Zalando and Delivery Hero: What European Platform Economics Really Look Like

Platform economics work when the marginal transaction is nearly free and fail when it involves moving a physical object across a city. Comparing these two companies isolates the variable that actually determines whether a consumer platform becomes profitable. This case study belongs to the software pillar of the Germany Company Stories hub.

What actually changed when Zalando moved to a marketplace?

The balance sheet. A retailer buys inventory, holds it, bears the risk that it does not sell, and marks it down at the end of the season. A marketplace lists a partner's inventory, takes a commission on sale, and holds no stock.

The effect on capital employed is dramatic. Fashion retail carries enormous working capital in seasonal inventory, and markdown risk is the single largest determinant of profitability in the sector.

The effect on assortment is equally important. A marketplace can offer far more products than a retailer can finance, which improves conversion because customers find what they want, and improves partner economics because brands control their own pricing and presentation.

The trade is control and margin per unit. Commission on a partner sale is lower than gross margin on an owned sale, so the model only improves profitability if the volume and capital effects outweigh the margin difference, which for a business at this scale they do.

Why does food delivery struggle to become profitable?

Because the dominant cost is a person travelling to a restaurant and then to a customer, and that cost falls only modestly with scale. Higher order density shortens trips and allows batching, which helps, and it does not approach the near-zero marginal cost that software platforms enjoy.

The second problem is that all three sides of the market have weak loyalty. Customers compare prices across applications, restaurants list on several platforms simultaneously, and couriers work for whoever pays more that hour. Multi-homing on every side prevents the winner-takes-most dynamic that justifies loss-making growth.

The third is regulatory. Courier employment classification has moved toward employment status in several jurisdictions, which raises the cost base structurally rather than cyclically.

The economics do work in specific configurations: dense urban markets with high order values, own-brand grocery formats with better margins, and advertising revenue from restaurants competing for placement. Those are the components that eventually produce profit, and they are a smaller business than the growth narrative implied.

What did geographic expansion actually cost?

Capital and focus. Expanding across dozens of markets requires local operations, local regulatory compliance, local competition and, frequently, acquisitions at prices set by competitive bidding.

The theory was that a portfolio of markets diversifies risk and creates a global platform with shared technology. In practice, delivery operations are almost entirely local: the technology transfers, the unit economics do not, and each market requires its own path to density.

The capital markets financed this while growth was the primary metric. When investor attention shifted to profitability, companies with many subscale markets faced a problem that could only be solved by exiting them, which crystallises losses and reduces revenue.

The resulting portfolio pruning across the sector was substantial and correct. The strategic point is that market entry is easy to finance and expensive to reverse, so entry decisions should be underwritten against the cost of exit rather than the cost of entry.

How do these businesses actually make money now?

Through advertising, subscriptions and services rather than through the core transaction. That is the consistent pattern across consumer platforms once growth funding ends.

For a fashion marketplace, the profitable layers are partner services, logistics fulfilment sold to brands, and advertising placement within search results. Each is high margin and each depends on the traffic the core marketplace generates.

For a delivery business, the equivalent layers are restaurant advertising, subscription programmes that raise order frequency, and own-format grocery with retail margins rather than delivery commissions.

The strategic implication is that the core transaction is a customer acquisition mechanism rather than a profit centre, which is a very different business to run and requires very different management capability than growth-stage expansion.

Why did European platforms not reach American scale?

Fragmentation and capital. Europe is a set of national markets with different languages, payment habits, regulations, labour law and logistics infrastructure, so a pan-European platform is closer to twenty national businesses than to one continental one.

Capital availability compounds it. Sustained loss-making growth requires investors willing to fund a decade of losses, and European growth capital has been thinner and more impatient, which pushes companies toward earlier profitability at smaller scale, a pattern examined in the startup ecosystem pillar.

There is a defensible argument that earlier discipline is better. Several European platforms reached profitability at moderate scale while American competitors were still consuming capital, and the survivors are durable businesses even if they are not global champions.

The unresolved issue is that platform markets tend toward concentration, and a regional player in a globally concentrating market eventually competes against someone with lower costs and deeper capital, which limits the terminal position.

What should an operator take from the comparison?

That business model design determines outcome more than execution quality does. Both companies were competently run; one had a model in which scale improved economics and the other did not.

The practical diagnostic is to decompose contribution margin per transaction into components and identify which of them fall with volume. Technology and customer acquisition typically do; physical logistics and human time typically do not.

The second point concerns pivots. Moving from retail to marketplace was a genuine model change executed while the business was already large, which is rare and difficult, and it required accepting lower revenue recognition per unit in exchange for better economics.

The third is that the capital market's preferred metric shapes behaviour. A management team should decide which metric actually reflects value creation and report it consistently, because the alternative is optimising for whichever number analysts happen to be using this cycle.

How does fulfilment become a competitive advantage?

Through speed and reliability that partners cannot match individually. A marketplace that operates its own warehouses and logistics can offer next-day delivery and free returns across the whole assortment, including third-party inventory, which no individual brand could provide alone.

That converts a cost centre into a service sold to partners. Brands pay for fulfilment because the platform executes it more cheaply than they can, and the platform earns margin on the service while improving the customer experience that drives its own traffic.

Returns are the specific economic problem in fashion. Return rates in some categories exceed half of items shipped, and the cost of processing, inspecting and restocking a returned garment is substantial and largely fixed per item.

The operators who solve returns economically do it through better size and fit prediction, which reduces the return before it happens. That is a data problem, and it is where the platform's scale genuinely creates advantage over any individual brand.

What does the advertising business actually look like?

A search auction. Brands bid for placement in product search results and category pages, competing for attention from a customer already intending to buy, which is the most valuable advertising inventory that exists.

The margins approach those of pure digital advertising because the platform incurs no additional cost when a sponsored result is shown rather than an organic one. Every euro of advertising revenue falls almost entirely to gross profit.

The constraint is trust. Excessive sponsored placement degrades result quality, customers find worse products, and long-term traffic suffers. Platforms that monetise search aggressively see the damage in retention metrics years after the revenue appears.

How should a brand decide whether to sell through a marketplace?

By separating reach from relationship. A marketplace supplies traffic and conversion that a brand's own site cannot match, and it stands between the brand and the customer data that builds long-term value.

The practical approach most brands adopt is a split: marketplace for reach and customer acquisition, own channel for full-price sales, loyalty and data. That requires pricing and assortment discipline so the two channels do not cannibalise each other.

The risk to monitor is commission inflation. Platforms that begin with attractive terms typically raise take rates as their traffic advantage grows, and a brand that has allowed its own channel to atrophy has no response.

What happens when growth capital stops?

The business model is revealed. Companies whose losses were funding genuine investment in fulfilment, technology or density reach profitability within a few years of cutting growth spending. Companies whose losses were subsidising the transaction itself do not.

The practical test during the funded period is contribution margin after variable costs at the order level in the most mature market. If a company's oldest, densest market is still negative at that level, scale is not the answer and additional funding delays rather than solves the problem.

That distinction became visible across European consumer platforms when capital tightened, and it sorted the sector more decisively than any strategic announcement.

How do subscription programmes change the economics?

By converting a price-sensitive occasional buyer into a habitual one. A customer paying an annual fee for free delivery orders considerably more frequently, which raises density, improves routing efficiency and reduces the customer acquisition cost per order.

The risk is adverse selection. Subscriptions attract the heaviest users, who are precisely the customers on whom the delivery benefit costs most, so the programme can lose money on its best customers while the intended cross-subsidy from light users never materialises.

The designs that work price the subscription against realistic heavy-user behaviour rather than average behaviour, and add benefits with low marginal cost such as exclusive assortment or early access.

Frequently Asked Questions