← All editions

The Commerce Shift · Week of August 20–26, 2026

The two-P&L retailer arrives

This week: Walmart and Target fund their raised guides with ads, marketplaces and membership; ChatGPT ads go continental with Criteo as the pipe; the first AI-native holiday takes shape; and Shein's IPO reprices cross-border commerce.

Retail

Walmart's quarter was really two quarters: soft stores, a roaring digital P&L

The Signal

Walmart's Q2 FY27 (Aug 20) delivered $187.9B in revenue (+5.9%), adjusted EPS of $0.81 vs. $0.73 consensus, and a raised full-year guide — but US comps grew just 2.6%, the slowest in six years, while global e-commerce rose 23%, US marketplace net sales jumped 52%, and global advertising grew 38% (Walmart Connect +43% ex-Vizio). Management said roughly half of incremental profit (excluding claims) came from advertising, membership and marketplace — effectively a two-P&L company; CEO John Furner added that Sparky AI-assistant users are up 70% and spend 40% more per order.

Why It Matters

The largest US retailer just made explicit what the industry has been circling for years: the store is the traffic engine, but ads, marketplace commissions, fulfillment fees and membership are the earnings engine. Every RMN gets benchmarked against 38% ad growth, every marketplace team gets asked why they aren't growing 52%, and every CPG funding Walmart Connect is now paying into a media business growing an order of magnitude faster than their own categories.

The Opportunity

Brands should renegotiate joint business plans with the two-P&L reality in view — Walmart needs your media and marketplace dollars to make its guide, which is leverage. Sellers should treat Walmart Marketplace plus WFS (already nearly half of marketplace volume) as the fastest-scaling Amazon alternative and get positioned before Q4 auction density catches up. And with Sparky moving real basket size, product content optimized for Walmart's AI surfaces is now a merchandising job, not an innovation project.

Beth's Take

Stop reading Walmart as a grocer with a side business. A 2.6% comp funding a 38-52% digital flywheel is the clearest statement yet that the scaled-retail operating model has flipped: the stores buy the traffic, and advertising, membership and marketplace convert it into earnings. Every large retailer will be forced toward this model — the only question is how quickly.

The leverage this creates is underpriced. When half of a retailer's incremental profit rides on vendor media budgets and seller fees, the vendors and sellers writing those checks have more negotiating room than most of them use — expect the sophisticated ones to start trading media commitments for margin relief, data access and shelf outcomes. And the number to watch from here is the comp itself: if store traffic softens further, the whole flywheel gets re-rated, because every high-margin business in it sits downstream of footfall.

Coverage

Commerce media

Target's turnaround is traffic-led — and Roundel and AI referrals are doing heavy lifting

The Signal

Target posted Q2 comps of +3.8% against a 2.4% consensus on a 3.6% traffic gain, with digital comps up 8.7% and same-day delivery up more than 25%; EPS of $4.11 doubled year over year (helped substantially by tariff refunds) and FY EPS guidance leapt to $9.90-$10.90 from $7.50-$8.50. Roundel ad revenue grew 28.6% to $279M, and CEO Michael Fiddelke said traffic sourced from external AI platforms is growing more than 3.5x faster than the industry, crediting agentic commerce partnerships including OpenAI; the company's first chief AI officer started Aug 24.

Why It Matters

After years of ceding share to Walmart, Target's inflection is coming from visits — not ticket — plus the same high-margin ads-and-marketplace layer powering Walmart's guide. The AI-referral stat is arguably the bigger tell: a top-five US retailer is now citing external AI platforms as a measurable, outperforming traffic source on an earnings call, which makes agentic referral share a KPI every retail board will ask about next quarter.

The Opportunity

CPGs and agencies should re-weight Roundel in holiday plans while its momentum gives Target room to invest in the network — and press for the measurement the moment demands. Retail leaders should instrument AI-sourced traffic now (referral tagging, agent-readable product data, partnerships with assistant platforms) so they can report — and grow — the number Target just normalized.

Beth's Take

Two consecutive quarters of traffic growth is a trend, not a bounce, and the composition matters more than the beat: visits up 3.6%, Roundel growing 28.6%, and management volunteering that AI-platform referrals are outgrowing the industry by 3.5x. The operators' question is no longer whether Target recovers — it's whether your holiday plan is positioned for a Target that wins visits and monetizes them twice, once at the shelf and once through Roundel.

The AI-referral stat is the one with legs. A top-five retailer just told Wall Street that external AI assistants are a measurable, outperforming acquisition channel — which makes agentic referral share a board-level KPI across retail within a couple of quarters, and hands the assistant platforms the beginnings of pricing power over retail traffic. For brands, the practical read is simpler: a network with this much momentum heading into holiday is usually the one investing hardest in capabilities and courting demand — that is when to press for the measurement you actually want.

Coverage

AI

ChatGPT ads go continental: 31 European markets, with Criteo as the commerce on-ramp

The Signal

OpenAI expanded ChatGPT advertising to 31 European markets starting Aug 24 — bringing the format to 40 markets roughly six months after its US debut — with ads shown only to Free and Go tier users and no advertiser access to chat histories. Criteo, ChatGPT Ads' first ad-tech partner, simultaneously extended its integration to all 31 markets, saying 2,000+ brands already buy ChatGPT ads through its platform, that Prompt Smart Ads drove roughly 4x higher spend after activation, and that more than 80% of ad-driven ChatGPT traffic comes from new customers.

Why It Matters

Conversational AI is now a scaled, monetized commerce-discovery channel on two continents — and a commerce-media incumbent has installed itself as the demand pipe. Criteo's new-customer stat is aimed straight at the incrementality question CPGs keep hammering retail media networks on, which means the high-intent shopper moment RMNs claim as their moat now has a credible competitor sitting upstream of the retailer entirely.

The Opportunity

Brands with European exposure need a ChatGPT media test in-market this quarter, while auctions are thin and pricing favors early movers; the Criteo pipe is the lowest-friction entry. Retail media leaders should sharpen their incrementality story now — 'new customer' proof is about to become the comparison point buyers bring to every RMN negotiation.

Beth's Take

The interesting move here isn't OpenAI's — it's Criteo's. The company that built its business on retailer relationships just positioned itself as the arms dealer for the channel most likely to sit between retailers and their shoppers, and its pitch — more than 80% of ad-driven ChatGPT traffic arriving as new customers — lands squarely on the incrementality nerve CPGs keep pressing with retail media networks.

Two things follow. Retail media sellers should assume the high-intent-discovery story is no longer exclusively theirs and get their own incrementality evidence in order before Q4 negotiations — that comparison is now coming to every renewal conversation. And watch whether other RMN-adjacent ad-tech players follow Criteo upstream: if they do, the conversational layer becomes the new front of the commerce-media stack rather than a curiosity at the edge of it.

Coverage

Consumer behavior

The first AI-native holiday: 65% of shoppers plan to use AI, 8% of retailers feel ready

The Signal

Narvar's 2026 Holiday Shopping Report (Aug 24) found 65% of US consumers plan to use AI for at least part of their holiday shopping — 43% for gift discovery, 33% for product comparison — while only 8% of surveyed retailers call themselves very confident using AI in the shopping experience. The behavior is already showing up in platform data: Shopify reports AI-referred sessions to merchant storefronts up 197% year over year in Q2, converting at roughly twice the organic rate in research-heavy categories, with more than half of AI-referred sessions landing directly on product pages.

Why It Matters

Consumer intent and merchant-level traffic data now tell the same story: AI-assisted shopping is a distinct, high-intent, fast-compounding acquisition channel with its own funnel shape — pre-qualified visitors skipping the homepage and landing on product pages. The 65%-vs-8% gap means that demand will route to whoever is prepared to catch it this Q4, including third-party assistants rather than retailers' own properties.

The Opportunity

Before holiday code-freeze: instrument AI referral traffic separately, harden product detail pages as landing pages (they are now the front door for half of AI-referred sessions), and get product data agent-readable. Shopify's data also shows organic search still leads total volume — this is an additive channel to fund, not a reason to gut SEO.

Beth's Take

Every 'first holiday season' for a channel — mobile, social — rewarded the retailers who instrumented it before it was material and punished the ones who filed it under next year. The 65%-versus-8% gap says most of the industry is choosing the second camp by default.

The Shopify numbers make the cost of that choice concrete: AI-referred sessions convert at roughly double the organic rate in research-heavy categories and land on product pages more than half the time, which means this funnel is short, product-data-dependent, and largely invisible to teams still reporting traffic the old way. Ten weeks out from peak, the achievable version of readiness isn't an AI strategy — it's tagged referral traffic, product pages treated as landing pages, and clean, agent-readable product data. That's a sprint, not a transformation.

Coverage

Marketplaces

Shein's Hong Kong IPO prices the post-de-minimis era: ~$27B, down ~70% from peak

The Signal

Shein opened the book on its long-delayed listing, offering 280 million shares at HK$47.60-49.50 to raise up to ~$1.8B at a valuation near $27 billion — roughly 70% below the ~$100 billion private valuation it reached in 2022 — with pricing set for Aug 31 and trading expected Sept 1 after failed attempts in New York and London. Revenue growth decelerated to 8% in 2025 from 20.7%, and the company posted a $99M loss in early 2026 driven partly by the end of the US de minimis exemption.

Why It Matters

This is the first public-market price on the direct-from-China cross-border model after the tariff and de minimis regime change — and the market says the pioneer is worth about a quarter of its peak. A completed listing also puts Shein under continuous disclosure, making its marketplace and logistics pivot visible to every competitor, regulator and brand deciding how seriously to take the next phase of cross-border commerce.

The Opportunity

Retailers and brands that spent three years losing price-sensitive share to Shein and Temu should read the S-curve honestly: the ultra-cheap import machine has structurally higher costs now, which reopens the value battleground for domestic players — the same fight Walmart and the clubs are visibly winning. Watch the Sept 1 debut and early disclosures for how aggressively Shein shifts toward local marketplaces and third-party sellers.

Beth's Take

A markdown of roughly 70% on the category-defining fast-fashion machine is the market grading the de minimis era's end, in public. The growth math tells the story more honestly than any policy debate: 20.7% growth decelerating to 8% in a year, and a $99 million loss sitting where the duty exemption used to be.

The model isn't dead — but its structural cost advantage is, and that reopens the value battleground for domestic players at exactly the moment Walmart's and the clubs' results show them winning it. Once Shein trades, its disclosures become the industry's quarterly window into cross-border unit economics; the question for everyone who lost share to the parcel flood is whether Shein's pivot toward local marketplaces and third-party sellers arrives fast enough to matter — and operators who priced their 2027 plans assuming an endless flow of duty-free parcels should revisit them now, not after the debut.

Coverage

Quick hits

  • Walmart ended its decade-long NFC holdout and will accept Apple Pay and Google Pay at all US stores by year-end — a capitulation that says checkout friction finally beat wallet-lock-in economics even at Walmart scale. TechCrunch ↗
  • Dick's Sporting Goods fell 31% — its worst day in three years — as Foot Locker's -3.6% comps dragged down the first combined quarter and management slashed guidance, flashing a warning for athletic footwear and mall retail. The Motley Fool ↗
  • The Trade Desk filed a mixed shelf while trading near 52-week lows, down ~65% YTD — the independent buy-side counterweight to Amazon and Google is signaling it needs financial flexibility. Benzinga ↗
  • Consumer confidence slipped to 89.4, its weakest since January, with the Expectations Index at 68.2 — well below the recession-signal threshold — even as present-situation readings improved; shoppers are spending but primed to pull back. The Conference Board ↗
  • Perion is buying in-store retail media firm PRN for up to $12M — ad tech consolidation is now chasing the in-store screen layer, where 80%+ of US retail commerce still happens. SalesTechStar ↗

Get The Commerce Shift

One email a week, when the edition publishes. No noise in between.

The Commerce Shift is compiled weekly from public reporting; every link goes to the original source. Nothing here draws on non-public information.