The End of the Ad: What Meta's Hatch Agent Means for DTC Brands

Meta is building something it's calling Hatch, an agentic shopping tool embedded inside Instagram. The way it reportedly works: a user expresses intent (typed, spoken, or inferred), and the agent browses, compares, and purchases on their behalf. No manual scrolling. No ad clicks. The consumer just says what they want, and the agent finds it.

The implications for DTC brands aren't just about a new ad format. They're about a fundamental shift in what determines who gets found, who gets bought, and who gets bypassed entirely.

Most of the commentary around AI shopping agents focuses on the user experience. That's the wrong lens. The right question is: what inputs does the agent use to decide which brand to surface?

How Will Targeting Disappear While Purchase Signals Take Over?

For the last decade, winning on Meta meant winning the targeting game, finding the right interest cluster, lookalike seed, or behavioral signal to put your creative in front of the right person at the right time. The advertiser controlled the match.

In an agentic commerce world, that match-making shifts. The agent draws on catalog quality, purchase history, fulfillment reliability, review velocity, and real purchase signals, not behavioral demographics. The advertiser doesn't target. The agent evaluates.

This is why Meta's framing of "catalog quality and purchase signals beat ad targeting" is both honest and alarming. It's honest because it reflects how LLM-driven agents actually make decisions. It's alarming because most DTC brands have spent the last decade optimizing for targeting while neglecting the product data layer that agents will actually read.

39% decline in Meta acquisition ROAS across 230+ brands tracked in Statlas, from February 2024 to April 2026. Targeting efficiency has been eroding, long before Hatch launches.

The numbers already show the trajectory. Across the 230+ brands CTC tracks through Statlas, Meta acquisition ROAS dropped from 2.57x in early 2024 to 1.56x in April 2026, a 39% decline in two years. Customer acquisition cost climbed from $40 to $53 over the same period. Ad spend on Meta grew 67%, but revenue grew only 31%. The gap between spend growth and outcome growth is widening, not closing.

Hatch isn't the cause of this erosion. But it will accelerate it for brands that don't adapt, and create a step-change advantage for those that do.

What Agents Actually Evaluate in Shopping Decisions?

If you want to understand how an AI shopping agent ranks and selects products, think about how a very good personal shopper makes decisions. They care about:

• Specificity of fit: Does the product description match what the person actually asked for? Vague catalog copy, "comfortable, stylish, versatile", is useless. Structured attributes (materials, sizing, use case, occasion) are what the agent can parse and match.
• Social proof density: Not just review count, but review recency, sentiment specificity, and verified purchase rate.
• Purchase signal depth: Has this product been bought by people with similar intent history? Repeat purchase rates, LTV patterns, and real transaction signals, not just ad clicks.
• Fulfillment reliability: Agents representing consumers will weight delivery confidence heavily. A brand with 98% on-time delivery is a safer recommendation than one with cheaper CPCs.

Notice what's not on that list: your ad creative, your lookalike audiences, your campaign structure. Those inputs don't survive the translation to agentic commerce. The data layer underneath, your catalog, your customer data, your fulfillment track record, does.

Why the Statlas Data Advantage Matters in an Agent-Driven World?

Here's the strategic framing that matters: when the discovery layer shifts from targeting to evaluation, the brands with the richest, most structured purchase signal data win.

This is precisely why CTC built Statlas, a cross-brand intelligence layer across 230+ brands that captures real purchase behavior, not just ad platform metrics. In a world where Meta's algorithm decides which products an agent recommends, the brands that have iROAS-level clarity on which products drive the highest-quality customers are not just better at advertising. They're better positioned to be surfaced by agents, because their catalog is backed by the kind of verified purchase signals that agents trust.

The brands that win in an agentic world aren't the ones with the best ad creative. They're the ones whose product data tells the clearest story about who buys and why.

What Should DTC Brands Do Before Q3?

You can't predict exactly how Hatch will behave, Meta hasn't published the agent's ranking criteria. But you can build toward the data layer that any credible AI shopping system will value.

1. Audit your catalog for structured specificity

Go through your top 20 SKUs and ask: could an AI agent accurately match this product to a specific intent statement? If your copy relies on brand voice and lifestyle language rather than structured attributes, you have a gap. Build product detail pages that answer the questions agents will ask: dimensions, materials, use cases, compatibility, repeat purchase rate.

2. Get iROAS clarity by product

Most brands know their blended ROAS. Almost none know which individual SKUs drive the highest incremental revenue per marketing dollar. In an agent world, you want to concentrate catalog optimization energy on the products that generate real purchase signals, not just clicks. iROAS-by-product is the data layer that tells you where to invest.

3. Treat your purchase history as a first-party signal asset

The brands that will have the most leverage in agent-driven commerce are the ones whose past customers have given them the richest signal about who buys what and why. That means actively building first-party data infrastructure now, email, SMS, post-purchase surveys, loyalty programs. Not for retargeting. For the data layer that agents will eventually query.

Is This About Giving Up on Ads?

Hatch, if it launches as described, doesn't eliminate advertising on Meta. It changes the model. Paid media may shift from direct-response clicks to catalog promotion, paying to be included in the consideration set that the agent draws from, rather than paying to interrupt a user mid-scroll.

Brands that treat this as a death knell for Meta advertising are overcorrecting. Brands that treat it as "nothing changes" are sleepwalking. The right response is to build the data layer that makes you a credible recommendation, whether the buyer finds you through an ad, an agent, or organic discovery.

The 39% decline in Meta acquisition ROAS we're already seeing across our network tells us the targeting-first era is winding down. Hatch is an acceleration of a trend that's already in the data. The window to build the catalog and purchase signal infrastructure that the next era requires is open right now, and it's smaller than most founders realize.

How Does CTC's Framework Apply to Agent Commerce?

CTC's methodology for Meta advertising is built on a foundational insight that directly maps to the Hatch challenge: Meta moved from an audience-first world to a creative-first world. The platform's Andromeda retrieval engine evaluates creative attributes and behavioral signals to determine who sees your ads. Your targeting choices are no longer the primary variable. What the system knows about your buyers is.

In the Hatch era, that principle extends one level further. Hatch does not select products based on who you told Meta to target. It selects based on the signals your catalog sends: product data completeness, pricing accuracy, purchase signals from real buyers, inventory availability. The brands with the strongest purchase signal infrastructure, not the best targeting, will win the agent-driven commerce layer.

CTC's signal quality checklist applied to Hatch readiness: CAPI integration transmitting server-side purchase events with high EMQ scores; product catalog data complete, accurate, and rich with use-case attributes; customer purchase signals properly segmented by new vs. existing buyer; and tROAS targets set against incrementality-adjusted benchmarks, not platform-reported figures. These are not new requirements. They are the same requirements that make Meta work, applied to the next surface where intent-to-purchase will be resolved.

Scale Your Agent-Ready Commerce Infrastructure

The shift to agent-driven commerce rewards brands with superior data infrastructure and purchase signal quality. When discovery moves from targeting to evaluation, your catalog becomes your competitive advantage.

Talk to Us