1. Retail media in numbers — a market that outgrew the forecasts
2026 is a breakthrough year for retail media for several reasons. First, scale: according to the report BCG "The $200 Billion Opportunity in Retail Media" (2026), global retail media ad spend reached $152 billion, up 28% year over year. For context, the market was worth barely $30 billion in 2020. No other digital advertising channel is growing at this pace.
Second, profitability. The McKinsey report "Retail Media Networks: The New Profit Engine" (Q2 2026) shows that operating margins from retail media for leading retailers run to 70–85% — levels comparable to pure-play technology platforms rather than traditional retail. For retailers operating on net margins of 2–4%, retail media can mean doubling the profitability of the entire business.
Poland's retail media market is also gaining momentum. Data from the IAB Polska "Retail Media 2026 Report" shows that the value of the Polish retail media market has exceeded PLN 1.2 billion, growing 40% year over year. The leaders are Allegro (Ads), convenience chains (Żabka with Żappka Ads), grocers (Biedronka, Lidl) and marketplace platforms. For context: the entire Polish online display advertising market is roughly PLN 5.5 billion — retail media already accounts for more than 20% of that pie, and the share is growing.
"Retail media is growing faster than social media and video advertising combined. In 2026 it is already more than 20% of the entire digital advertising market. For low-margin retailers this is not extra revenue — it is a new foundation for profitability."
2. Two models: broker versus operator — a strategic choice, not a technical detail
The fundamental decision every retailer entering retail media faces concerns the choice of operating model: broker or operator. The difference between them is not cosmetic — it determines the revenue structure, control over data, supplier relationships and the scalability of the business.
The broker model (reseller / managed service)
In the broker model, the retailer resells its advertising inventory to an external partner (e.g. Criteo, CitrusAd, PromoteIQ) that takes on the technology, sales and operations. The retailer provides only the audience (customer data) and the inventory (space on its own channels). This model is quick to launch — a pilot can be running within 3–4 months — but it has three fundamental limitations.
First, control over data: the technology partner operates on the retailer's customer data, creating a risk of losing the first-party data advantage to competitors. Second, profitability: the broker takes 20–40% of ad revenue, leaving the retailer a share that, at low scale, does not cover operating costs. Third, dependency: the retailer becomes dependent on a technology partner whose roadmap does not necessarily align with its needs.
The operator model (self-service / owned platform)
In the operator model, the retailer builds its own retail media network: its own DSP/SSP platform, its own sales team, its own ad server. This is a far more costly and time-consuming path — a full implementation takes 12–18 months and an investment of several million zloty — but it gives full control over data, margins (70–85% instead of 30–50%) and supplier relationships.
According to the GroupM report "This Year, Next Year: Retail Media 2026", of the 50 largest global retailers, 68% chose the operator model as their target, even if they started with a broker. The trend is unambiguous: the operator model is not a matter of choice but of timing.
"The choice between the broker and operator model is the most important strategic decision in retail media. A broker gives you a quick start at the price of control over data and margins. An operator builds a long-term advantage at the cost of higher initial investment. There is no third way."
3. Amazon, Walmart and the Polish players — lessons from different speeds
The global retail media market offers three clear reference models, each carrying different lessons for Polish retailers.
Amazon — the operator model in its extreme form
Amazon Advertising is today the world's third-largest digital advertising player (after Google and Meta), with ad revenue exceeding $56 billion in 2025 (up 24% year over year). The key to Amazon's success is not the technology — though it is sophisticated — but the closed data ecosystem: Amazon combines purchase data (what the customer bought), intent (what they searched for), browsing (what they watched) and demographics into a single unified profile that enables precise targeting in a closed-loop model — from ad exposure to purchase completion.
For Polish retailers the lesson is clear: retail media without your own transactional data is just another advertising channel. Retail media built on first-party data is an advantage no pure-play adtech can copy.
Walmart — from broker to operator in 3 years
Walmart Connect is the most instructive case study for mid-sized and large retailers. In 2022 Walmart operated largely on the broker model, working with The Trade Desk as an external DSP. Within 3 years it completed the transformation: it took over its own ad server (through acquisition), built its own sales team (more than 500 people), launched a self-service Walmart DSP platform and — crucially — opened its ecosystem to off-site advertising (ads outside its own channels, targeted with Walmart data). The result? Walmart's ad revenue grew from $2.1 billion in 2022 to $5.9 billion in 2025, and the advertising segment's margin topped 75%.
The lesson for Polish retailers: the transformation from broker to operator can be done in stages, but it requires a clear roadmap, budget and — above all — the board's courage to invest in the team and technology before revenue covers the costs.
Polish players — the pilot phase with ambitions
Poland's retail media market is at an earlier stage of development than the US or UK markets, but it is growing faster in percentage terms. Allegro Ads — the largest Polish retail media network — generated an estimated PLN 800–900 million in ad revenue in 2025, roughly 15% of the platform's total revenue. Żabka with Żappka Ads has built a loyalty-and-advertising model that combines mobile app data (more than 10 million users) with ad targeting across the convenience ecosystem. Biedronka is testing in-store formats (digital signage screens) as a retail media channel — a direction that is only just emerging in the West and that, in Poland, could become a first-mover advantage.
The challenge for Polish retailers is not awareness of retail media's potential but operational readiness: a lack of in-house adtech teams, immature data infrastructure (most retailers do not have a unified customer profile), limited budgets for technology investment and — crucially — an underestimation of the investment required to move from the broker to the operator model.
In practice this means Polish retailers face a decision that cannot be postponed. International suppliers — Procter & Gamble, Unilever, Nestlé — are already allocating 15–25% of their trade marketing budgets to retail media, and by 2027 that share will rise to 30–40%. A retailer that cannot offer them its own retail media network simply will not get those budgets — they will go to competitors or to marketplace platforms (Allegro, Amazon) with mature advertising ecosystems. In other words: retail media is no longer an option. It is a condition for maintaining relationships with key suppliers.
4. Five trends that will shape retail media in 2026–2028
Drawing on BCG, McKinsey, IAB and GroupM reports and on my own consulting projects, I identify five trends that will change the rules of the game in retail media over the next 24 months.
Trend 1: Off-site retail media — retailer data beyond its own channels
The biggest architectural shift in retail media. Retail media ads have so far appeared mainly on retailers' own (on-site) channels: the website, the mobile app, search results. In 2026 market leaders — Amazon, Walmart, Carrefour — are opening their data to ads across the open web (off-site): social media, programmatic display, CTV. This lets a supplier target a retailer's customers not only at the moment of purchase but across the entire journey, from brand awareness to conversion. For retailers this is a new revenue stream that does not require growing on-site traffic.
Trend 2: In-store retail media — digital signage as an advertising medium
The digitalization of the store environment opens retail media up to the physical world. Digital signage screens, interactive shelves, audio retail media — these are channels that reach the customer at the moment of highest purchase intent: in the store. According to the IAB "In-Store Retail Media Guide 2026", in-store retail media will grow at a 45% CAGR through 2028. For Polish retailers investing in store modernization, this is a chance to build an advantage — provided the screens are not just a vehicle for their own content but an advertising platform for suppliers.
Trend 3: AI in retail media — from targeting to creative
Artificial intelligence is transforming retail media on three levels. First, AI-driven targeting: predictive algorithms analyze transactional data in real time, optimizing bidding and budget allocation. Second, generative creative: GenAI tools produce ad creative variants (images, copy) tailored to audience segments — Amazon is testing this through its "Creative AI Ads" program. Third, AI-powered measurement: closed-loop attribution using machine learning that shows the real ROI of a campaign rather than vanity metrics (impressions, clicks).
Trend 4: Measurement standardization — the end of the "walled gardens" era
The retail media industry suffers from a lack of common measurement standards. Every retailer reports campaigns differently, making it hard for suppliers to compare effectiveness across platforms. In 2026 IAB launched the Retail Media Measurement Standards v2.0, which define unified metrics: incremental ROAS, share of shelf uplift, category halo effect. For Polish retailers this is a signal: if they do not adapt to the standards, they will fall out of international suppliers' budgets, which expect comparable data.
Trend 5: Platform consolidation — the "retail media OS"
The retail media market is entering a phase of technological consolidation. A new category is emerging — the "retail media operating system" — integrated platforms that combine ad serving, DSP, SSP, DMP/CDP, measurement and billing in a single solution. Leaders (Criteo, Epsilon, PromoteIQ) are building end-to-end platforms that let retailers launch a retail media network in a hybrid model — their own data and supplier relationships, external technology. For Polish retailers this is a real alternative to the full operator model: faster, cheaper, yet still offering control over data.
5. The technology stack — what you really need to get started
Conversations with the boards of Polish retailers often surface the same problem: retail media seems technologically complicated, so the decision is postponed "until we are ready". Yet the minimum viable retail media stack is simpler than it appears. Here is what you actually need to run a pilot within 4–6 months.
- Unified Customer Profile (CDP/DMP) — combined transactional, behavioral and demographic data. Without this there is no targeting. Minimum: integrating data from POS, e-commerce and the loyalty program in one place.
- Ad Server + SSP — the technology for managing campaigns and auctions. To start, a white-label solution (Kevel, Adswizz) or a partner platform (Criteo Retail Media, PromoteIQ) is enough.
- Measurement & Attribution — tools for measuring campaign performance in a closed-loop model. Minimum: integration with the transactional system showing whether ad exposure translated into sales.
- A retail media sales team — at least 2–3 people dedicated solely to selling advertising space to suppliers. This is the most commonly overlooked element — technology without a seller generates no revenue.
- Ad formats — to start: sponsored listings in search plus on-site display. Expansion: off-site programmatic, in-store digital signage (phase 2).
The key principle: do not try to build everything from scratch. In 2026 the market offers mature platforms that let you launch a retail media network in a hybrid model — your own data and relationships, external technology. The full operator model is a goal for 2028, not a condition for starting.
It is also worth planning your data architecture now for future requirements. A retail media network that today runs on basic demographic segments will, in 18 months, need to support real-time behavioral targeting, integration with the programmatic ecosystem and incremental ROAS measurement. Investing in a clean, structured data layer at the start costs 20% more but saves 200% in migration costs later. This is the lesson Polish retailers should draw from British retailers (Tesco, Sainsbury's), who built retail media on legacy systems and then paid twice over to modernize them.
6. Recommendations for Polish retailers
Eight strategic recommendations for the boards of Polish retailers that want to build retail media as their primary margin engine — not a side project.
- Treat retail media as a business, not a marketing project. A retail media network is a separate business unit with its own P&L, team and goals. It cannot be parked under the marketing department — there it will always lose the fight for resources to day-to-day campaigns.
- Start with data, not technology. The unified customer profile is the foundation. Without it, even the best platform will not generate an advantage. Invest in integrating data from POS, e-commerce and the loyalty program before choosing a technology provider.
- Choose the hybrid model to start. Do not build your own DSP in the first year. Use partner platforms (Criteo, PromoteIQ, Epsilon), keeping control over data and supplier relationships. Build the operator model in stages — the way Walmart did.
- Hire a retail media sales team. This is the fastest-returning investment. Two or three salespeople with experience in retail media or programmatic advertising who know the suppliers, their trade marketing budgets and can translate the value of the retailer's data into the language of ROI.
- Standardize measurement from day one. Adopt the IAB Retail Media Measurement standards. Report incremental ROAS, not just clicks and impressions. Suppliers — especially international ones — expect comparable data across platforms.
- Do not ignore in-store. If you are modernizing stores, design digital signage screens as retail media channels from the start — not as an afterthought. In-store retail media is the fastest-growing segment, and Polish retailers have a chance at first-mover advantage here.
- Build supplier relationships on a consultative-sell model. Do not sell "ad space" — sell access to insights about their category and consumers. The best retail media networks show suppliers how to grow the whole category, not just their own brand.
- Plan off-site as phase 2 — but plan it from the start. Off-site retail media (ads targeted with retailer data beyond its channels) is the biggest growth market. Start on-site, but build off-site into the 18-month roadmap from the beginning. Suppliers already expect to reach a retailer's customers across the open web.
Retail media in 2026 is not a fad, not an experiment and not a side project. It is a structural change to the retail business model, in which customer data and advertising space become an asset that generates higher margins than selling products from the shelf. Retailers that grasp this in 2026 and make investment decisions before the end of Q4 will build an advantage that is hard to copy. Those that stay with the broker model in the hope that "the market will grow and things will work out" will hand their most valuable asset — customer data — to intermediaries.
If you are wondering which model to choose for your network and how to build a retail media network that genuinely increases margin — I help retail boards design and implement retail media strategies from concept to first revenue. Let's talk.