Cut the spend and the whole machine shrank within the year.
2024 to 2025, company-reported figures.
Official. HelloFresh FY 2025 results. Note 2. · gabrielgruter.com
This is the first of a series: I take a brand I have never worked on, plot it on the triangle from public signals only, name where it over-invests, and say what I would do about it. Everything below is public record, read from the outside. Where a judgement is mine, I say so.
The shape of the machine
HelloFresh buys its demand a unit at a time: discounts recruit, churn removes, the machine buys again. Marketing runs at roughly a fifth of revenue, which at that scale is a cost of goods.
HelloFresh built one of the great acquisition engines. Introductory discounts bring a customer in, the box arrives weekly, and when the customer churns, the machine goes out and buys another one. In 2024 the group turned over €7.7 billion, and in the third quarter marketing has been running at roughly a fifth of revenue: 18.7% in 2024, 20.0% a year later.
When a customer leaves, the machine buys another.
The brand is buying its demand a unit at a time, at market price.
A fifth of revenue is not a marketing budget. At that scale it is a cost of goods.
The tell
In 2025 HelloFresh cut more than €200 million of marketing spend. Active customers went from about 7.3 million to about 6.9 million, orders from roughly 123 million to 115 million, and revenue fell 9% in constant currency to €6.76 billion.
A brand with equity can pause its marketing and coast on memory for a while, because people keep turning up out of habit and preference. HelloFresh reduced the buying and shrank inside the same year. Marketing fell by more than €200 million, yet marketing as a share of revenue went up, from 18.7% to 20.0% in the third quarter. Revenue fell faster than the cut.
Marketing fell €200 million and still grew as a share of revenue.
Revenue fell faster than the cut.
Official. HelloFresh Q3 reporting. Note 2. · gabrielgruter.com
The company calls the shrinkage deliberate: a pivot toward fewer, longer-tenure customers, and the margin numbers back the profitability half of that story. Fair enough. But note what the pivot concedes. A large slice of the revenue was demand the discounts made. The brand did not own it. They said the customers were discount-chasing. Calling the revenue manufactured is my read.
Plotted from the outside
The point sits deep in the Activation corner. Identity is partly funded, the green box is distinctive; Equity, the reason to pay full price, is what goes unfunded.
Nearly everything HelloFresh spends sits in activation.
The brand’s distinctive look and sound are not the problem. The green box is recognisable, and the brand has bought its share of jingles and sponsorships. What the public numbers say is missing is the reason to pay for the box at full price, remembered by people who are not currently holding a voucher. When a fifth of revenue goes to buying the next order, that memory is what goes unfunded.
What I would do
Keep the machine. Pick the full-price reason lapsed customers miss, fund it as an asset, step the introductory discounts down while the memory builds, and measure the rebuild in share of search.
Keep the machine. It is a good machine, and nobody should turn off an engine that efficient out of principle. The work is on the other two jobs, and it starts with the question the vouchers have been papering over: what is this brand for, at full price? Convenience, learning to cook, waste, the weekly ritual. Pick the one the lapsed customers miss, fund it like an asset rather than a campaign, and hold the distinctive green to it. My bet is the ritual, the box that settles the what’s-for-dinner question every week. And step the introductory discounts down while the memory builds.
- Keep the machine Nobody should turn off an engine that efficient out of principle.
- Pick the full-price reason Convenience, learning to cook, waste, or the ritual. The bet here: the ritual, the box that settles the what's-for-dinner question.
- Step the discounts down While the memory builds.
- Measure it in search Share of search against the category, and the branded share of the query mix. If memory returns, it shows there first.
A brand that only ever recruits at half price teaches people what it costs.
Then measure the rebuild the way I would measure any equity work: share of search against the category, and the branded share of the query mix. If the memory is coming back, it shows there first.
Where this argument is weakest
Part of the decline belongs to the category, the churn may be structural, and the margin half of the pivot is working. The growth model was still activation-shaped, and the 2025 numbers show it.
A teardown from the outside owes you the caveats. Meal kits are a trial-prone category, and every player in it took the same post-Covid normalisation. Some of this pain belongs to the category, and the clean test, HelloFresh against the category’s own decline, is not public in a comparable form. The churn may be partly structural, because a weekly cooking commitment is a hard product, and no amount of equity fixes a product problem. And the pivot is young: margins improved in 2025, and if tenure lengthens the way management intends, the smaller base could prove the right trade. The growth model was still activation-shaped, and the 2025 numbers show it.