Sponsored Ad Formats in Retail Media: The Operator's 5-Type Playbook (2026)

Last updated: August 2026. Reviewed by Kunal Damgude, Growth and Product Marketing Manager, Osmos.

Sponsored ad formats in retail media split into five retailer-relevant categories: Sponsored Products (onsite search), Sponsored Display (onsite browse and retargeting), Sponsored Video (onsite and offsite shoppable video), Offsite (DSP and programmatic extension to Meta, Google Shopping, and DV360), and In-Store / DOOH (digital screens and connected store environments). Each format carries a different revenue profile, implementation cost, and measurement maturity, and the order in which a retailer or marketplace launches them determines how much revenue per advertiser dollar the network ultimately captures. For the platform-by-platform performance numbers each of these formats actually earns once they ship, across Amazon, Walmart, Instacart, and beyond, see our pillar guide on ROAS benchmarks by platform and ad format in 2026; this article is the format-mechanics and retailer launch-decision guide that sits underneath it.

This is a retailer’s guide. If you run a retail media network, marketplace ad business, or quick-commerce monetization stack, and you are deciding which sponsored ad formats to launch, in what sequence, with which tech, and at what pricing model, every section below is written from your seat. Brand and agency readers will still find the format mechanics useful, but the calls to action assume you own the inventory rather than buy it.

Why Sponsored Ads Define the Retail Media Revenue Stack

US advertisers spent $60.32 billion on retail media in 2025 and will spend $71.09 billion in 2026, according to a December 2025 EMARKETER forecast (eMarketer, January 2026). The concentration underneath that total is sharpening rather than levelling: by 2028 Amazon’s retail media revenues alone will pass $75 billion, more than $65 billion ahead of the next-largest network (eMarketer, May 2026). Inside the spend itself, the sponsored-ads category dominates: on the most recently published global breakdown, sponsored products take 40% of retail media budgets, sponsored brands 24%, display 20%, and video 16% (Nielsen, June 2025).

That mix is the single most important benchmark a new retailer will encounter, and it is a moving target. Offsite is growing about twice as fast as onsite across 2026 (Advertising Week, February 2026), and the multiple itself is the news: a year earlier the gap was closer to three times. It has compressed as onsite matured and offsite stopped being an experiment, which means the land-grab window on offsite is narrowing rather than widening. Video is where the steepest curve now sits: retail media CTV spend is expanding roughly three times faster than retail media search (eMarketer, February 2026). The format mix that produced your revenue base in 2024 is not the format mix that will produce it in 2027.

"RMNs can now deliver value across every stage of the buying journey, building awareness, driving consideration and strengthening loyalty." (AdExchanger editorial, The Full-Funnel Retail Media Network, October 2025)

Underneath those market-size numbers, Osmos is the format stack this article maps to: a single white-label system covering product, display, video, offsite, in-store, gamified, story, and influencer-live formats, so a retailer can sequence launches without stitching together a point solution per format.

The Five Sponsored Ad Format Categories in Retail Media

A sponsored ad in retail media is paid commercial inventory that appears inside a retailer or marketplace’s owned digital or physical surfaces (search results, browse pages, product detail pages, video carousels, in-store screens), typically priced per click or per thousand impressions and attributed back to the retailer’s first-party purchase data. The word "sponsored" distinguishes paid placements from organic results; the words "retail media" distinguish them from open-web display or social ads bought against third-party audiences.

There are two ways to organize the format taxonomy. The Amazon trichotomy of Sponsored Products, Sponsored Brands, and Sponsored Display is widely cited but pulls every retail media network into a single retailer’s product naming. The retailer-agnostic taxonomy used by IAB Europe and Nielsen is more useful when you are deciding what to build. IAB Europe’s updated 101 Guide to Retail Media (July 2025) splits the universe into digital on-site, digital off-site, and digital in-store formats, with sponsored products, display, and video as the three primary on-site sub-types. We extend that into a five-category retailer framework:

Inside each section below we cover the mechanics, the revenue profile, the launch prerequisites, and the retailer decision framework for that format. Two things sit slightly outside this scope and have dedicated coverage elsewhere: native ads in editorial and recommendation feeds, and the emerging engagement formats (gamified, story, carousel) that live as capabilities inside categories 2 and 3 rather than as separate categories. On the European baseline, on-site retail media ad spend grew 22.2% to over €10 billion in 2024, with over 90% of advertisers now partnering with retailers (IAB Europe, July 2025). The format-portfolio question is no longer hypothetical for any non-trivial retailer.

A note on vendor naming. We deliberately do not use Amazon, Walmart, Criteo, or Target product names in the section headings of this article. They appear inside the sub-sections as worked examples, because they are the most documented format launches in the public record, but the framework is retailer-agnostic. Amazon’s specific Sponsored Products, Sponsored Brands, and Sponsored Display trichotomy is one implementation of categories 1 and 2, not the universal map.

1. Sponsored Products: The Onsite Search Revenue Anchor

Sponsored Products are paid product listings that appear in a retailer’s onsite search results, category browse pages, and sometimes on product detail pages, ranked by a combination of bid (typically CPC) and relevance to the shopper query. They are the highest-intent, lowest-friction sponsored format and the entry point for every retail media network we have seen launched.

Sponsored Products carry no uploaded creative. The ad is auto-assembled from the retailer’s own product listing: catalog image, title, price, rating. The advertiser does not supply or optimize an ad asset for it, which means the only quality levers available are listing completeness and product-detail-page quality. That single mechanical fact separates this format from every other one in this article. Sponsored brands, onsite display, and onsite video all take an advertiser-uploaded asset, so an onsite programme is not creative-free; sponsored products are simply the one format inside it that is.

The mechanics are simple in concept and complex in execution. A retailer sets aside a configurable share of search results, often two to four positions per page, for paid placements. Advertisers bid on keywords, or more commonly today on auto-targeted product attributes, and the auction selects winning ads on a blend of bid value and predicted relevance. Inventory is theoretically infinite, because every search query is a new auction, but practically constrained by query depth and ad load: at low query volumes the retailer runs out of bidders, and at high ad loads organic relevance suffers.

The market-leader benchmark on ad load comes from Amazon: 99% of Amazon searches display sponsored products, with an average of 20 sponsored listings per page load (eMarketer, January 2025). Mid-market RMNs sit far below that line. Staples, by way of contrast in the same eMarketer report, "increased sponsored listing coverage by 17.2% in H2 2024" but "only 68% of searches showed sponsored products." The 30-point gap is the bid-density problem in one number: most non-Amazon networks have far more empty inventory than they have advertisers willing to bid on it.

That is also the retailer’s opportunity. Search still claims the majority of retail media spending and remains the entry point for advertisers into the channel (eMarketer, January 2026), which means the incremental dollars flowing into a $71.09 billion 2026 market land disproportionately in this format first. The networks that capture a disproportionate share of them will be the ones that solve bid density first, through better contextual targeting, smarter advertiser onboarding, and self-serve campaign tools that lower the activation bar.

"Is it a distracting, intrusive ad? Or is it something equally as relevant to an organic search result?" (Andreas Reiffen, CEO, Pentaleap, quoted in eMarketer, January 2025)

The relevance test is the long-running retailer dilemma for sponsored products. Push ad load too aggressively and search quality degrades; under-monetize and you leave revenue on the table. Most networks land somewhere between Staples’ 68% and Amazon’s 99% and tune from there.

Retailer launch checklist for Sponsored Products:

• A first-party product catalog with structured attributes (brand, category, price, availability) refreshed at least daily
• A keyword or attribute auction with a relevance scoring function, because pure bid auctions destroy search quality within weeks
• A self-serve advertiser console with budget caps, daypart controls, and pacing
• A reporting layer that returns impressions, clicks, ad-attributed sales, and ROAS at SKU granularity
• A bid-density growth plan: contextual auto-targeting, advertiser incentives for under-bid categories, house ads to fill remaining inventory

Osmos’s Product Ads module covers the auction, contextual targeting, inventory sync, and one-click campaign setup that closes the gap between catalog and revenue-generating inventory. ML-powered contextual promotion is the lever that disproportionately improves bid density on non-Amazon RMNs where keyword coverage is sparse: the network synthesizes bidder demand by matching advertiser intent to query context rather than waiting for advertisers to manually keyword-load.

Pricing model: CPC auction is the dominant pricing for Sponsored Products. Some networks add reserve floors per keyword cluster or premium placement uplifts; both are pricing tweaks rather than separate models.

2. Sponsored Display: Onsite Browse, Retargeting, and the Banner Renaissance

Sponsored Display, also called sponsored banner or onsite display, covers paid placements that appear on browse pages, category pages, product detail pages, the cart, and the post-purchase confirmation surface. Targeting is contextual (the page or category being browsed), behavioral (recent onsite shopper actions), or audience-based (first-party shopper segments). Pricing is typically CPM with a layer of fixed-tenancy reservations for premium placements like the homepage carousel. Unlike sponsored products, every unit in this category takes an advertiser-uploaded creative asset, which is why creative review, spec compliance, and production support become real operating costs the moment a retailer launches it.

Placements that match the look and feel of the surrounding page are a fast-growing slice of this category rather than a category of their own; see our native advertising in retail media benchmarks guide for format-level performance data.

Sponsored Display is where most retailers see their first revenue diversification beyond sponsored products. Where sponsored products are high-intent and conversion-led, display is upper-funnel and brand-led, which is why the buyer pool is different (more brand budgets, more category-marketing budgets) and why CPMs are typically richer on a per-impression basis. For retailers, display has three structural advantages: a small number of premium placements can carry six- and seven-figure tenancy deals from CPG brands and category captains; ad-load tuning is more forgiving than search; and creative production cycles are familiar to brands that already buy banner display elsewhere.

A specific worked example: Target’s Roundel generated nearly $2 billion in value as of mid-2025, with a stated goal to double that over the following five years (Path to Purchase Institute, July 2025); no more recent public figure has been published since. Roundel’s portfolio "spans on-site sponsored listings, off-site programmatic, social, CTV, video, audio, influencer, out-of-home, and in-store formats," a multi-format retailer pulling on every category in this taxonomy. The Precision Plus buying solution announced in 2025 layers first-party data and real-time shopper behavior on top of the display inventory. This is the trajectory most ambitious RMNs follow once their sponsored-products engine is producing reliably.

Hybrid formats live inside this category as well. The Product Display Ad (PDA) sits between sponsored product and display: a banner-shaped placement with product context (price, image, "Add to cart") that competes for the same surfaces as banner display but converts more like sponsored product. PDAs are useful for retailers whose advertisers have catalog feeds but limited creative production capacity, because the unit assembles from the feed rather than from a bespoke asset.

For deeper history on how display formats evolved from generic banners into the targeted placements in use today, our companion piece on the display ad evolution in retail media covers the design and inventory shifts in detail.

Retailer launch checklist for Sponsored Display:

• An inventory catalog of every browseable surface (homepage, category pages, PDP, cart, post-purchase) with placement IDs and slot specs
• An ad server that supports both auction and tenancy fulfillment in the same surface, which most legacy GAM deployments do not
• A creative review layer covering brand safety, ad-spec compliance, and accessibility
• First-party audience segmentation tied to onsite browse and purchase signals
• A pricing strategy that mixes CPM auction floors with reserved tenancy for premium surfaces

Osmos’s Display Ads and PDA modules cover both the auction-driven and tenancy-driven surfaces, with first-party targeting and geo targeting built in. An AI content-validation layer checks every uploaded creative before it ships, which removes one of the more expensive ops-cost line items in display advertising.

Pricing model: CPM with fixed tenancy on premium surfaces. A CPC option is available on some hybrid PDA placements but is not the dominant model.

3. Sponsored Video: The Fastest-Growing Format Category

Sponsored Video covers shoppable video ads served on a retailer’s onsite surfaces (search results, category pages, PDP carousels) and increasingly on offsite environments (CTV, partner ecommerce sites, social) using the retailer’s first-party data. It is the format category with the most launch activity of the past two years and the strongest documented performance evidence in the retail media research base. Like display, it is creative-bearing: the advertiser supplies the asset, and the retailer’s spec and review workflow is what determines whether mid-market brands can participate at all.

The single most cited retailer-side video test is Albertsons’ beta of Criteo Onsite Video (April 2025): a 280% increase in click-through rates and a 460% sales lift when Onsite Video was paired with Sponsored Products, plus a 5.6x lift in new-to-brand customers when video was combined with display and sponsored products in a unified plan. Digital Commerce 360’s independent coverage of the same test confirms the headline numbers (Digital Commerce 360, April 2025). Six months later, Kroger Precision Marketing launched Onsite Video Carousels on Kroger.com and the Kroger app, combining short-form video with shoppable product listings inside search results. Kroger did not publish performance figures at launch, but the format’s existence on a top-three US grocer’s primary search surface tells you where this category is heading.

"Video has always been a powerful storytelling tool but rarely a direct driver of commerce, until now." (Stephen Howard-Sarin, Managing Director, Retail Media for the Americas, Criteo, press release, April 2025)

The new-to-brand framing matters more than the conversion stats for a retailer thinking about format mix, and it is worth being precise about what it means: new-to-brand counts a shopper who has not previously bought the advertiser’s brand, not a shopper who is new to your store. A weekly regular who has never bought a given brand is new-to-brand for that advertiser. Walmart Connect’s Sponsored Videos product page reports that 91% of sales from Sponsored Videos come from new-to-brand customers for SMB advertisers in FY2025 (Walmart Connect, Sponsored Videos); the qualifier matters, because this is the SMB-advertiser cohort across February 2024 to January 2025, not a universal rate across all brand sizes. For Sponsored Brands, 69% of orders come from new-to-brand buyers in the same period (Walmart Connect, Sponsored Brands). What both figures tell a retailer is the same thing: video and brand-led formats reach buyers an advertiser has not previously transacted with, which is the audience incremental retail-media budget is meant to capture.

CTV is the offsite extension of the same format category, and retail has already become the largest single ad-spending category on connected TV, at just under one-fifth of all CTV spend (eMarketer, February 2026). That is the scale a retailer is stepping into rather than creating. Format mechanics differ (full-screen 15-second to 30-second video, no in-feed competition, household-level targeting) but the retailer-side decision, whether to license data to a CTV platform or package the inventory jointly, ladders into the same category. Walmart Connect’s April 2026 launch of Connect Select, a curated premium CTV marketplace inside Walmart DSP carrying Vizio, Paramount, and Warner Bros. Discovery inventory, is the clearest sign of where a scaled network takes this next (PYMNTS, April 2026), and in June 2026 Walmart extended retail media video into YouTube through DV360 (Chain Store Age, June 2026).

Retailer launch checklist for Sponsored Video:

• A creative onboarding workflow that accepts vertical, square, and landscape variants with auto-trim
• An ad server that serves video without an external SDK on mobile web, since the SDK requirement is a frequent advertiser blocker
• An attribution layer that ties video impressions to subsequent ad-attributed sales, ideally with a new-to-brand cut
• A creative cost-share program for advertisers without in-house video production, which is the bottleneck most networks under-invest in
• A measurement story that handles both view-through and click-through attribution per format

Osmos’s sponsored video module covers autoplay, video analytics, and ML-served creative variants. Story ads and shoppable influencer-live units extend the category into full-screen and creator formats, both useful additions once base video is producing reliably.

Pricing model: CPM premium, often two to three times the CPM of static display on the same surface. Tenancy options exist for category-takeover formats.

4. Offsite (DSP and Programmatic Extension)

Offsite formats extend a retail media network’s first-party data and ad inventory into environments the retailer does not own: Meta, Google Shopping, DV360, the open programmatic web, and increasingly CTV. The retailer is the data and audience source; the inventory comes from external publishers and platforms. From a retailer’s perspective, offsite is the format where revenue per advertiser dollar is lower per impression but the addressable spend pool is meaningfully larger.

The growth differential is the most-cited stat in the category. Offsite retail media spend is expanding at roughly twice the onsite rate through 2026 (Advertising Week, February 2026), a multiple that has narrowed from closer to three times a year earlier. The direction is unchanged, the urgency is not: offsite is still where incremental retail media budget is moving, and retailers that can offer both surfaces capture a structurally larger share of advertiser planning, but the window in which offsite alone counts as a differentiator is closing.

The retailer’s case for offsite is built on three pieces of evidence. First, demand is there: 96% of brands and agencies are open to buying on-site retail media through a DSP, 80% say DSP buying would make it easier to shift more budget to retailers, and 66% cite more cost-efficient buying as the top DSP advantage, per Koddi’s survey of 126 Fortune 1000 brand and agency decision-makers (Koddi, May 2025). The 2026 picture is consistent: Stratably’s 2026 State of Retail Media report, produced in partnership with Skai and surveying 166 brand and agency organizations, finds that retail media leaders now allocate 27% of media budget to retail media against 23% for laggards, and activate across 7.2 networks on average against 6.2. Advertisers are not choosing between networks so much as adding them, and the multi-network buyer needs a programmatic route in. Second, performance evidence is public: AdExchanger’s case study of Proximo Spirits documents 60% new-to-brand sales, 200% ROAS, and 10x higher off-site impression volume than on-site in an integrated campaign combining onsite and offsite formats (AdExchanger, October 2025). Third, audience expansion: offsite is where a retailer reaches the shopper who visits the site occasionally or not at all. Our piece on audience monetization beyond onsite ads covers that demographic logic in more detail.

Two structural facts about offsite are worth carrying into the pricing decision. Social advertising accounts for nearly half of offsite retail media spending, and 72% of retail media advertisers name video as one of their main offsite spending areas (eMarketer, January 2026). Social is also not a single creative model: some of it is dynamic, catalog-assembled inventory, and a large share of it is advertiser-uploaded reels, videos, and story units. A retailer planning offsite creative operations needs to support both, and the specs and policies for the uploaded half are set by platforms the retailer does not govern.

"The fragmented, one-off nature of today’s RMNs is seen as a barrier, not a benefit." (Koddi programmatic report, May 2025)

The retailer implication: advertisers are tired of stitching one-off integrations across four to six retail media networks. The networks that win offsite share will be the ones that make data and inventory accessible through the DSPs advertisers already use (Meta Ads, Google’s Shopping ad stack, DV360) rather than asking each advertiser to learn yet another vendor console. That is also why 63% of the largest RMNs now offer offsite search capabilities, per eMarketer’s January 2026 FAQ (eMarketer, January 2026): the format is becoming table stakes rather than a differentiator.

Retailer launch checklist for Offsite:

• A first-party data layer mature enough to export shopper segments to external platforms with privacy and consent enforcement
• Activation integrations with at least Meta, Google Shopping, and one DSP (DV360 or The Trade Desk)
• A clean-room or matched-cohort capability so advertisers can measure offsite-driven sales back to the retailer’s POS
• A creative operations path for both catalog-assembled and advertiser-uploaded social formats, each with its own spec and review flow
• A pricing strategy that handles both CPM programmatic and managed-service models, because the two buyer types for offsite are different

Osmos’s offsite retail media module activates Meta, Google Shopping, and DV360 from a single retailer console, a multi-channel design that addresses the fragmentation pain Koddi documents. For most mid-market RMNs, offsite is the phase-two launch after a stable sponsored-products engine, not the phase-one launch.

Pricing model: CPM programmatic for the bulk of inventory, with managed-service overlays for advertisers that need full-funnel campaign management.

5. In-Store and DOOH: The Newest Format Category

In-store retail media covers digital screens, smart carts, aisle endcaps, and connected store environments: paid placements served at the physical point of sale. DOOH (digital out-of-home) extends the category into store-adjacent surfaces like fuel stations and pickup areas. It is the youngest format category in this taxonomy and the most operationally complex, because it requires hardware deployment, in-store attribution capability, and a creative spec that differs from every digital sibling format.

AdExchanger has noted the format diversity already in market: "smart carts, gas pumps, Coinstar machines, digital price tags, aisle displays, and in-store audio" (AdExchanger, December 2024). Target Roundel’s portfolio expansion includes "in-store hubs in high-traffic areas with demos, sampling, and digital screens" (Path to Purchase Institute, July 2025). Osmos’s Advertima partnership covers the in-store audience-measurement layer that connects screen exposure to sales.

The clearest 2026 evidence on both scale and proof comes from Albertsons. The Albertsons Media Collective is expanding in-store digital screens to 800 of its 2,200-plus stores during 2026, up from an 80-store pilot the previous summer, and has signed more than 50 advertising partners since launch (Digiday, January 2026). A beta campaign for Mondelez’s Sargento Cheese Bakes across 116 stores returned a $2.41 matched-market incremental ROAS and a 14% lift in in-store sales (MediaPost, January 2026). That pairing, a third of the estate wired and a matched-market lift number to show for it, is the template a retailer should be aiming at: footprint first, then a controlled test that produces a number a brand’s finance team will accept.

"The way that we’re planning it is to get as large of a footprint as quickly as possible." (Liz Roche, VP of media and measurement, Albertsons Media Collective, quoted in Digiday, January 2026)

The rest of the US grocery and pharmacy field is moving the same way. CVS expects roughly 11,000 digital screens nationwide by the end of 2026 and is doubling front-entrance screens from 500 to 1,000, while Kroger has framed its own commitment as "an opportunity for the in-store environment to be used as a brand-building channel" (Christine Foster, SVP commercial strategy, Kroger, in Modern Retail, January 2026). Best Buy takes a different route into the same budget: rather than pricing screens individually, it offers the whole store as one takeover package spanning windows, entrance, TV walls, and PC monitor displays, bundled with its year-round NFL partnership and a Dude Perfect content deal, and backs it with the claim that 93% of its transactional revenue connects to a customer ID (Retail Dive, September 2025).

Two retailer caveats. First, in-store launches typically lag onsite launches by 18 to 24 months, because hardware and creative-spec problems are not solvable inside an ad server. Second, measurement is earlier in maturity than for any other format here, and the industry is still settling on what counts as an in-store impression, which is why the category is priced on store count and campaign duration rather than on CPM. Format-specific depth on measurement standards, targeting hardware, and attribution lives in our dedicated in-store retail media monetization playbook.

Onsite vs Offsite: The Retailer Revenue Trade-off

The single most consequential format-mix decision a retailer makes is how to balance onsite and offsite. The trade-off is structural, not preference-driven.

According to Mirakl, on-site retail media advertising margins often reach 70-80%, while off-site margins are typically lower due to external media costs (Mirakl, November 2025). That single stat captures the retailer economics: every onsite impression keeps the media cost inside the network, while every offsite impression pays a media tax to Meta, Google, or a DSP. So why launch offsite at all?

Three reasons drive retailers across the bridge. First, the addressable advertiser base is meaningfully larger offsite, because the long tail of brands without the assortment or price points to win onsite auctions can still be activated against retailer audiences in offsite environments. Second, reach against buyers who have never purchased a given advertiser’s brand is structurally higher offsite (the Proximo Spirits 10x impression volume figure is the canonical example), and that is the audience a brand is paying incremental money to find. Third, advertisers want the convenience of programmatic buying through DSPs they already use; declining to offer it cedes share to networks that do.

The retailer sequencing recommendation is consistent across the research base: build onsite first, then expand offsite once the first-party data infrastructure can support clean-room or matched-cohort activation (Mirakl, November 2025). Our playbook on monetizing onsite traffic covers the seven plays that fill that inventory before any offsite extension is worth attempting. Networks that try to launch offsite before they have a stable onsite measurement story tend to under-monetize the offsite extension, because they cannot prove incrementality back to the retailer’s POS.

Format Performance Metrics and Measurement

Performance benchmarks by format type (ROAS, CPM, conversion rate) are covered in depth in our ROAS benchmarks by platform and ad format in 2026 pillar. We do not duplicate that table here. Instead, three retailer-relevant measurement realities by format:

Sponsored Products is the format with the cleanest closed-loop attribution, because the click and the purchase happen on the same retailer surface. Standard reporting includes impressions, clicks, ad-attributed sales (last-click within a configurable window), and ROAS. Retailer decision: how to handle the view-through credit window for high-consideration purchases like electronics and apparel, where the click-to-buy path stretches over multiple sessions.

Sponsored Display and Sponsored Video require incrementality measurement to be meaningfully evaluated. Click-through attribution alone systematically under-credits these formats, because their primary value is upper-funnel: driving consideration that converts later through other paths. The Albertsons and Criteo Onsite Video test result (280% CTR, 460% sales lift, 5.6x new-to-brand lift when paired with Sponsored Products) is meaningful precisely because the test was built to measure the combined-format effect, not the video format in isolation.

Offsite formats require closed-loop or clean-room measurement to tie offsite-served impressions back to onsite or in-store purchases. IAB Europe’s Commerce (incl. Retail) Media Measurement Standards V2.1 is the current version, superseding V2, and it codified a clearer measurement funnel plus flexible ad-size guidelines that make cross-environment measurement comparable across networks (IAB Europe). The compliance grace period that allowed reporting against either V1 or V2 ran to the end of July 2026 and has closed, so a network still reporting on the older definitions is now outside the standard rather than inside a transition window. Retailers that adopt the current version remove one of the largest friction points in advertiser conversations: whether the network measures the same way the buyer’s plan does.

"Marketers in most industries today tend to rely more on in-platform attribution tools than independent third-party measurement solutions to gauge the success of their full-funnel retail media campaigns." (Nielsen, The future of retail media, June 2025)

The measurement gap Nielsen identifies is the retailer’s opportunity to differentiate. Networks that ship a third-party-verified measurement story (incrementality, new-to-brand, geo-holdout, MMM-compatible) capture a structurally larger share of brand budgets than networks relying on in-platform self-reporting alone.

Pricing Models by Format Type

Revenue-model architecture is covered in depth in our companion spoke on retail media monetization models. Here is the pricing proxy for each of the five categories.

Sponsored Products price on a CPC auction, with relevance scoring alongside the bid as a non-negotiable input; reserve floors per keyword cluster are the most common retailer tweak. Sponsored Display prices on a CPM base with fixed tenancy layered onto premium surfaces, and tenancy is where the six- and seven-figure category-takeover commitments live. Sponsored Video carries a CPM premium over static display on the same surface, which is the clearest evidence that the market already treats attention-heavy formats as a different product. Offsite prices programmatically on CPM, with managed service layered on for buyers who want the campaign run for them.

In-store and DOOH are the exception, and CPM is the wrong unit for them. Price in-store on a cost-per-store, cost-per-campaign, or cost-per-day basis tied to store count and campaign duration. Genuine impression-based pricing requires the multi-sensor deployments most retailers have not installed, so a CPM quoted against an estimated footfall figure is a modelled number wearing the clothes of a measured one, and a brand’s analytics team will eventually work that out. The underlying economics are hardware amortization across screen-uptime hours, which is precisely why duration and store count are the honest units to price on. The same logic applies to targeting: in-store inventory is bought by place and time, which store, which day, which hour, not by the individual shopper standing in front of the screen.

Auction mechanics (bid evaluation, second-price against first-price, floor-pricing dynamics) are a deep topic in their own right; our pillar on retail media auction mechanics covers the full retailer playbook for auction design and yield optimization.

Beyond the Five Formats: Building and Pricing a Custom Ad Format

The five categories above are the taxonomy. What sits across all five is a pricing decision: whether to design a differentiated, high-impact unit that exists on your surfaces and nowhere else, and price it above the standard rate card. This is not a sixth format category. It is a strategy a retailer applies inside categories 2 and 3, and in 2026 it is the clearest route into brand awareness budgets that a CPC search auction structurally cannot reach.

A custom ad format is an ad unit a retailer designs, names, and controls on its own surfaces, which an advertiser cannot buy in the same shape on any other network. Two 2026 launches show what that looks like in practice.

Costco Velocity: Reserved Display. In March 2026, Costco’s retail media arm unveiled Reserved Display, the first new ad format built natively on the Costco Velocity stack, in partnership with the AI adtech firm Moloco. Placements run on the highest-traffic pages Costco owns, the homepage and search results, and the company describes them as "personalized digital endcaps": highly visible, highly relevant, and modelled on the in-aisle merchandising device Costco members already recognise from the warehouse floor (eMarketer, March 2026; Path to Purchase Institute, March 2026). Moloco’s machine learning is trained on members’ real purchase behavior and integrates with Costco’s identity-resolution and audience-building infrastructure, so an advertiser can activate a custom Costco audience without a separate setup step. Optimization is tied to sales and incremental growth rather than raw impressions, which is the detail that keeps a high-visibility unit from becoming a vanity-impressions play. Beta partners got access in Q2 2026, with broader availability and an SSP and DSP programmatic connection following through Q2 and Q3.

"We aren’t building a media network to hide margin. We are building an engine to move merchandise." (Mark Williamson, Associate Vice President, Costco Retail Media, quoted in eMarketer, March 2026)

Home Depot Orange Apron Media: Category Leaderboards. Home Depot took the same idea onto category landing pages. At its third annual InFronts in April 2026 it announced "high-impact Category Leaderboards on category landing pages," launched alongside a self-serve Brand Page Portal, geotargeted banner ads, and sponsored brands display formats built for more immersive brand connections that were still in development at announcement (Home Depot corporate newsroom, April 2026; Forbes, May 2026). The detail that matters for a retailer pricing this kind of unit is how Home Depot reports on it: results are framed against the campaign’s stated goal, whether that is awareness, product-launch support, shopper education, or consideration, rather than forced into a single sales-return formula for every placement. A format priced on awareness has to be reported on awareness, or the buyer falls back to judging it on ROAS and the premium collapses.

Neither company publishes a rate card, and neither says in its own words that it prices these units at a premium. The pricing argument is Osmos’s, and it is this: a format that exists on exactly one retailer’s surfaces is not comparable to anything else on a media plan, and non-comparable inventory is the only inventory a retailer can price on value rather than on the prevailing CPM. A sponsored-products auction clears at whatever the second-highest bidder will pay for a click. A leaderboard that exists only on your category pages clears at whatever an awareness budget will pay for that position. Those two revenue lines behave completely differently, and the second one draws brand budget rather than trade budget, which is the money most mid-market RMNs never see.

Three constraints separate a durable custom format from an expensive one-off:

• It has to be creative-bearing and spec-stable. These are display-class and video-class units, so the advertiser uploads an asset. Publish the spec, hold it still for at least a season, and give mid-market advertisers a production path, or only the largest three brands in each category will ever buy it.
• It has to be measured against the goal it is priced on. If the unit is priced for awareness, report reach, new-to-brand share, and lift, not last-click ROAS. Home Depot’s goal-based reporting is the working example.
• It has to survive contact with the ad server. A bespoke unit that cannot be trafficked, capped, brand-safety-reviewed, and reported alongside the rest of the portfolio becomes a one-off worked by hand, and hand-worked exceptions do not scale past a handful of deals per quarter.

Retailers that would rather adapt vendor-supplied high-impact inventory than build their own should read our companion guide on native advertising in retail media benchmarks, which covers the auction-based display route in depth. Osmos supports the build route inside the onsite channel: gamified and engagement units, full-screen story ads, carousel, and shoppable influencer-live formats ship as part of the onsite format set rather than as a bespoke build, and each of them carries the retailer’s own branding rather than a vendor’s.

Format Comparison: Retailer Coverage at a Glance

This table compares format portfolio coverage across major retail media platforms. Pricing is covered separately in the monetization sibling spoke; this is purely a portfolio-completeness lens. Available means the format is in market and broadly accessible; Partial means it exists but in limited inventory or selected categories; Enterprise-only means it requires a direct managed-service engagement; Not available means no documented format offering.

Sources: format coverage compiled from Walmart Connect Sponsored Videos and Sponsored Brands (FY2025 new-to-brand data), PYMNTS coverage of Walmart Connect Select (April 2026), the Criteo Onsite Video launch (April 2025), Path to Purchase Institute Roundel coverage (July 2025), and eMarketer retail media search (January 2025) for Amazon sponsored-products coverage.

A note on competitor strengths. The table reads as a portfolio map, not a quality judgment. Each named competitor has documented strengths the table does not capture. Amazon’s bid density (99% query coverage) is unmatched and gives advertisers a confidence floor no other RMN can match. Walmart Connect’s new-to-brand rates for SMB advertisers demonstrate incremental audience reach that rivals Amazon for performance brands, and its curated premium CTV marketplace is a route into brand budgets most networks do not have. Criteo’s onsite video general availability and the Albertsons test results give it the most documented video performance evidence in the industry. Target Roundel runs the most complete vendor-managed multi-format portfolio in the US market, spanning onsite, offsite, and in-store.

The differentiation Osmos brings is format completeness for retailers in one white-label stack: gamified formats, story ads, carousel, influencer live, and multi-channel offsite activation in a single retailer console rather than stitched across vendor partnerships, deployed as a retail media operating system the retailer brands as its own.

Format Launch Sequencing: The Retailer’s Roadmap

If you are launching a new retail media network, or rationalizing an existing format mix, sequencing is the single biggest decision after build-versus-buy. The sequence below is synthesized from the research base and from the launch decisions we have watched retailers make on Osmos deployments.

Phase 1: Sponsored Products (months 0 to 3). Lowest tech complexity, highest bid-density potential, fastest ROAS for advertisers, fastest revenue recognition for the retailer. This is the entry-point format for every retail media network. Get it stable, get the relevance scoring tuned, and onboard 50-plus advertisers before adding format complexity.

Phase 2: Sponsored Display and PDA (months 3 to 6). Inventory expansion across homepage, category pages, and PDP, upper-funnel revenue capture, and access to category-marketing and brand budgets that never show up for sponsored products. Tenancy reservations on premium surfaces start producing six-figure deals once display is stable.

Phase 3: Sponsored Video (months 6 to 12). Premium CPM tier, new-to-brand audience reach, and the format that unlocks brand-led budgets for awareness campaigns. Creative cost-share programs are essential; without them only the largest brands launch and the format under-monetizes. The Albertsons and Criteo test (280% CTR lift, 460% sales lift when video was paired with sponsored products) is the retailer-side proof point most advertisers accept.

Phase 4: Offsite and DSP (months 9 to 18). Requires first-party data infrastructure mature enough to export segments with consent enforcement. Activate Meta, Google Shopping, and a DSP (DV360 or The Trade Desk) in parallel, because single-channel offsite under-delivers on the addressable spend pool. Build the clean-room or matched-cohort capability before launching, not after.

Phase 5: In-Store and DOOH (months 12 to 24). Hardware deployment, audience capability, and in-store attribution are 12 to 18-month build cycles in their own right. Most networks layer in-store on top of a stable digital portfolio rather than sequencing it before digital is producing. Albertsons’ path, an 80-store pilot in 2025 followed by an 800-store expansion in 2026, is the shape this normally takes.

A differentiated custom format is not a sixth phase. It is something a retailer can attempt from Phase 2 onward, once display inventory and a creative review workflow exist, and it is usually the fastest way to convert an established sponsored-products business into one that also collects awareness money.

The revenue case for moving through this sequence rather than camping at Phase 1 is documented in our companion piece on retail media margin contribution in a tight economy and in our argument for offsetting customer acquisition costs with retail media. The short version: Phase 1 alone produces a profitable RMN; Phase 1 through Phase 4 produces an RMN that materially shifts the retailer’s overall margin structure.

"Measurement is being used as an incentive for a lot of these partnerships, because it’s something that’s in such high demand." (Anthony Costanzo, Chief Analytics Officer, Mile Marker, quoted in Digiday, July 2025)

The gating constraint at every phase is measurement. Each new format launches more cleanly when the retailer can prove cross-format incrementality, which is why the measurement story (closed-loop, clean-room, standards-compliant) is the actual operating system underneath the format portfolio.

The implementation tooling matters too. Osmos covers the ad operations layer, wallet management, advertiser onboarding, campaign review, AI brand-safety validation, and feature gating, which disproportionately determines whether new formats onboard advertisers fast enough to clear bid density. A format suite without an ops layer under it tends to under-monetize, because advertisers cannot self-serve. Sitting above that layer are the revenue-management decisions this section is about: yield management across format types, demand generation for under-bid categories, advertiser insights, bring-your-own-traffic, and house-ads management. Retailers making format-mix calls at the revenue-management level need that layer inside the same stack rather than in a spreadsheet beside it.

Format Mix by Retailer Vertical: Worked Examples

Different retail verticals optimize for different format mixes. Three retailer-side worked examples:

Grocery and quick commerce. Sponsored Products dominates because purchase frequency and category breadth produce massive query volume; the Amazon, Kroger, Albertsons, and Instacart playbooks all start here. Onsite video is the upper-funnel layer for CPG brand spend that does not fit neatly into the sponsored-products auction. The Albertsons and Criteo test and the Kroger Precision Marketing Onsite Video Carousels launch (October 2025) are the canonical grocery proof points, and grocery is also where in-store is scaling fastest. Offsite CPG activation through Meta and Google Shopping is the Phase 4 expansion. Osmos serves the grocery vertical with onsite, in-store, and offsite capability deployable from one stack.

Fashion and beauty. A video-forward format mix, because the buying decision is visual and aspirational. Story ads and shoppable influencer-live units carry disproportionate weight, since the creator economy is where this category’s discovery happens. Sponsored Products still drives conversion, but the ratio tilts more heavily toward video and full-screen formats than it does in grocery. Indian beauty marketplace Purplle, an Osmos client, has built a fashion and beauty vertical RMN with this format mix.

Pharmacy and health. Sponsored Products and Sponsored Display dominate, because category-level targeting (allergy season, cold-and-flu, OTC categories) is the retailer’s primary monetization lever. Video plays a smaller role, since category-marketing budgets in pharma are more constrained by regulatory creative review. Apollo 24x7, also an Osmos client, runs a pharmacy-vertical RMN built around that core with sponsored search expansion.

Retail Media Ad Formats in India and Emerging Markets

India’s retail media format landscape is roughly 18 to 24 months behind the US on absolute inventory volume but is moving fast on CTV and streaming integration. Amazon India’s 2026 advertising trends (February 2026), authored by Priyanka Khaneja Gandhi, Head of Marketing and Creative Solutions at Amazon Ads India, reports that India’s streaming audience has passed 600 million, with connected-TV users growing 87% year on year to 129 million, citing the Ormax OTT Audience Report 2025. The implication for retailers: streaming and shoppable-video formats are arriving in India faster than display or offsite DSP integration.

"Technology should amplify human creativity, not replace it." (Priyanka Khaneja Gandhi, Head of Marketing and Creative Solutions, Amazon Ads India, About Amazon India, February 2026)

Indian marketplaces have built competent sponsored-product engines. Flipkart Ads supports sponsored product listings, banner ads, video ads, and offsite formats; Amazon India has the most mature format suite by virtue of being a multinational platform with India-localized inventory; JioMart and Meesho have launched sponsored-product modules but lag on video and offsite extensions. Indian beauty and pharmacy retailers such as Purplle and Apollo 24x7 are the early adopters of multi-format stacks beyond sponsored product, often using white-label tooling like Osmos to deploy onsite, offsite, and in-store formats from one retailer console rather than stitching together vendor partnerships.

The India playbook differs from the US in three places. First, sponsored-product bid density is a binding constraint earlier: fewer advertisers, lower keyword coverage, higher reliance on contextual targeting to fill auctions. Second, offsite via Meta and Google Shopping is a higher share of the addressable advertiser spend pool than in the US, because programmatic display infrastructure is more concentrated. Third, video formats are launching directly into CTV and streaming environments rather than passing through a desktop-display intermediate phase. The format leapfrog is real.

Format Implementation Challenges (and What to Do About Them)

The four challenges most retailers hit when launching beyond sponsored products:

Bid density sparsity. Most non-Amazon RMNs see sponsored-product coverage in the 50% to 80% range of search results against Amazon’s 99%; the eMarketer Staples comparison at 68% is representative. The fix is contextual auto-targeting (turning keyword-poor categories into auction-eligible inventory algorithmically), advertiser incentives for under-bid categories, and house-ads fill-in for residual unfilled inventory.

Advertiser adoption bottleneck. Advertisers default to sponsored products because they are cheap to test and easy to attribute. Convincing them to test display, video, or offsite requires proving incrementality, which most networks struggle to deliver because they lack the measurement infrastructure. The fix is to ship third-party-verified incrementality measurement before pushing format expansion, not after.

Creative quality and velocity. This is the constraint that separates the four creative-bearing formats from sponsored products, and it is usually under-planned. One-size-fits-all creative for display under-performs, and video requires separate production workflows and spec compliance. Most networks under-invest in creative cost-share programs, and the result is that only the largest brands can launch the new formats. The fix is retailer-funded creative production for mid-market advertisers, often delivered through automated creative platforms.

Network fragmentation as a barrier. Advertisers work with four to six retail media networks simultaneously. Each new format a retailer launches competes for the same limited advertiser attention and campaign budget. Buyer-side fatigue with one-off network integrations is well documented (Koddi, May 2025), and every additional format asks the same overstretched team for more attention. The fix is making the retailer’s stack accessible through the platforms advertisers already use: DSP integrations for offsite, standardized measurement for cross-network comparability, and self-serve tooling that lowers the activation cost for every new format.

Frequently Asked Questions

How does Koddi’s format offering compare to a full-stack retail media platform?

Koddi is an ad-server and DSP-enabled platform optimized for programmatic access to retail media inventory, and it is strong on auction mechanics, DSP buying, and retailer-side bid optimization. It is not a full-stack format suite: gamified formats, story ads, in-store hardware, and white-label retailer branding sit outside its core. Retailers evaluating a stack should compare Koddi against full-stack offerings on three axes: format portfolio breadth (does the platform cover all five categories in this article?), the ad operations layer (advertiser onboarding, brand safety, wallet management), and the revenue strategy layer (yield management, demand generation, bring-your-own-traffic). Koddi excels on the first axis within programmatic; full-stack alternatives such as Osmos extend the second and third.

What technology do retailers need to launch in-store retail media formats?

In-store retail media requires four technology layers that pure digital formats do not: digital screen hardware deployed in-aisle, in-cart, at endcaps, or at checkout; an audience layer that decides which store, which day, and which hour a given ad runs in, rather than reacting to whoever happens to be walking past; an attribution capability that ties in-store exposure to in-store or post-visit purchases, typically through QR-code interaction, loyalty-card scan, or matched-market testing; and a creative spec that handles short dwell time, no audio, and high ambient light. Osmos’s partnership with Advertima covers the audience layer. The operational complexity is why in-store typically lags onsite digital launches by 18 to 24 months.

How is in-store retail media measured?

In-store measurement is the youngest discipline in this taxonomy. The dominant approaches are screen-uptime logs and computer-vision audience counting for exposure; QR-code scans, in-cart-display interactions, or NFC taps for interaction; matched-market or geo-holdout test and control panels comparing stores running the format against stores that are not; and loyalty-program closed-loop attribution where shopper IDs link in-store exposure to subsequent transactions. Matched-market testing is currently the method producing numbers brands accept, as in the $2.41 incremental ROAS and 14% sales lift Albertsons published from its 116-store Mondelez beta. The IAB and IAB Europe in-store standards defined store zones and reporting metrics that retailers are now starting to adopt, but harmonization across networks is still in progress rather than settled.

What are the five sponsored ad formats in retail media?

The five retailer-relevant sponsored ad formats are Sponsored Products, the onsite search placements that anchor most retail media revenue; Sponsored Display, covering onsite browse, retargeting, and banner placements; Sponsored Video, the fastest-growing category spanning onsite and shoppable offsite video; Offsite, the programmatic extension of retail media audiences to DSPs, Meta, Google Shopping, and DV360; and In-Store and DOOH, digital screens and connected store environments. Sponsored Products typically launch first and carry the largest share of spend, while In-Store and DOOH are the newest and most operationally complex.

What is the difference between sponsored products and sponsored display?

Sponsored Products are keyword and category-targeted placements in onsite search results, billed on CPC through an auction; they capture high-intent shoppers at the moment of search and are the single largest sponsored-format revenue line for most networks. Sponsored Display covers the paid placements spread across the rest of the shopper journey, from browse and category surfaces through to the cart and the post-purchase page, targeted contextually, behaviorally, or by first-party audience, and usually billed on CPM with fixed-tenancy options for premium inventory. The operational difference that matters most to a retailer is the creative model: a sponsored product is auto-assembled from the product listing and takes no uploaded asset, while a sponsored display unit takes an advertiser-supplied creative and therefore needs a spec, a review workflow, and a production path for smaller advertisers.

Where to Go From Here

If you are a retailer weighing format order, tech stack, and revenue model, three next reads close out the framework:

1. Benchmark depth. Our pillar on ROAS benchmarks by platform and ad format in 2026 covers the platform-by-platform performance numbers for every format described here, the "what does this earn once it ships?" companion to the mechanics in this guide.
2. Revenue model mechanics. Retail media network monetization models covers CPM, CPC, and tenancy revenue architecture, and how to model format-level yield.
3. Auction design. Retail media auction mechanics and bidding covers the bid-evaluation, floor-pricing, and yield-optimization mechanics that determine how much revenue any given format actually produces.

Which formats to launch first, how to sequence them against the data and inventory you already have, and how to model format-level revenue contribution before anything gets built: that is the conversation Osmos has with retailers every week, across onsite, offsite, and in-store. For the wider strategic frame around it, our guide to building and benchmarking a retail media network covers the build, buy, and benchmark decisions end to end.

Sources

1. eMarketer: FAQ on retail media networks: How marketers should allocate budgets in 2026 (January 9, 2026)
2. eMarketer: Retail Media Ad Spending Forecast H1 2026 (May 5, 2026)
3. eMarketer: FAQ on CTV advertising: Trends, formats, and platforms to watch in 2026 (February 15, 2026)
4. Advertising Week: Why Off-Site Retail Media Is Growing Twice as Fast as On-Site (February 26, 2026)
5. eMarketer: Costco Velocity retail media Reserved Display launch coverage (March 20, 2026)
6. Path to Purchase Institute: Costco's Retail Media Network Adds AI Onsite Ads (March 24, 2026)
7. The Home Depot: Orange Apron Media Unveils New Partnerships, Onsite Offerings, and Certification Program at Third Annual InFronts (April 22, 2026)
8. Forbes: Home Depot's Orange Apron Media Targets High Intent Shoppers And Pros (May 4, 2026)
9. Digiday: Albertsons is putting digital screens for ads in more than a third of its stores (January 22, 2026)
10. MediaPost: Albertsons Aims To Show What In-Store Ads Actually Deliver (January 6, 2026)
11. Modern Retail: Retailers like Kroger, CVS plan many more screens for in-store ads in 2026 (January 6, 2026)
12. Retail Dive: Best Buy builds retail media with in-store takeovers, sports deals (September 22, 2025)
13. PYMNTS: Walmart Simplifies Streaming Ads for Mid-Sized Business (April 27, 2026)
14. Chain Store Age: Walmart extends retail media network ads into YouTube streaming video (June 11, 2026)
15. IAB Europe: Commerce (incl. Retail) Media Measurement Standards V2.1
16. IAB Europe: Updated 101 Guide to Retail Media (July 15, 2025)
17. Nielsen: The future of retail media (June 4, 2025)
18. Criteo: Criteo Introduces Onsite Video to its Retail Media Mix (April 23, 2025)
19. Digital Commerce 360: Albertsons rolls out shoppable video ads with Criteo, sharing early test results (April 29, 2025)
20. Kroger Precision Marketing: NEW Onsite Video Carousels Turn Inspiration into Conversion (October 21, 2025)
21. AdExchanger: The Full-Funnel Retail Media Network: Bridging On-Site And Off-Site For Seamless Customer Journeys (October 7, 2025)
22. AdExchanger: Is 2025 The Year Of Retail Media Standardization? (December 19, 2024)
23. Digiday: Retail media's mid-2025 reality: Why advertisers are going all in on full-funnel (July 8, 2025)
24. eMarketer: Retail media search in 2025: Balancing sponsored ads with the customer experience (January 21, 2025)
25. Koddi: The state of programmatic retail media: 2025 trends and insights (May 13, 2025)
26. Mirakl: A retailer's guide to key retail media ad formats (November 6, 2025)
27. About Amazon India: The future of advertising in India: 5 trends brands need to know for 2026 (February 13, 2026)
28. Path to Purchase Institute: Target's Roundel Eyes Growth With New Tools, Precision Strategy (July 23, 2025)
29. Walmart Connect: Sponsored Videos (FY2025 data, February 1, 2024 to January 31, 2025)
30. Walmart Connect: Sponsored Brands (FY2025 data, February 1, 2024 to January 31, 2025)
31. Stratably, in partnership with Skai: The 2026 State of Retail Media: What Separates Leaders From Laggards, survey of 166 brand and agency organizations (January 26, 2026)