Insights - Category strategy

Private label took half the units. Not on price.

Private label now sells half the units in Europe's six largest grocery markets, and it does it on less than half the promotional support brands use to defend themselves. The discount is not the defence. It is the delivery mechanism.

The verdict first, because it inverts the standard brief. Private label did not take half of Europe's grocery units by being cheaper. It took them while discounting at less than half the rate brands do, which means the brand is losing a price war that the other side is not fighting. Every promotional cycle a brand runs to hold the line re-stages the purchase as a price comparison, and a price comparison is the one contest where the retailer's own label has a structural cost advantage it did nothing to earn. The promotion is not failing to work. It is working perfectly, for the other side.

Private label is the market now, not the alternative

Circana reported in April 2026 that private label had crossed 50 percent of unit sales across Europe's six biggest grocery markets for the first time, tracking millions of SKUs across more than 230 FMCG categories. Value share was 42 percent, worth 324 billion euro, and unit share has risen every year since 2021, by more than three percentage points cumulatively.

The country spread matters more than the headline. On units: Spain 59 percent, the Netherlands 56, the UK 52, Germany 52, France 46 and Italy 36. On value the same six run 52, 55, 44, 44, 36 and 31. Poland sits well below all of them, and both of the panels that measure it disagree about by how much.

NielsenIQ, in the data PLMA publishes for 17 European markets, puts Polish private label at 25.1 percent of value and 28.6 percent of units in 2025, up from 24.6 and 28.0 in 2023. YouGov's Shopping Monitor 2026 puts the value share at 23.5 percent, 3.8 percentage points higher than five years earlier, on spending of 67 billion zloty and growth of 6.6 percent. Both numbers are right for what they measure, and the useful read is the shape rather than the level: steady annual gains, decelerating from roughly 20 percent a year in 2022 and 2023 to about 7 percent now, in a market where discounters took 44 percent of retail FMCG spending.

Deceleration is where the comfortable story usually gets written. It is the wrong read. A share that grows 7 percent a year from a quarter of the market compounds faster in absolute points than 20 percent growth did from a sixth of it, and the PLMA report names Poland explicitly as a market with future growth potential because its shares are still relatively low. Spain is what low looks like after the runway is used.

The asymmetry nobody prices: 34 against 14

One number should have ended the price debate, and did not. In the same six markets, Circana found that about 34 percent of branded unit sales moved on promotion, against 14 percent of private label unit sales.

Sit with the direction of that for a moment. The side that is winning on price promotes 14 percent of its volume. The side that is losing promotes 34 percent. If price were the mechanism of the shift, the ratio would run the other way, because the attacker would be buying the trial it needs and the defender would be holding its line. Instead the defender is doing almost all of the discounting, and losing unit share every year since 2021 while doing it.

The reason is not complicated once you stop treating promotion as a price event and start treating it as a framing event. A promotion does two things at once. It moves a few units this week, and it tells the shopper what the decision is about. Run it on 34 percent of your volume and you have told them, roughly every third purchase, that the relevant question in this category is what things cost. That question has a structurally correct answer and it is not your brand, because the retailer's label carries no advertising, no innovation pipeline, no listing fees and no sales force, and the retailer decides whose price sits on the shelf edge.

Private label's 14 percent is not restraint. It is the correct play for the side that already owns the price position. You do not pay to argue a case the other party is arguing for you.

The brand is losing a price war that the other side is not fighting. Thirty-four percent against fourteen is not a defence. It is a subsidy paid to the question you cannot answer.

What the defence costs, counted

Put a number on it, because the number is the argument. Take a brand whose volume would be worth 100 million zloty at full shelf price. Apply the two figures above as the working assumption: 34 percent of units sell on promotion, at an average depth of 20 percent off.

Revenue realised is 66 million at full price plus 34 million at 80 percent, which is 93.2 million. The brand gave away 6.8 million zloty, or 6.8 percent of full-price sales value. Because the units sold either way, that 6.8 million does not come out of a budget line. It comes out of contribution, in full.

Now set it against the thing that does have a budget line. Working media at 7 percent of net sales on that base is about 6.5 million zloty. So the brand runs two budgets of almost exactly the same size. One of them is briefed, pre-tested, reviewed by three people and defended in a planning meeting. The other is approved in a trade negotiation one customer at a time, never aggregated, and reported as a deduction rather than an investment. We set out the mechanics of how that deduction disappears into net revenue realisation in the discount nobody costed; what matters here is that the brand's largest marketing instrument is the one it does not call marketing.

The obvious objection is that promotion buys volume, so the give-away is not a loss. Test it with the only arithmetic that settles it. A 20 percent discount has to be repaid in volume out of a thinner contribution per unit, and the multiple required is the contribution margin divided by the margin that survives the cut.

Contribution margin at full price	Volume needed on promoted units to stand still
50 percent	1.7x, or 67 percent more
40 percent	2.0x, or 100 percent more
35 percent	2.3x, or 133 percent more
30 percent	3.0x, or 200 percent more
25 percent	5.0x, or 400 percent more
20 percent	Never. The discount is the entire margin

Break-even volume multiple for a 20 percent discount, derived as the contribution margin divided by the margin remaining after the cut. Substitute your own margin and depth.

At a 35 percent contribution margin a 20 percent promotion has to more than double the volume of every unit it touches just to leave contribution where it was. Some mechanics clear that bar. A third of annual volume does not, and the part that does not is a transfer to shoppers who were going to buy anyway. The formula for that part is simple enough to run in a meeting: 6.8 million zloty multiplied by the share of promoted units that would have sold at full price. At half, the brand paid 3.4 million zloty for nothing. The share is measurable and almost nobody measures it, which is the whole reason the budget survives.

Poland: there is no price gap left to defend

This is where the Polish numbers stop being reassuring. The PLMA report gives the underlying sales, not just the shares: in 2025 Polish private label took 14.9 billion euro on 11.7 billion units, and brands took 44.6 billion euro on 29.2 billion units.

Divide. The average private label unit sold for 1.27 euro and the average branded unit for 1.53. Private label is priced at 83 percent of the brand, so the average structural discount in the Polish basket is about 17 percent, not the half-price gap the category still argues about. Run the same division on Circana's six-market shares and private label there sits at roughly 72 percent of the branded unit price. Polish private label is relatively more expensive against brands than Western European private label is, which is the opposite of the usual assumption.

Then put the promotion back in. A branded unit at 20 percent off sells for 1.22 euro. The average private label unit sells for 1.27. On these averages the promoted brand is already the cheaper option, and it has been losing share every year regardless.

One honest caveat, because this number will get quoted. These are basket averages across all categories, not like-for-like SKU comparisons, and the mixes differ: private label is over-represented in cheap, high-volume categories, which drags its average unit price down. That cuts against the precision of the 83 percent, though arguably in the direction that makes the real like-for-like gap narrower still. What survives the caveat is the conclusion, and it is not a small one. The question on the table in most Polish commercial reviews is whether to close a price gap. The arithmetic says the gap is already closed, the brand closed it with promotion, and closing it did not work. Continuing is not a strategy; it is the absence of one. We laid out the conditions under which a cut is nonetheless the right move in price wars: when not to cut price, and private label meets almost none of them.

Your competitor sets your shelf price

There is a structural reason this fight is unlike any other competitive fight, and it has nothing to do with price. In Porter's terms the retailer is a buyer with concentrated power. In practice it is also a competitor, and the two roles share one set of books.

No other rival gets to decide your shelf price, your facings, your promotional calendar and your listing, while reading your sell-in data and watching which of your lines actually turn. A brand competing with private label is negotiating with its own competitor for access to the only place the competition happens, and the competitor earns a better margin on the alternative it is offering in the same bay. The retailer's loyalty scheme completes it: the shopper data sits on the retailer's side, and so does the decision about whose offer the shopper sees.

That asymmetry is permanent and there is no clever answer to it. It does, however, settle which levers are real. Anything that depends on the retailer's goodwill, shelf position or promotional support is rented. Anything that lives in the shopper's memory before they reach the aisle is owned. The brands that hold up under private label pressure are the ones that moved their investment from the rented half to the owned half before they had to.

Where private label cannot go, and what that proves

The Polish category data is the most useful evidence in this whole argument, because it draws the line in the only place that matters.

Per YouGov, private label holds under 10 percent of spending in cola-type drinks, margarines, deodorants and razors. In price-sensitive categories such as toilet paper, paper towels, groats, rice, frozen fruit and ready meals, it runs from 50 to over 70 percent.

Look at what those two lists actually have in common, because it is not price. Cola is cheap, heavily promoted, trivially copyable sugar water sold next to a retailer version at a visible discount, and private label still cannot get past 10 percent of it. Groats are cheap, lightly promoted and also trivially copyable, and private label owns them. The difference is not cost, quality or discount depth. It is whether the category purchase is attached to an occasion and a brand that answers it, or is a replenishment decision made on price per kilo.

Which gives a diagnostic worth more than any share report. For each of your categories, ask what the shopper is solving for. If the answer is a stock level, private label will take it, and the only sane responses are cost leadership or exit. If the answer is an occasion, a ritual, a social risk or an identity, private label stalls well below its category average, and it stalls there even when it is substantially cheaper. A cheaper retailer version has sat on the shelf beside the cola for years, at a visible discount, and it has bought under 10 percent of the category. That is the clearest proof available that the brand premium in those categories is not paying for price, and that defending it with price is a category error, in both senses of the phrase. This is the same mechanism as the penetration argument we set out in penetration, not loyalty: what grows a brand is being the answer that comes to mind at more occasions, and what private label buys is the occasion where nothing comes to mind at all.

Its method is copying, so stop handing over category codes

Private label does not do positioning. Its method is mimicry: it reads what the category's leading brand has made familiar, reproduces the recognisable parts at a lower cost base, and places them at the shelf edge beside the original. That makes its capability precise, and the precision is exploitable.

It can copy anything that has become a category code. The colour that now means decaf, the pack silhouette that means premium, the typography that means natural, the claim that every third pack carries. None of that is yours in any functional sense, because recognition without ownership is a public good you paid to build. It cannot copy an asset that is both famous and uniquely attributed to you, because the uniqueness is the part that makes the copy read as a copy rather than as the category. The two numbers that govern this are the ones we set out in distinctive beats different: how many people recognise the asset, and how many of them name you and only you.

So the practical defence is uncomfortable for most brand teams, because it is not a campaign. It is an audit that usually finds that three quarters of the brand's visual equity is category code, that the genuinely unique assets are the ones the last two agencies wanted to modernise, and that the pack architecture has been drifting toward the category mean for years in the name of looking contemporary. Every step toward the category mean is a step toward being copyable at a lower cost base by the company that controls the shelf.

How we count it

Three of our lenses carry this, and they answer different questions in a fixed order.

Shopperology goes first, because the decision about where to fight has to precede the decision about how. It builds the occasion inventory per category and scores what the shopper is solving for at each one, which separates the categories where a brand can hold from the categories where the honest answer is cost leadership or exit. Brand X-Ray then scores the assets on fame and uniqueness and marks each one as category code or genuinely owned, which tells you what private label can take from you for free and what it cannot take at all. Pricemore closes it on the money: promotional frequency against depth, the break-even multiple above run on your real margin, the share of promoted volume that was never incremental, and the reference price repair path, which is almost always fewer and shallower events rather than a list price move. The pack and price ladder that comes out of it is the architecture we described in price-pack architecture.

Five things to do, in order. First, aggregate a year of promotional spend into one number and put it next to working media on one page, because almost no brand has ever seen those two figures side by side and the comparison does more work than any deck. Second, measure the non-incremental share of promoted volume on the mechanics you repeat most, and multiply it by the give-away to get the figure you are paying for nothing. Third, cut promotional frequency before depth, because frequency is what teaches the reference price and depth is what clears the stock. Fourth, run Brand X-Ray on your own pack and label every element as code or owned, then stop modernising the owned ones. Fifth, re-cut your category list by what the shopper is solving for and move money out of the replenishment categories into the occasion ones, which is the only reallocation on this list that changes where you compete rather than how well.

And one thing not to do. Do not answer a 50 percent unit share with a value range of your own at a lower price. A third tier under your own brand hands the retailer a cheaper comparison inside your own portfolio, funds it out of your contribution, and makes the price axis the explicit basis of the category. If the shopper needs a cheaper option, the retailer already sells one, and it is better at selling it than you will ever be.