Retail Media Automation and CPC Auctions: What a Retailer Should Automate in 2026

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

A retailer scales its retail media network by automating five layers in order: the cost-per-click auction that prices every ad slot, the bidding tools advertisers use inside that auction, campaign setup and approvals, reporting back to the SKU, and the catalog feed and budget pacing that keep ads accurate and on plan. Automate the auction and its floors first, because every other layer depends on what the auction charges and which ad it lets win.

Osmos is a white-label retail media operating system for retailers and marketplaces. Its modular apps for onsite, offsite and in-store ads sit over your existing stack, so you can break your monetisation ceiling faster without rebuilding it.

What does retail media automation cover?

Retail media automation is software that runs the decisions inside a retailer's ad business, from pricing a placement to pausing an out-of-stock product, so that more advertisers can run campaigns without the retailer adding ad operations headcount for each one. Osmos builds this automation into the retailer's own network: Osmos Product Ads prices clicks in an auction the retailer controls, and Osmos Orchestration takes setup, approvals and billing off the ad operations team's desk.

Osmos groups the work a retail media network automates into five layers, listed in the order a retailer should automate them:

• Auction and pricing: the cost-per-click auction ranks eligible ads, applies floors and charges the winner, with no person approving each placement.
• Bidding: advertisers set a goal or a budget and the platform adjusts bids, while the retailer keeps control of floors, eligibility and relevance.
• Campaign workflow: onboarding, briefs, creative checks, wallets and billing run as one workflow inside the platform.
• Reporting: order-level results reach the advertiser by SKU, category and keyword without an analyst building each report.
• Catalog, feed and pacing: the product feed builds the ads, stock changes pause them, and pacing rules keep spend on plan through the day.

When a retail automation platform built for retail marketing, pricing or store operations is shortlisted for a media project, check that it ranks and charges for ad placements and that it reports ad-driven sales back to the advertiser. For the wider case for automation as the reason networks scale beyond their top advertisers, see Automation & Auctions: How Retail Media Scales.

Which retail media solutions run a cost-per-click auction model?

Osmos is a retail media solution with a cost-per-click auction model: on Osmos Product Ads, advertisers bid per click for sponsored listings across search, category and product pages, floor CPCs adjust to demand on every ad slot, and premium floors apply automatically on popular slots. Among the largest in-house networks, Amazon has the best-documented CPC auction: it says it has priced clicks on Sponsored Products using a form of generalized second-price auction since 2006 (Amazon, 31 August 2026). Topsort and Kevel, which power other retailers' networks, offer their own auction designs.

A cost-per-click auction is a pricing model in which advertisers bid the most they will pay for a click on a placement. The platform ranks bids by price and relevance, and the winner pays only when a shopper clicks. For the mechanics of first-price and second-price clearing, floor pricing and bid shading, see How Retail Media Auctions Work.

Does relevance count as much as the bid? Amazon states that in 2024 approximately 92% of the Sponsored Products ads it selected were not the highest bid, often by a wide margin, and describes the trend in its own words: "Over time, our ranking formula has increasingly given greater weight to relevancy over bid amount" (Amazon, 2026). Osmos Product Ads ranks sponsored listings by shopper intent, scoring each impression on dozens of behavioral signals that include shopper intent, category, price affinity, brand affinity, cart and purchase history alongside the bid. For how ad servers turn those signals into ranking, see AI-powered ad serving and behavioral targeting.

Can floors and reserve prices move with demand? A fixed floor gives advertisers a predictable entry price; a floor that moves with demand also captures more on busy keywords and lowers the bar for small advertisers on quiet ones. Amazon says it is testing soft reserve prices, "a real-time minimum value that seeks to better reflect what each placement is actually worth," alongside a hard reserve, "the minimum a bid must surpass to enter an auction" (Amazon, 2026). Osmos's dynamic yield control does this per slot on Product Ads, lifting yield without the ad operations team retuning floors by hand.

Can the retailer explain every price it charges? Advertisers renew budgets with networks they trust, and auction transparency is now an industry standard. The fairness side of that question is covered in Transparent Auctions: How Retailers Can Stay Fair and Profitable.

Can premium inventory run as an auction too? The most valuable keywords and slots are often still priced by negotiation, which caps both the price and the number of advertisers who get a chance at them. A platform that can put fixed placements into a scheduled auction, with a base price and visible bidding, removes that cap.

How BigBasket moved fixed tenancy ads into a pre-auction

BigBasket, one of India's largest grocery platforms, used Osmos to move its Fixed Tenancy Ads from one-by-one negotiation to a self-serve pre-auction, with a 93% average uplift from base price to final auction price. Before the change, the team had identified about 300 high-volume, high-intent keywords; media sales called advertisers individually and prices were set through separate negotiations run on spreadsheets. With Osmos, each keyword or keyword group became an ad package with a base price, auction dates and flight dates, visible to eligible advertisers, who bid through a white-labeled advertiser portal while BigBasket saw every bid. Fixed Tenancy Ads generated 50% of total ad revenue in the first 30 days, and advertiser participation grew 27% within two months (How BigBasket rebuilt fixed tenancy ads to unlock scalable search monetization). The packaging layer behind it is Osmos Ad Packages, and the pricing logic for premium bundles is set out in How to Build Retail Media Ad Packages Brands Will Pay a Premium For.

Which CPC-auction platform fits which retailer?

How do you automate bidding on a retail media network?

Automated bidding on a retail media network offers advertisers goal-based bidding, either a target return or a budget to spend as productively as possible, on top of the CPC auction. The platform refreshes suggested bids for advertisers who set their own, and the retailer keeps control of floors, eligibility and relevance rules. AI-driven CPC bidding software does the per-keyword arithmetic at auction speed.

On Osmos, planning and bidding start from the catalog. Osmos Product Ads launches a campaign in one click, auto-selecting products from the catalog based on behavioral signals. Its Media Forecaster recalculates spend, impressions, orders, revenue and ROAS in real time as campaign settings change, so an advertiser sees the effect of a bid before committing. Osmos's Sofie AI Agents, listed as coming soon on the Osmos site, are designed to diagnose underperformance and rank fixes to keywords, bids and product selection; advertisers review each suggestion, and one-click auto-apply is on the roadmap.

Instacart's current docs offer optimized bidding to maximize sales or new-to-brand sales alongside manual bidding, fall back to 1.5x of the target goal when the advertiser sets no input, and state that "We only support exact match keyword bids" for optimized bidding (Instacart Ads docs). Walmart's developer changelog says suggested bids are refreshed daily for Sponsored Products, Sponsored Brands and Sponsored Videos, and that suggestions now extend to keywords already in a campaign (August 2026) (Walmart Connect release notes).

The retailer keeps its own levers around the bidding algorithm, and the floor is the most visible of them. From 23 July 2026 Walmart cut the minimum bid for Sponsored Videos to $0.50 from $0.80, a change that lets smaller advertisers into a format they were priced out of (Walmart Connect release notes). Relevance thresholds, brand distancing, budget caps and the choice of which formats accept goal-based bidding are the others. Bidding automation should keep tail advertisers spending without a person tuning their bids, while the network's pricing rules stay intact.

How do you troubleshoot automated bidding and auction errors?

When automated bidding is working, spend tracks the daily plan and win rates stabilize after the learning period. When results drift, the cause is usually a setting or a data input, so check those before blaming the algorithm.

1. Respect the learning phase. Instacart's docs put it at about 3 to 7 days, and judging, pausing or re-targeting a campaign inside that window resets it.
2. Check that the target is realistic. Instacart warns that "Target ROAS that's too high might result in lower spend, lower sales, and a lower win rate" (Instacart Ads docs). A target ROAS set too high lowers the bid the algorithm can submit, which reduces the campaign's win rate.
3. Look at relevance before raising bids. When an ad keeps losing, the listing or keyword match is often the problem. Because ranking weighs relevance, a higher bid on a poorly matched product buys little, so bidding automation that only raises bids loses to fixes that improve the match between ad and query.
4. Compare floors with what tail advertisers bid. If the floor is higher than tail advertisers bid, the network loses their spend.
5. Check the catalog and match type. Out-of-stock products, broken listings and missing keywords look like bidding failures, and they are the most common cause of a campaign that will not deliver. Work through them in order: stock status first, then the listing (title, image, price), then the keyword list. Where a platform supports exact match only, as Instacart's optimized bidding does, a query variant missing from the keyword list gets no delivery at all. Raising the bid on that campaign changes nothing, because the ad never enters the auction for that query. Floors that adjust to demand per slot remove a separate cause, covered in step 4.

Which tools automate retail media planning and campaign execution?

Osmos Orchestration automates retail media planning and campaign execution for retailers: advertisers plan, launch and report through one white-labeled portal, while setup, approvals and billing run as a single workflow shared by the ad operations, finance and sales teams. The Osmos Orchestration page lists a 3x advertiser capacity lift and 5 to 10 FTEs of ops effort saved, with 20+ systems integrated and 60+ manual touchpoints removed.

Osmos Orchestration splits campaign execution across these apps, starting with the one advertisers use and ending with finance:

• Brand Self-Serve: brand-facing ad buying on top of the retailer's existing ad server, so the long tail of advertisers runs campaigns without a person on the retailer's side setting each one up.
• Campaign Manager: within Brand Self-Serve, ad operations teams configure creative, targeting, budget and schedule once, and a single brief runs across every selected channel with per-channel overrides.
• Inventory Studio: every ad placement across formats and channels is added and configured in one place.
• BYO Media Channels: campaigns are planned, run and reported across the retailer's own channels and vendor-fulfilled placements in one system.
• Wallets & Billing: brand wallets, MDF and incentives reconcile against real campaign delivery, so finance and ad operations stop keeping separate books.

Advertisers on Osmos see real-time campaign status and per-channel brief overrides today; full self-serve mid-flight configuration for advertisers is a future direction on the Osmos roadmap. Orchestration also sits on top of the tools a retailer already uses and connects to them through pre-built connectors. A retailer that wants to outsource campaign execution can use Osmos Ad Ops as a Service, where a dedicated Osmos team runs delivery, ROAS and wallet share for the network.

The same planning questions apply beyond a retailer's own site. Commerce media, the wider category covering any company that monetizes purchase data, needs the same plan, launch and report cycle; What Is Commerce Media? explains how it relates to retail media. A single brief matters most when a campaign spans onsite and offsite channels, covered in the guide to offsite channels for a retail media network. As agents take on more of the planning, the system that coordinates them decides how well they work, the argument made in Multi-Agentic AI Orchestration in Retail Media.

Retail media network growth comes from torso and tail advertisers. When a network can serve only the advertisers its account managers call, headcount sets its ceiling and every manual handoff lowers it. Automating onboarding, briefs and billing lets the same team carry many more advertisers, and the margin effect of that shift is worked through in The Retail Media Business Case: Margin, Not Revenue.

Which enterprise tools automate brand-safe ads for a retailer?

A retailer automates brand-safe ads on two layers: auction rules that control which ads may appear next to each other and next to which products, and creative review that checks uploaded assets before they reach the queue. On the auction layer, Osmos Product Ads applies brand distancing on every auction call, so same-brand products never appear side by side, lets retailers configure brand-set rules in the Control Center, and pauses out-of-stock products automatically.

Creative review applies only to formats that take an uploaded asset, such as display, video, sponsored brand units and offsite placements; Product Ads are assembled from the product listing, so there is no asset to review. For upload formats, the practice that scales is to score every creative against the retailer's format, quality and policy specs at submission and send only exceptions to a person. Osmos Creative Review, which is coming soon, is built on that model. Walmart introduced Moderation Assist in March 2026, "a new set of tools available through the Display API for Creative," under its beta terms, while stating that "all ads still go through the standard review process to ensure full compliance with all ad policies" (Walmart Connect, March 2026).

How should a retailer automate reporting?

A retailer automates reporting by tying every ad to the orders it influenced and publishing results to advertisers by SKU, category and keyword, without an analyst building each report. Osmos Product Ads provides order-level reporting with SKU, category and keyword performance, brand halo included, and customizable attribution windows, and Osmos Measurement adds studies such as brand lift for questions that last-click reporting cannot answer.

Advertisers renew budgets faster when they can see which SKUs their ads moved. An advertiser who can see which SKUs an ad moved, at what cost, inside the same portal where the campaign was booked, needs fewer calls with the retailer's account team and makes the next decision faster. Automated reports stay trusted when the attribution window sits next to every figure, one definition of an attributed sale holds across formats, and advertisers see the same live status the retailer's own team sees.

How do you automate catalog feeds, SKU targeting and creative at scale?

You automate catalog-driven retail media by treating the product feed as the ad: titles, images, prices, ratings and stock status flow from the catalog into every Product Ad, so feed quality decides ad quality, and automation should check and correct the feed before it touches bids or creative. Osmos Product Ads auto-selects products from the catalog and pauses out-of-stock items from the auction, which keeps ads in step with the feed without manual edits.

Retail media feed management. The feed for a retail media network is the retailer's own catalog. The work is to keep it complete and current: every SKU with a correct title, a primary image, price and stock, refreshed often enough that an ad never promotes an item the shopper cannot buy. Rules that flag missing attributes at ingestion catch most problems before they reach an auction.

SKU-level targeting. For high-volume SKU campaigns, the platform has to target and report at SKU level. Osmos Product Ads targets on retailer-owned first-party signals (purchase history, browsing, category affinity, geo, store, keywords and key-value pairs) with no third-party cookies, and reports results back to the SKU. Retailers looking for commerce media agents or partners with SKU-level targeting usually want both together. How contextual and behavioral signals combine in that targeting is covered in Contextual vs Behavioral Targeting in Retail Media.

Automating SKU-level content fixes. Because a Product Ad shows what the product page shows, content automation for retail media means fixing the listing: flagging products with missing images, short titles, absent ratings or broken variants, and routing them to the catalog or brand team. A rule that holds a SKU out of ad eligibility until its listing passes a quality check protects both the shopper experience and the advertiser's return, a link explored in Mastering the Shopper Journey.

Product image automation at scale. A retailer can architect product image automation around one master image per SKU held in a digital asset system, with derived sizes generated automatically and validated at ingestion for dimensions, background and text overlays. Product Ads draw on the catalog image directly; display and video formats take derived or brand-supplied assets that pass the same checks.

Creative automation in retail. Creative automation applies to formats that accept uploaded creative: display, video, sponsored brand units and offsite placements. Product Ads have no uploadable creative, so their lever is catalog and product-page quality. On upload formats Walmart reported that its Automated Creative Generator cut creative production time by about 80%, according to Digital Commerce 360 (January 2026). A retailer that offers several onsite formats should therefore run a feed-quality pipeline for Product Ads and an asset generation and review pipeline for everything else.

How do budget pacing and anomaly detection automation protect advertiser spend?

Budget pacing automation spreads an advertiser's spend across the day and the flight so campaigns spend evenly through the day and finish the flight on budget, and anomaly detection flags the moments when spend runs ahead or falls behind. On Osmos, Campaign Manager gives one pacing view across every channel, with auto-adjustments that keep delivery on track, and surfaces exceptions such as out-of-stock products, rejected creative and pacing misses; routine exceptions follow preset resolution paths, and complex ones reach ad operations with the context needed to fix them. That exception handling runs on built-in rules, which makes its behavior predictable.

Osmos Wallets & Billing handles the money side of pacing. It auto-debits brand wallets from escrow before they reach zero, so campaigns keep running without a top-up, and it triggers incentives by rule. Advertisers see fewer campaigns paused for reasons unrelated to performance, and spend that would have been lost to an empty wallet stays in the retailer's auction.

Should a retailer build or buy its retail media automation?

Most retailers buy the automation layers and build only what makes their network distinctive, such as custom formats, house rules and pricing policy. Building an auction, self-serve portal, billing and reporting in house takes engineering capacity that few retailers want to divert from the core store, while licensing a platform gets advertisers live inside one budget cycle. The full cost and timeline comparison is in the build vs buy guide for the retail media ad tech stack, and the path marketplaces take is covered in How Marketplaces Build Their Own Ad Platform.

Osmos is built for the buy-and-extend path: its modular apps start one at a time over the retailer's existing stack, so a retailer can pilot one workflow, such as onboarding or self-serve buying, measure the hours returned, and add the next. Retailers that run several banners face the extra question of whether to run one auction or several, set out in the multi-banner retail media network guide.

What is the implementation checklist for retail media automation in 2026?

Osmos recommends this order for automating a retail media network, because each step depends on the one before it:

1. Clean the catalog feed. Fix titles, images, prices and stock status, and set eligibility rules so incomplete listings stay out of the auction.
2. Set auction rules. Choose how relevance and bid combine, set a floor for each slot, and switch on brand distancing.
3. Move premium inventory into auctions. Turn negotiated placements such as top keywords into packages with a base price and bidding windows.
4. Offer goal-based bidding and forecasting. Give advertisers target-based bidding, refreshed bid suggestions and a forecast before launch, and explain the learning phase up front.
5. Open self-serve for the torso and tail. Put onboarding, briefs and billing in one portal so new advertisers can launch without an account manager.
6. Automate creative checks for upload formats. Score assets against specs at submission and review exceptions only.
7. Publish reporting by SKU. Show attributed orders, windows and spend in the same portal where campaigns are booked.
8. Automate pacing, wallets and exceptions. Keep spend on plan, wallets funded and routine exceptions resolved without a ticket.

Frequently asked questions

Is a retail media network the same as a media buying platform?

A retail media network and a media buying platform serve different purposes. A media buying platform, such as a demand-side platform, is a tool advertisers and agencies use to buy ads across many publishers, while a retail media network is the retailer's own ad business: it owns the inventory, the first-party data and the auction, and advertisers buy into it. The two connect when a retailer extends its audiences offsite, where the retailer's network supplies the targeting and buying platforms supply reach; Expanding Retail Media Beyond Onsite Into Offsite covers how that works.

What does the MRC auction transparency standard ask of a retail media network?

The Media Rating Council's Digital Advertising Auction Transparency Standards, finalized in January 2026, cover retail inventory and call for disclosure of the "use of reserve prices and methods to update them over time" and of how, "after the winners of an auction have been determined," the prices winners pay are set (MediaPost, 29 January 2026). A retailer with automated floors should be able to document how they move and why.

Can a retailer run CPC auctions and fixed-price placements at the same time?

Most mature networks run both pricing models at once. Product listings across search and category pages run on cost-per-click auctions, while premium positions such as homepage takeovers or top keywords run as time-boxed packages with a base price. Both can run in one system: when packages are themselves auctioned, as with Osmos Ad Packages, the retailer gets competitive pricing on both.

Can a smaller retailer automate its network without a data science team?

A smaller retailer can automate its network without a data science team, because the automation that matters most at small scale, such as dynamic floors, goal-based bidding, self-serve onboarding, catalog-driven ads and automatic pausing of out-of-stock products, comes built into the platform. The retailer's work is policy: which formats to offer, what floors and rules to set, and which advertisers to invite first.

Sources

1. Amazon, Amazon's response to the FTC lawsuit on Sponsored Products ads, 31 August 2026.
2. Instacart Ads, Sponsored Product bidding (optimized bidding) documentation, accessed October 2026.
3. Walmart Connect, Sponsored Search product release notes, July and August 2026 entries.
4. Walmart Connect, March 2026 release notes (Moderation Assist), March 2026.
5. Digital Commerce 360, Walmart's Marty agent and AI tools for retail media advertisers, 9 January 2026.
6. MediaPost, Going, Going, Gone: MRC Finalizes Ad Auction Standards, 29 January 2026.
7. Topsort, "Auctions" (knowledge base), docs.topsort.com/knowledge-base/ad-server/auctions, 9 February 2025; customer case studies accessed 12 October 2026.
8. Kevel, "Customers," kevel.com/customers, accessed 12 October 2026.

Ready to automate your network's auction and operations? Get a demo of Osmos.