The Advertising Data Squeeze: Retail Media and Privacy 📡

Something quiet happened to online advertising.

The tracking cookie that followed you around the web mostly went away.

Advertisers needed a replacement. Shops had one.

They knew exactly what their logged-in customers had bought.

US advertisers are projected to spend $71.09 billion on retail media in 2026, up from $60.32 billion (Improvado, 2026).

That is a new advertising industry, built from the wreckage of an old one. 📡

🧾 Principaux résultats en bref

Mesure	Chiffre	Source
US retail media spend, 2026	$71.09 billion	Improvado (2026)
US retail media spend, 2025	$60.32 billion	Improvado (2026)
US growth rate, 2026	+17.2%	Improvado (2026)
Social media ad growth, for comparison	Below 11%	Improvado (2026)
Global retail media spend, 2025	$184 billion	Improvado (2026)
Global projection for 2030	Over $312 billion	Improvado (2026)
Competing 2026 estimates	$62B to $203.9B	Varies by definition

🍪 What Actually Changed

For twenty years, advertising ran on third-party cookies.

A cookie set by one site could follow you to another.

That is how an advert for shoes chased you across the internet.

Those cookies have largely left the advertising stack (Adtelligent, 2026).

Browsers blocked them. Regulators discouraged them. Users disliked them.

Why advertisers panicked

Targeting and measurement both depended on that tracking.

Without it, you cannot easily tell who saw an advert and later bought.

Advertising without measurement is just spending.

Why shops were the answer

A retailer does not need to follow you across the web.

You are logged in. You are on their site. You are buying.

They own the one signal advertisers want most: what you actually purchased.

That is called first-party data, and privacy rules treat it very differently.

📈 The Growth, in Context

Retail media is now the fastest-growing major advertising channel by percentage.

US spend rises 17.2% in 2026, while social media advertising grows below 11%.

The money moved from somewhere

Advertising budgets rarely grow as fast as any single channel.

So retail media is taking share, not just adding spend.

Every dollar here came out of search, social or display.

That is why the shift matters to anyone selling online.

Estimates disagree here too

You will see 2026 figures from $62 billion to $203.9 billion.

The spread comes from whether the figure is US or global, and what counts as retail media.

Some include in-store screens. Some include marketplace search ads only.

As with social commerce, check the definition before quoting.

💵 The Climb, Year by Year

US retail media spend went from $60.32 billion to $71.09 billion in a year (Improvado, 2026).

That is nearly $11 billion of new advertising money in twelve months.

Where that money is concentrated

A small number of very large retailers take most of it.

Scale matters here more than in most advertising markets.

You need millions of logged-in shoppers before the data is worth selling.

That is why retail media rewards the biggest players first.

The second wave

Mid-sized retailers are now launching their own networks.

Many will struggle, because advertisers do not want fifty separate ad platforms.

Expect consolidation, or aggregators that buy across several at once.

🌍 How Big Is It Globally?

Global retail media spend was around $184 billion in 2025 (Improvado, 2026).

Projections put it above $312 billion by 2030.

That implies roughly 11% growth a year over the period.

Why the global rate is lower than the US rate

The US market is further ahead, so its 17.2% reflects a maturing channel.

The global figure averages fast and slow markets together.

A slowing growth rate is normal and usually a sign of a real market.

🏪 How Retail Media Actually Works

The mechanics are simpler than the jargon suggests.

Format	What it looks like
Sponsored product listings	Paid placement in search results
Display ads on the retailer site	Banners on category pages
Off-site targeting	Retailer data used to target elsewhere
In-store digital screens	Ads at the shelf or checkout
Email and app placements	Sponsored slots in retailer messages

The first row is the bulk of the money.

When you search a marketplace, the top results are often paid.

Why brands pay for it

Because it sits at the exact moment of purchase.

Someone searching a retailer for “running shoes” is not browsing idly.

They are choosing between options, with a card in reach.

That intent is worth far more than a banner on a news site.

Why retailers love it

Advertising margins are enormous compared with selling goods.

A supermarket makes a few percent on groceries.

Advertising revenue can carry margins many times higher.

For low-margin retailers, this can be the most profitable part of the business.

⚖️ The Privacy Trade Nobody Voted For

Here is the part worth thinking about carefully.

Third-party cookies were removed to improve privacy.

What replaced them is data retailers already hold about your purchases.

That is not obviously more private. It is differently private.

Why it is arguably better

The data stays with a company you chose to buy from.

There is a real relationship, not an invisible network of trackers.

Rules on first-party data are clearer, and consent is easier to locate.

Why it is arguably worse

Purchase history is more revealing than browsing history.

What you bought says more about you than what you looked at.

And retailers now have a strong financial reason to collect more of it.

The incentive to track shifted rather than disappeared.

Résumé honnête

Cookie deprecation reduced one kind of surveillance.

It also created a market where shopping data became directly valuable.

Both statements are true, and most coverage reports only the first.

🧼 Clean Rooms: The Term You Will Keep Hearing

Data clean rooms are becoming standard in retail media deals (Rockbot, 2026).

The idea is straightforward once explained.

Two companies want to compare data without handing it over.

How they work

Both sides put data into a controlled environment.

Queries run across it, but neither party can extract the other’s raw records.

Only aggregated results come out.

It allows measurement without transferring personal data.

Pourquoi elles sont importantes

They let a brand answer the question advertising has always struggled with.

Did the people who saw this advert actually buy the product?

Without clean rooms, that question needs cross-site tracking.

The limits

Clean rooms are expensive and complex to run.

They favour large advertisers and large retailers.

Small sellers get the benefits only if their platform provides them.

🛒 What This Means If You Sell on Marketplaces

For small sellers, retail media is not an opportunity. It is a cost.

Effect	What it means for you
Top results become paid	Organic visibility falls
Competitors bid on your terms	Your brand searches get intercepted
Advertising becomes near-mandatory	Effective fee increase
Large brands outbid small ones	Costs rise over time
Data advantage sits with the platform	You see less than they do

The third row is the one that changes your economics.

When paid placement becomes necessary to be seen, it functions as a higher commission.

Que faire face à cela

Calculate your true cost of selling on a marketplace.

Include commission, fulfilment fees and advertising together.

Many sellers find the real rate is far above the headline commission.

Our guide des logiciels pour petites entreprises covers tracking these costs.

The case for owning your own channel

This is the strategic argument beneath all of it.

On your own site, you keep the first-party data.

You also keep the customer relationship and the email address.

Retail media makes the case for a direct channel stronger than it has been in a decade.

🔐 The Data Risk Nobody Mentions

Retail media rewards holding more customer data. That has a cost.

Data you hold is data that can be breached.

The average breach now costs $4.99 million, up 12% in a year (IBM, 2026).

A retail media business is, by design, a large customer database.

Why this connects directly

Advertising value comes from detailed purchase histories.

Detailed purchase histories are exactly what attackers want.

The same asset that earns advertising revenue creates breach exposure.

Our breach cost analysis covers what that exposure costs.

The retention tension

Good security practice says delete data you no longer need.

Retail media economics say keep everything, because history improves targeting.

Those two incentives point in opposite directions.

Most organisations resolve that tension in favour of revenue.

🏛️ What Regulators Are Watching

Three areas are drawing attention, and each could reshape the market.

Concern	What is being examined
Self-preferencing	Do platforms favour their own products?
Ad transparency	Is paid placement clearly labelled?
Purchase data use	Was consent meaningful?
Market power	Can sellers realistically opt out?

The last row is the sharpest question.

If advertising is effectively required to be visible, is it really optional?

Regulators tend to look closely once a choice stops being a choice.

Why this is a real risk to forecasts

Most projections assume current rules continue.

A ruling on self-preferencing or consent could change the economics quickly.

None of the growth forecasts quoted here price that in.

What sellers should do about it

Do not build a business that only works on one platform’s terms.

Keep a direct channel alive even if it is smaller.

That is insurance, not idealism.

📊 Where the Numbers Get Slippery

Retail media statistics need the same scepticism as any fast-growing category.

Question	Why it changes the number
US or global?	Roughly a threefold difference
Does it include in-store screens?	Adds substantially
Marketplace ads only, or all retailers?	Large effect
Gross spend or net revenue?	Agency fees included or not
Who published it?	Ad platforms benefit from big figures

That last row applies unusually strongly here.

Much retail media research is published by companies selling retail media services.

The number and the sales pitch often come from the same organisation.

What is well established

Despite the spread, the direction is not disputed.

Retail media is growing faster than social and search advertising.

Its growth is driven by privacy changes and by purchase-intent data.

Those conclusions hold across every source.

🔮 Where This Goes Next

Global spend was around $184 billion in 2025, with projections above $312 billion by 2030.

That implies roughly 11% annual growth over five years.

Note that this is slower than the 17.2% expected in 2026.

Why the forecast slows

Fast growth from a new category always decelerates.

The easy budget shifts happen first.

A forecast that slows over time is usually more credible than one that does not.

Straight-line projections of very high growth should raise suspicion.

What could change the picture

Regulation is the biggest variable.

If rules on using purchase data tighten, the model gets harder.

Competition regulators are also examining whether platforms favour their own products.

Neither risk is priced into most forecasts.

🧭 What Advertisers Should Actually Do

Three practical conclusions, regardless of size.

Measure incrementality, not attribution. Ask whether the advert caused the sale.

Retail media reports look excellent partly because they capture people already buying.

Someone searching your brand on a marketplace was likely to buy anyway.

The incrementality problem explained

Imagine paying to advertise to people already walking to your till.

The reported conversion rate would be superb.

The actual added sales would be near zero.

That is the central measurement risk in retail media.

Why platforms report such good results

They measure what happened after someone saw an advert.

They cannot easily measure what would have happened anyway.

That is not deception. It is a genuinely hard measurement problem.

But it means reported returns are almost always flattering.

Treat platform-reported performance as an upper bound.

How to test it cheaply

Pause a campaign for two weeks. Watch total sales, not campaign sales.

If total sales barely move, the adverts were capturing existing demand.

Very few advertisers run this test, because the answer is often uncomfortable.

🧾 A Worked Example: What Retail Media Really Costs a Seller

Abstract fees are easy to underestimate. Here is the arithmetic.

Take a seller with $100,000 of yearly marketplace sales.

Coût	Typical rate	Montant
Marketplace commission	15%	$15,000
Fulfilment fees	10%	$10,000
Advertising to stay visible	8%	$8,000
Total cost of the channel	33%	$33,000

The headline commission was 15%.

The real cost of selling was closer to a third of revenue.

Why advertising creeps upward

Bidding is competitive. As more sellers advertise, the price of visibility rises.

Sellers who do not bid lose placement to those who do.

The result is a slow, structural increase in cost.

This is not anyone behaving badly. It is how auctions work.

The comparison worth running

Work out the same total for your own direct channel.

Include payment fees, hosting, and the marketing needed to get traffic.

Many sellers find direct is cheaper than they assumed once ad costs are counted.

Our ecommerce growth analysis covers the wider market picture (U.S. Census Bureau, 2026).

📉 What Happened to the Advertising That Left

Retail media grew by taking budget from somewhere else.

Understanding where helps predict what happens next.

Channel	What changed
Display advertising	Hit hardest by cookie loss
Social advertising	Still growing, but under 11%
Search advertising	Resilient — intent-based already
Retail media	Absorbing the shift

Notice which channel held up.

Search survived because it never depended much on cross-site tracking.

Channels built on intent rather than surveillance were least affected.

The lesson underneath

Advertising that works because someone is actively looking is durable.

Advertising that works because you followed someone is fragile.

Privacy rules will keep tightening. Intent-based channels will keep winning.

What that means for small sellers

Focus on being findable when someone is searching for what you sell.

That is cheaper and more durable than any targeting strategy.

It is also the one advantage that does not depend on a platform’s rules.

🚫 What This Data Does Not Tell You

It does not prove retail media works better. Reported returns include existing demand.

It does not settle whether privacy improved. One form of tracking replaced another.

It is heavily US-weighted. Other markets are at different stages.

Much of it is vendor-published. Sellers of retail media publish the size of retail media.

It cannot see the counterfactual. Nobody knows what those budgets would have earned elsewhere.

🏁 La version courte

Third-party cookies largely disappeared, and advertising needed new targeting data.

Retailers had the best available substitute: what their customers actually bought.

US retail media spend reaches about $71.09 billion in 2026, growing 17.2%.

Social advertising grows below 11%, so budgets are moving.

For big brands this is a new channel. For small sellers it is a rising cost of being visible.

And the privacy story is more complicated than it appears.

Surveillance did not end. It moved somewhere with a better legal footing and a stronger business case.

That is worth saying plainly, because the usual telling is simpler.

Cookies were bad, they went away, privacy improved.

What actually happened is that the data moved closer to the till.

Whether that is better depends on what you think privacy is for. 📡

❓ Questions fréquemment posées

What is retail media?

Advertising sold by retailers on their own sites, apps and stores, targeted using their own customer purchase data.

How big is retail media in 2026?

US spend is projected around $71.09 billion, up 17.2%. Global figures differ widely depending on definition.

Why is it growing so fast?

Third-party cookies largely left the advertising stack, and retailers hold first-party purchase data that replaces them.

Is retail media better for privacy?

It is different rather than clearly better. Data stays with a company you chose, but purchase history is more revealing than browsing history.

What is a data clean room?

A controlled environment where two companies can measure across combined data without either extracting the other’s raw records.

Is retail media good for small sellers?

Usually it is a cost. As paid placements take the top results, advertising becomes necessary to stay visible.

How do I know if my retail media ads work?

Test incrementality. Pause a campaign and watch total sales rather than campaign-reported sales.

Why do published figures vary so much?

Whether the number is US or global, and whether in-store screens and off-site targeting are included.

What does retail media really cost a seller?

More than the headline commission. A 15% commission plus 10% fulfilment plus 8% advertising is a 33% cost of channel.

Why did search advertising survive cookie loss?

Because it was already intent-based. Someone searching tells you what they want without needing to be tracked (Adtelligent, 2026).

Does holding more customer data carry risk?

Yes. The average breach costs $4.99 million (IBM, 2026), and retail media works by keeping detailed purchase histories.

What are regulators looking at?

Self-preferencing, ad labelling, consent for purchase data, and whether sellers can realistically opt out of advertising.

Should I sell direct instead?

Retail media strengthens that case, because on your own site you keep the data, the margin and the customer relationship.

📚 Références

Improvado. (2026). Top retail media networks 2026: Rankings and benchmarksConsulté le 8 août 2026 sur https://improvado.io/blog/top-retail-media-networks

Rockbot. (2026). Retail media trends 2026Consulté le 8 août 2026 sur https://blog.rockbot.com/retail-media-trends-2026

Adtelligent. (2026). Retail media market outlook: Key data and growth forecastConsulté le 8 août 2026 sur https://adtelligent.com/blog/retail-media-market-outlook/

Bureau du recensement des États-Unis. (2026). Ventes trimestrielles du commerce électronique de détail : Premier trimestre 2026Consulté le 8 août 2026 sur https://www.census.gov/retail/ecommerce.html

IBM. (2026). Coût d'un rapport sur une violation de données en 2026Consulté le 8 août 2026 sur https://www.ibm.com/reports/data-breach

Lectures complémentaires sur ce site

Our social commerce analysis applies the same definition-checking to another contested market. The ecommerce growth piece gives the wider context, and our breach cost analysis covers the risk of holding more customer data.

À propos de cette analyse

Retail media research is unusually often published by firms that sell retail media services, and that is stated in the text rather than hidden. Where 2026 estimates range from $62 billion to $203.9 billion, the range is shown and the reasons explained. Figures were checked on August 8, 2026.