Retail Media Intel

Executive summary

Retail media is consolidating fast around full-funnel, cross-market scale. ShopX unified Shopfully, Flipp and MEDIA CENTRAL into a 27-market group reaching 400 million shoppers to aggregate multi-retailer inventory and connect digital search signals to physical checkouts. At the same time Affle 3i agreed to buy Perion Network for $420 million in cash at $11.70 per share, a 34.5% premium, pairing Affle mobile CPCU strength in emerging markets with Perion strength in Retail Media, CTV and DOOH to sell CTV-to-Cart and DOOH-to-Cart across 100+ countries. Specialty retail is conceding build-it-yourself: Criteo is taking over Notino retail media tech in European beauty. Commerce is being rewritten by Google from both ends of the funnel. In the US, Universal Cart is now widely visible in Search product panels with Add to cart and Buy now under Visit site plus a header cart icon, with transactions run via Google Pay while retailers including Nike, Sephora, Target, Ulta Beauty, Walmart, Wayfair and Shopify see nothing until cart transfer and ad-click to Google-checkout attribution remains undocumented. In India, Google Pay is piloting Discover Shop, an AI-personalized fashion feed over a 60 billion-product Shopping Graph with virtual try-on, visual search and YouTube Shorts, but redirecting to merchant sites for checkout — a deliberate contrast: take checkout in the US, take discovery without checkout in India to defend engagement against PhonePe. The cross-domain connection is that discovery and checkout are collapsing into one surface, whether that surface is CTV, search, payments or store. Affle-Perion pitch is CTV-to-cart; ShopX pitch is search-to-store-checkout; Google pitch is AI-answer-to-Google-cart. Underpinning all three is the product feed: Google says AI queries are now 3x longer with one in six multimodal, ads in AI surfaces match to both query and AI response, and retailers need six optional product data fields. Pinterest Visual Language Models, Instagram Shops catalog tagging and Google Lens AI Mode visual fan-out all point to the same conclusion: keyword coverage is no longer enough. My take: Google is winning the commerce capture while retail media aggregators race to defend store-led budgets before Google owns the transaction. Perion fall from a $2 billion peak to a $420 million sale, net only ~$192 million ex-$228.3 million cash on ~$400 million revenue after Microsoft Bing policy changes, is a warning on platform dependency. Trade Desk cutting ~15% of staff with consensus expecting revenue to decline 1.9% annually to ~$2.8B by 2029 shows open-web programmatic is paying for this shift. First week signal is clear: own closed-loop proof from impression to cart, or be intermediated.

Market dynamics

Consolidation is now about combining upper-funnel reach with lower-funnel proof. Affle brings zero-debt Perion with ~$228 million cash, ~600 developed-market employees, Perion One and Outmax AI agent, financed via Citibank debt plus cash with no financing condition, targeting net debt-free post-close and $700M+ FY28 revenue run-rate with EPS accretion in one year. ShopX centralizes technology, measurement and cross-border audience data while leaving commercial operations local — the aggregator playbook to unlock in-store budgets single-retailer networks cannot. Both mirror Criteo-Notino outsourcing: scale wins on tech and measurement, not on retailer logos. Tech shift is AI + visual + checkout compression. Google is expanding ads in AI Mode to 10 new countries including Australia, Canada, India and Nigeria, with Brazil, Mexico and UK planned for Q4 2026, plus multiturn ads for follow-ups in the US, with Search advertisers auto-opted in and no standalone opt-out. It is testing grouping paid products under one Sponsored heading and unpaid under All products, with paid first and blue titles as the only per-card marker, normalizing blended paid/free carousels. Meanwhile measurement is degrading: Gemini briefly tagged cited links with utm_source=gemini in September 2026 then removed it overnight Sept 29, web clicks still pass gemini.google.com referrer but iOS app clicks strip it and land in Direct in GA4. Centralized retail media measurement promises and Google checkout opacity are on a collision course. Budget flow favors closed-loop commerce outcomes over open-web reach. Soft auto and CPG budgets plus competition are blamed for Trade Desk pressure even as it migrates to owned data centers and pushes Kokai and Koa Assistant. CPG response is to in-house commerce media talent, e.g. Reckitt hiring Ankit Kishore from Zenith/Carat/GroupM with Nestle/Pepsi/Philips e-commerce media experience. Fluent Q3 results Nov 12 will be a small-cap read on first-party-data + ML commerce-media demand across Fluent, AdParlor and Call Solutions.

Key developments

Strategic implications

If you build retail media or commerce products, prioritize closed-loop proof over reach. ShopX and Affle-Perion both sell on connecting an upper-funnel signal — search, CTV, DOOH — to a cart or store till with measurement. Google Universal Cart and AI Mode show the bar: one-tap purchase from discovery with managed payment and orders. Your roadmap needs CTV-to-cart and search-to-store attribution, multi-retailer basket resolution, and log-level AI referrer handling to survive Gemini UTM stripping and iOS Direct bucketing, not just another onsite sponsored-product bidder. Second, treat the product feed and visual assets as shared infrastructure across both domains. The same enriched title, image, variant, availability and lifestyle data that wins Instagram tagging and Pinterest Product Pins now determines Google AI Mode matching, blended Sponsored vs All products placement, and virtual try-on eligibility in GPay Discover. Invest in feed enrichment for Google six optional fields, high-resolution multi-angle imagery for visual search, and daily inventory accuracy for 50-60 billion-listing scale matching. The Criteo-Notino outsourcing signal says mid-tier retailers will buy this rather than build — sell them turnkey feed-to-visual-to-measurement, not point tools. Third, design for two checkouts: Google-owned and merchant-owned. In the US Search flow Google owns payment and order until transfer; in India GPay flow merchant owns transaction. Build order ingestion, identity stitching and incrementality measurement that works in both modes, with consent-first personalization controls like GPay shopping settings. If you cannot tell a brand what Universal Cart stole versus created, you will lose budget to whoever can.

Competitive landscape

Google is winning across both domains by owning discovery, AI answers and increasingly checkout, from 60 billion-listing Shopping Graph to Universal Cart to non-optional AI Mode ads. Affle is bidding to be the full-funnel challenger if it closes Perion in early 2027, survives the 40-day go-shop through Dec 11 and shareholder/US antitrust approvals, and delivers $700M+ FY28 scale. ShopX is the in-store aggregator to watch at 400M shoppers/27 markets, Criteo the infrastructure consolidator for specialty verticals like Notino beauty, while Trade Desk and standalone search-dependent players like legacy Perion are losing to walled gardens and commerce-closed-loop.

Action items

Trend lines

Contradictions

• Google advances on-platform checkout via Search Add to Cart while scaling off-platform discovery in Google Pay India that redirects to merchants and YouTube, simultaneously coupling and decoupling transaction.
• AI search promises 3x longer higher-intent queries but consolidates Sponsored presentation and auto-enrolls advertisers, reducing transparency and control as relevance rises.
• Global retail media pitches unified cross-market measurement via ShopX and Affle-Perion in the same week Gemini UTM stripping weakens independent third-party attribution.
• Outcome-based CPCU and CTV-to-cart growth is framed as efficiency gain while Trade Desk cost cuts and selloff suggest efficiency is being forced by margin pressure on independent intermediaries.

Blind Spots

• No detail on ShopX in-store measurement methodology, retailer POS integration, or shopper consent across 27 markets.
• Missing economics of Google Universal Cart and Affle CPCU: take rates, return attribution, Google Pay payment flow, and merchant margin impact.
• Absent Amazon, Walmart Connect, TikTok Shop and Meta responses to Google and Affle-ShopX moves, leaving competitive reaction unmapped.
• No regulatory view on AI Mode auto opt-in, cross-border shopper data pooling, or virtual try-on biometric privacy in India test.

Connections

• Google Pay fashion feeds in India act as visual-query training data for longer multimodal Search ads and Shopping Graph, linking payments discovery to search monetization.
• ShopX in-store audiences plus Affle-Perion CTV-offsite enable a living-room-to-aisle loop where CTV creates demand and in-store media closes it under one measurement plan.
• UTM loss plus AI Mode auto opt-in pushes brands toward Google-owned checkout to recover attribution, turning a measurement constraint into a commerce conversion driver.
• Pinterest and Instagram image tagging plus Google Lens normalizes merchant-redirect shoppable while Google Search Buy Now normalizes on-platform checkout, forcing brands to run dual fulfillment logic.

Article index