In 2025 your North America business, meaning the United States and Canada, let go of 17% of its orders on purpose. It chased fewer customers who were each worth more. It made more money anyway: sales fell 13% and operating profit rose from €368.8m to €385.3m. That is a hard trade to execute and you executed it.
Two notes on the numbers, used throughout. Sales changes are quoted in constant currency, which strips out exchange-rate movement so the comparison reflects the business rather than the euro. Operating profit is your AEBITDA line: profit before interest, tax, depreciation and one-off costs.
The first three months of 2026 ran the same play and got a different answer. Sales fell 10%. Operating profit fell 45%, from €61.3m to €33.4m, which is more than four times faster than sales. Nothing in your reporting flags what changed, because the number that explains it is printed eleven lines below the number that hides it.
Start with the encouraging figure. In the first three months of 2026, every €1 of North America sales left 28.9 cents after paying for the food and for getting the box to the customer's door. A year earlier it left 23.0 cents. That is a gain of 5.9 percentage points, earned while sales were falling, which reads like a business getting sharper under pressure.
Now the same line measured a second way. Your statement also reports that figure with impairments removed. An impairment is an accounting write-down: the company judges that some assets, here idle production sites, will never earn back what they cost, and books the loss. No cash moves, but the loss lands in the profit figure.
On that second measure North America went from 30.3 cents to 29.3 cents. It fell.
This matters because it is not what happened over the full year. Across 2025 the reported margin rose 3.6 points, from 24.5 to 28.1 cents, and the write-down-adjusted margin rose 2.6 points, from 27.9 to 30.5. Roughly three quarters of that gain was real operating improvement. In the first quarter of 2026, none of it was.
Your efficiency programme is the reason. It targets €300m of annual savings and was about 80% implemented by the end of 2025, on your own account. That leaves roughly €60m still to land. The programme carried last year's result. The first quarter shows the underlying margin turning over while its final fifth is still being delivered.
You do not need a competitor to see what North America could do differently. You run a second business under the same brand, pursuing the same declared strategy of fewer and better customers, and in 2025 it produced a very different result.
Both succeeded at the same half of it. North America lifted average order value, meaning what a typical customer spends per delivery, by 4.8%. International lifted it by 4.7%. Practically identical, which says the pricing and add-on work lands equally well in both.
| Full year 2025 | North America | International | The gap |
|---|---|---|---|
| Sales (€m) | 4,207.1 | 2,553.7 | North America is 1.6 times larger |
| Sales change, constant currency | −13.0% | −1.4% | nine times the decline |
| Orders (millions) | 53.66 | 46.87 | |
| Order change | −17.0% | −6.2% | 2.7 times the decline |
| Order value change | +4.8% | +4.7% | the same |
| Average order value | €78.30 | €53.60 | North America is 46% higher |
| Kept from every €1 of sales | 28.1c | 23.0c | North America keeps more |
| Marketing per €100 of sales | €19.80 | €14.70 | North America spends 5.1 more |
| Marketing per order | €15.69 | €8.17 | 1.92 times as much |
| Operating profit (€m) | 385.3 | 209.3 |
That is what makes this a demand problem rather than a margin problem, and it is why the rest of this document spends very little time on cost.
A 17.0% fall in orders against a 4.8% rise in order value produces the 13.0% sales decline you reported for 2025. The two numbers reconcile precisely, which means every play below has to move one of them, or the cost of serving them.
| Forward play | What the record prices it at | Verdict |
|---|---|---|
| Close the marketing productivity gap against your own International segment | North America spends €19.80 per €100 of sales against International's €14.70 and loses orders 2.7 times faster. The 5.1 point gap is about €216m, which is 56% of North America's operating profit and 3.6 times what remains of the €300m efficiency programme. Both halves raised order value equally, so execution quality is not the variable. | Pursue |
| Outsource North American production and fulfilment | The one competitor on your own list that publishes a US segment did exactly this and keeps 42.9 cents per euro against your 28.1. €273.2m of write-downs in two years is the cost of not doing it. Held back by the counter-evidence: that competitor is in a covenant-bound restructuring with going-concern dependency. | Second |
| Shift marketing investment from North America toward International | International held sales to a 1.4% decline on 5.1 points less marketing intensity. But it is 1.6 times smaller and already near flat, so it cannot absorb enough spending to offset a miss in a segment that size. | Second |
| Finish the efficiency programme | About €60m of the €300m target remains. Worth completing, but it is a fifth of one programme against a €216m benchmark gap, and the adjusted margin fell in the first quarter while it was still being delivered. | Hold |
| Raise order value faster through premium add-ons | Needs +16.3% to hold sales flat at the first quarter's order decline, against the +4.8% you managed in your best year for it, partly by removing discounts, which works once. The requirement is 3.4 times the achievement. | Rejected |
| Raise menu prices | Food and preparation costs already rose 2.2 points of sales on deliberate quality investment. A price rise on a base losing 14% of its orders works directly against the retention that investment was meant to buy. | Rejected |
| Cut marketing spending further | Marketing discipline is your own stated cause of the order decline. North American marketing fell 16.6% in 2025 and orders fell 17.0%. Cutting further buys margin with the volume that is already the binding constraint. | Rejected |
| Restore marketing spending to 2024 levels | 2024 spending of €1,009.2m preceded a 12.3% group order decline the following year. Re-buying the low-return customers you deliberately exited reverses the one part of the strategy that demonstrably improved margin. | Rejected |
| Shift the North American mix from ready-to-eat back to meal kits | Meal kits run at a 9.0% operating margin against ready-to-eat's minus 5.9%, so the near-term arithmetic favours it. But ready-to-eat improved 2.5 points year on year and is the category you are building the US around. This trades the 2027 position for a 2026 number. | Rejected |
| Close or sell the smaller brands | Good Chop and The Pets Table carried €33.8m of first-quarter sales at a minus 17.1% operating margin, so roughly €5.8m of drag. Real, and immaterial next to a €216m gap. Not where a quarter should go. | Rejected |
| Expand retail and grocery distribution | Order value is disclosed excluding retail, which confirms the channel exists but leaves it unsized. No sales, margin or volume figure is published for it, so it cannot be priced from outside. | Untestable |
| Increase how often each customer orders | You disclose orders and order value by segment but not customer counts. Frequency cannot be separated from customer numbers in the public accounts, so no claim about it would be checkable. | Untestable |
At the first quarter's rate of order decline, 14.0%, order value would have to rise 16.3% to hold sales flat. You delivered 4.8% in your strongest year for it, partly by removing discounts, which is a lever that can only be pulled once. The requirement is 3.4 times the achievement.
Run it the other way and it is clearer still. A 4.8% gain in order value exactly offsets an order decline of 4.6%. North America is running at 14.0%. Order value cannot carry this segment at any plausible rate of improvement, and that removes the three plays most companies reach for first.
Holding order value growth where it landed last year, here is what each rate of order change produces for North America.
| If North America orders change by | Sales, constant currency | What that would take |
|---|---|---|
| −17.0% | −13.0% | 2025 repeated. No change in trajectory. |
| −14.0% | −9.9% | the first quarter, held for the year |
| −10.0% | −5.7% | the decline slows by roughly a third |
| −7.0% | −2.5% | North America performs as International did in 2025 |
| −4.6% | 0.0% | break-even, where order value exactly offsets volume |
| 0.0% | +4.8% | orders stabilise |
Group ready-to-eat sales were €465.9m in the first quarter, down 6.9% in constant currency, at an operating margin of minus 5.9%. That is an improvement on minus 8.4% a year earlier, but it is still lossmaking, and it is still shrinking. Meal kits ran at plus 9.0% over the same quarter.
So the category you are building the US business around is currently consuming profit rather than producing it. Any 2026 plan that depends on ready-to-eat returning to growth is assuming that recovery rather than forecasting it, and the public record gives no basis to do either.
International's lower marketing intensity partly reflects a decade of brand-building in Germany, the Benelux and the Nordics, where HelloFresh is close to being the category and organic demand is correspondingly cheap. Winning a customer in the United States is structurally more expensive, more auction-driven and more contested, by Home Chef inside Kroger, by Tovala and CookUnity in ready-to-eat, and by grocery delivery generally. A gap of this kind can persist without anyone doing anything wrong.
Ready-to-eat is concentrated in North America and is in acquisition mode, which loads that segment's marketing line with spending whose return arrives later. North American order value is 46% higher than International's, so a higher absolute cost per order is expected rather than surprising. Some share of 5.1 points is simply what a larger, newer, more contested order costs, and no amount of execution recovers that share.
Marley Spoon is the only company on your own competitor list that publishes a United States segment, and its US business posts better margins than yours: 42.9 cents kept from every euro against your 28.1, and a 12.3% operating margin against your 9.0%. It got there by outsourcing US production and fulfilment entirely.
Grant every objection and the gap does not reach zero. International raised order value 4.7% against North America's 4.8%, the same commercial motion executed equally well, and lost 6.2% of orders against 17.0%. Neither auction pricing nor mix explains that difference. What it reflects is what the spending buys, and that is the part worth attacking.
North America spends €19.80 of every €100 of sales on marketing. International spends €14.70 and defends its revenue nine times better. Each point of that gap is worth about €42m a year at North American sales.
| Lever | What is left in it | On what basis |
|---|---|---|
| Marketing productivity against International | about €216m at full closure, about €54m at a quarter of it | 5.1 points of €4.26bn of North America sales |
| The remaining efficiency programme | about €60m | the final 20% of the €300m target |
| Faster order value growth | not enough at any plausible rate | needs +16.3%, achieved +4.8% |
| Further marketing cuts | negative | the cuts are what produced the 17.0% order decline |
To be clear about what this is not: it is a claim about where the next point of margin is cheapest to find. It is not a claim that North American marketing is being run badly, and nothing in the public record would support that.
Take North American marketing spending and orders by channel and by product category, and compare against International on the same cut. The question is narrow: how much of the gap survives once ready-to-eat acquisition and order value mix are held constant? That residue, not €216m, is the real target. It is a two-week analysis on data you already hold, and it can show the recommendation is wrong.
Both versions are already published, so this costs nothing. Managing to the reported line in the first quarter would have shown a 5.9 point improvement in a quarter when the business gave up 1.1 points. The adjusted line is the one that tells you whether the efficiency programme is still working, and right now it is the one saying it has stopped.
Part 4 shows the whole 2026 outcome sitting between a 14% and a 7% order decline. Order value is already doing its share at 4.8% and cannot be pushed much further. Whatever the year's target is, it is an order count target, and it should be stated as one.
€273.2m of write-downs across 2024 and 2025 is the running cost of a production footprint sized for demand that has not returned. Marley Spoon is the reason not to move quickly. But knowing what third-party fulfilment would cost per order against your current owned cost is worth having before the next impairment review, whether or not anything changes.
We are Scalable OS. We work with public records: audited annual accounts, segment disclosures and quarterly statements. From those we reconstruct from them what is actually happening inside a business and the market around it. Then we send that to the business, unsolicited, before there is any relationship at all.
The reason is straightforward. The analysis in this document is the kind that normally arrives after a retainer, a discovery phase and a scoping call, which means most operating teams never see it at any point in their working lives. It is not expensive to produce, because the underlying records are public and free. It is that nobody has a reason to produce it for you until you are already a client. We would rather demonstrate the work than describe it.
There is a second reason, and it is the one that decided the shape of this document: nothing in it was requested. A search engine or an assistant answers the question you thought to ask. This report exists to raise the ones nobody inside your business has had a reason to ask, because the two numbers that answer it sit eleven lines apart on the same page, and nobody is paid to read them together.
That is also its limit, and the limits are specific. Four ordinary internal reports would change the analysis:
Send us The figures below. Each one is an input to a number this report could not compute from your filed accounts: contribution margin by brand and by channel for the last eight quarters, excluding all purchase-accounting entries; the acquisition model's original revenue and margin forecast against actual, by year; the cash operating expense of the acquired brand separated from amortisation of acquired intangibles; inventory on hand and sell-through rate for the acquired brand, by month; monthly paid marketing spend and newly acquired customers, by channel, for 24 months; first-order contribution margin after discounts, shipping, payment fees and returns. and we will send back this same review rebuilt on your actual numbers: