the back-to-school campaign.