Every single data point verifiable at a public URL.
The result is this: in a saturated market, up to three-quarters of offers do not move. In online pharmacy it is 75%. In other sectors the number varies, but the structure does not: the majority of offers are never repriced, never respond to competition, never generate dynamics. They are not mispositioned. They are structurally invisible. No repricing algorithm, no campaign, no channel switch moves them. The market structure has placed them there, and there they stay.
This is not a model. It is a measurement.
Who has the data
I enrolled at university in 1995, when the Web arrived in Italy. Since 2005 I have always worked in e-commerce. I watched Amazon arrive, I helped Yoox when it was still in Bologna, followed emerging brands going online, managed the digital transformation of companies that now turn over hundreds of millions. In every one of these projects I saw the data. And in the data I saw something that none of my clients, none of my tool providers, none of the pricing consultants could ever explain to me: theoretical combinations are not all of the real combinations. There are configurations the market never produces. Not by chance. By geometry. By constraint.
To measure what I was seeing, in 2015 I founded Competitoor, which became Skipper: a price intelligence company. Not a repricing tool. An observation infrastructure. Complete weekly crawls of every merchant's public offering in a sector, every product identified by code (MINSAN, EAN), every price dated, every URL verifiable. Not samples. Not estimates. The entire observable market, week by week, for years.
Skipper closed in 2024. Not for lack of data or results, but for lack of investors. In a market where people prefer to spend on "analysing consumer psychology" rather than looking at their own numbers, a company that showed you the bare structure of your market found no one to fund it. The machine stopped. But the data remain. And what those data revealed is not erased by shutting down the company that collected them. If this sounds like the story of someone who lost everything and now speaks from the ruins, you are right. But the view from the ruins is better: there is nothing left to sell, only something to show.
Stop for a moment on this point, because it is the thing nobody tells you: who can analyse 2.5 million offers across 926 categories in 6 different markets? Who has the data? An academic works with samples. A consultant works with client data. A repricing tool sees the merchant's catalogue, not the market. To see the entire market you need to build the machine that observes it. I built it. I ran it for nearly ten years. The data I present here do not come from a download or an API. They come from an infrastructure I built, operated, and paid for out of my own pocket until the market decided it did not need it. Who else has this dataset? Amazon. But Amazon does not publish it.
The discrete bands
With those data I was able to measure something I could previously only sense: repricing frequency is not continuous. It does not shade from one sector to another. It jumps.
| Sector | Weekly repricing |
|---|---|
| Wine | 5-12% |
| Luxury fashion | 8-15% |
| Pharmacy and beauty | 15-22% |
| Electronics | 25-35% |
There are no stable intermediate values. These are discrete bands, not a gradient. The sector determines the band. The algorithm runs inside the band.
People who work in e-commerce know these things by intuition. But intuition does not tell you where the wall is. It does not tell you that the reset threshold is at 13% new catalogue. It does not tell you that the same structure replicates between Italy and the UK to the eighth of a percentage point. Intuition says the wall exists. Measurement tells you where it stands.
Where all of this comes from
I spent the last two years understanding why. The answer was not in e-commerce. It was in the structure of constraints. I wrote a mathematical framework, validated it across 22 independent domains (from number theory to music, from chemistry to linguistics, from artificial intelligence to inhabited space), and only at the end did I apply it to the domain that started everything. I kept e-commerce for last, not first, because it is the domain where everyone says everything and nothing has ever been validated with rigour. I wanted to arrive with a structure already tested on systems where the numbers do not lie.
A market needs only three elements
Remove everything superfluous (brand equity, elasticity, seasonality, channel, predictive models) and a market needs exactly three things to exist: a difference (the offer relative to the market's structural response), a limit (the available margin), and a boundary where the two interact.
Remove the difference: everything is equal, no dynamics. Remove the limit: prices diverge to infinity, no structure. Remove the boundary: no transaction possible. Three elements. Not four, not ten, not the 47 parameters of your pricing dashboard.
I verified this reduction across all six markets. The structure holds every time. What changes is the speed at which margin is consumed by competition.
In smartphones, margin compresses rapidly: no brand protects the price from the repricing war. In perfumes, the brand absorbs the competitive pressure: margin rises from €9.50 to €10.80 over five months. Black Friday in perfumes produces a one-week dip (from €76 to €73), reabsorbed immediately. Margin does not always fall. But margin is always the limit.
And the competitive trajectory that brought it there is irreversible: you can change the price, but you cannot erase the history of competitive actions that defined your position.
Dual saturation
In a category with 27,000 offers and 2,400 brands, how many products can a consumer actually evaluate? Research on human attention says roughly 50 to 100. The rest does not exist within their classificatory capacity. It is not that the consumer does not see your product because it is low on the page. It is that their attentional system, under finite capacity, cannot classify more than a certain number of alternatives.
Two converging saturations: on the supply side, the market has already positioned the majority of products. On the demand side, the consumer cannot classify any more. The 75% is invisible from both sides.
The visibility paradox
At this point someone will say: so only visibility matters. I invest in ads, SEO, promotions, and I push my product into those 50-100 that the consumer manages to see.
No. It is exactly the opposite.
If you sell supplements and spend to make your product visible among 27,000 competing offers, you are paying to give visibility to the product, not to yourself. The consumer sees the promotion, searches for the product on Google or Amazon, and buys it from whoever sells it at the best price. Often that is Amazon. You paid for the advertising. Amazon collected the sale.
The problem is structural: there is no boundary between your investment in visibility and your conversion. The money you spend to promote a product that anyone can sell is an investment in the product, not in you. And the product is not yours. It belongs to the entire market.
The consequence for multi-brand retailers is devastating. If you are an e-commerce selling 2,400 brands, your competitive advantage is the cart: the ability to offer assortment. But you are not promoting the cart. You are promoting the individual product. And the individual product has no boundary with you. Anyone sells it. Advantage gone.
And the alternatives? First on Google Shopping? That means being the cheapest. Below cost. The boundary vanishes by definition. "But I acquire customers." Sure. Customers who buy below cost. Try raising the price and see how many come back. The quality of the customer acquired below cost is proportional to the margin you sacrificed to get them: zero.
First on Google organic? You capture all the clicks from people who are "evaluating": pure infocommerce. People who search, compare, and then buy elsewhere. Instead of letting them navigate organically, you pay for them. Brilliant.
Amazon itself nearly went under more than once. It makes no margin. Ever. And not through incompetence: because it understood that no boundary exists even for the largest retailer in the world. Amazon's margin does not come from e-commerce. It comes from services (AWS, Prime, advertising). The marketplace is the machine that captures traffic generated by merchants' advertising investments. The sellers make the revenue. The platform makes the profit. Structure, not strategy.
And the brands? Brands do not talk about product. Everyone knows this. They talk about representation, values, stories, community. Because they have understood (by instinct, not by structure) that the only boundary that holds is the one between the consumer and the brand identity, not between the consumer and the product. When the brand is you, the visibility you buy comes back to you. For everyone else, every euro spent on visibility in a saturated market is a euro donated to whoever holds the best position.
The market structure transforms the advertising spend of the small into the revenue of the large. Not through malice. Through geometry.
Three implications for online sellers
The first: aggressive repricing in a saturated market does not shift the invisible 75%. It makes them invisible faster. If your product is in the structurally untouchable zone, lowering the price consumes margin without changing position. In pharmacy, the offers that do not move do not move regardless of price. Full stop.
The second: the only event that resets the structure is a reset of the competitive alphabet. The launch of a new smartphone model resets the cycle: new products exceed 13% of the catalogue, price levels reopen, dynamics restart. A fashion collection change does the same. Black Friday does not: it opens margin temporarily (from €30 to €48 in household appliances) but recompresses within weeks. It is not a structural reset. It is a massage. Product innovation is not creativity: it is the only way to change the alphabet under constraint. Not all changes are resets. Changing channel shifts the boundary. Changing price consumes margin. Only a new product changes the alphabet.
The third: the structure is invariant across markets. Women's jewellery in Italy and women's jewellery in the UK produce the same repricing frequencies (8.4% vs 7.7%), the same discount rate (7.9% vs 8.0%), the same offer variability (8.9% vs 7.4%). On completely different scales: average price €255 versus £1,106, active merchants 15-25 versus 6-7. The regime is identical. The scale is not. You cannot solve a structural problem by changing market, channel, or currency. The structure follows you.
How to verify I am not talking nonsense
Take any e-commerce market. Look at the weekly distribution of repricing by category. If you find a smooth, gradual transition between sectors, this article is wrong. If you find discrete bands that jump, with no stable intermediate values, it is confirmed.
Then look at the percentage of offers that do not change price in a month. If it is significantly below 50% in a saturated market, this article is wrong. If it is around 70-80%, structural saturation is confirmed.
Every data point I cite is verifiable. Every URL is public. Every product code is traceable. In a sector where anyone says whatever they like, I chose the only possible path to validation: remove everything that is not essential and read what remains.
The framework
E-commerce was the last domain. Not the first.
The framework is called Sub-Limit Dynamics. To date it comprises more than 30 published papers with verifications across 23 independent domains. A USPTO patent has been filed on one of the applied mechanisms. Two articles have been published in Tom's Hardware Italia.
The full paper on the economic domain ("Margin as Limit", E1) contains the formal derivation, the data from all six markets, and the placement in the constraint atlas.
The framework and all papers are accessible at dailui.com.
If this seems impossible, verify the data. They are all there.