• The DTC Times
  • Posts
  • What Meta Told 1% of Advertisers Behind Closed Doors

What Meta Told 1% of Advertisers Behind Closed Doors

What you need to know from the 2025 Performance Summit: AI agents, VBO, and the end of ROAS as a metric

The 2025 Meta Performance Summit was this week, and we’ve got the goods for you…

If you’re still optimizing for ROAS, chasing attribution hacks, or running the same old ASC campaigns, you’re flying blind in a system that’s already moved on.

Let’s go deep on what actually matters from the Summit.

Grab a cup of coffee…☕️

Here’s what’s inside:

📉 MacroMeta Just Declared War on ROAS
Why attribution-first strategies are going away, and what rising CPMs + stagnating inventory mean for performance marketers

📈 TrendsCreators. Custom Conversions. Controlled Chaos.
How brands are winning with better signals, modular UGC, and AI-personalized ad delivery

🛠 TacticsVBO. Incremental Delivery. Profit Optimization.
The new Meta stack explained, from GPT targeting to AI agents that handle objections before checkout

📢 Quick Hitsfunnel analyzers, product seeding hacks and how to target boomers with your ads. 

🚨 Scaling on Google in 2025? Don’t Wing It.

Many brands have given up on leveraging Google SEM.

On June 10, join the teams behind Obvi and DigiCom as they break down the exact playbook that makes profitable Google Ads possible in 2025:

  • The branded vs non-branded strategy shift

  • The campaign structure keeping ROAS alive

  • Why most brands fail to scale (and how to fix it)

🎯 Featuring Ash Melwani (Obvi CMO) + the DigiCom growth braintrust
🎥 100% tactical. Live Q&A + account audit included.

📅 Save your seat here → lu.ma/2zts7xw0

Macro — Meta Just Declared War on ROAS

The Old Meta Is Dead. ROAS Worship Died With It.

Meta didn’t just roll out new ad tools last week. They issued a manifesto: incrementality > attribution. Profit > ROAS.

For years, marketers were rewarded for optimizing toward what the system could measure, not what actually grew the business.

  • ROAS looked strong, even if those purchases would’ve happened anyway

  • ASC campaigns recycled the same audience, but still got credit for conversion

  • Spray-and-pray testing inflated "performance" without driving lift

  • Cost caps and last-click metrics made results look good on paper, while masking stagnation

It wasn’t that people were doing things wrong, it’s that the system itself rewarded proxy metrics instead of actual business outcomes.

But now the rules are changing. And here’s why that matters:

Meta’s inventory growth is slowing.

Platform growth in North America and Europe has flattened. At the same time, advertiser demand keeps rising. More brands. Bigger budgets. More competition.

Result? CPMs are increasing — even if your ads aren’t performing any better.

The supply-demand imbalance is real, and Meta knows it.

That’s why the company is now retooling its entire performance ecosystem around incrementality, profit, and business-level outcomes, not recycled attribution.

Meta isn’t hiding it anymore: they want brands focused on cleaner data, higher-value acquisition, and long-term business impact. 

Also worth noting: AI isn’t just a buzzword here.

It’s the backbone of Meta’s new ad system. From delivery optimization to bid modeling, AI is how Meta plans to stretch increasingly limited inventory into scalable, profitable growth.

Every shift — from value-based bidding to incremental delivery — is powered by AI decision-making. Meta’s no longer asking, “What ad works best?” It’s asking, “What outcome is most valuable and who is most likely to deliver it?”

🥡 Takeaway — ROAS Is Dead. Profit and Lift Are the New North Star.

If you’re still optimizing for last-click ROAS, you're misaligned with the direction of the entire Meta ecosystem.

The new playbook rewards brands that optimize for profit per impression, not credit per click.

📊 Want proof? Read the full recap from Herd or watch Charley’s breakdown for the full behind-the-scenes strategy shift.

Creators. Custom Conversions. Controlled Chaos.

The most important creative shift coming out of Meta’s 2025 Performance Summit has nothing to do with style and everything to do with signal.

Meta is no longer just asking what your ads say. It wants to know who they’re for, why they work, and how they contribute to profitable growth. That means brands are rethinking:

  • What they feed the algorithm

  • Who delivers the message

  • How creative variation affects targeting and delivery

Let’s break down the 3 big creative trends operators need to act on now:

1. Creators Unlock Creative Diversity (and Preferential Delivery)

“Meta is giving partnership ads preferential treatment. And it’s not subtle.”
Charley The Disruptor

Creator-led ads aren’t new, but what’s changed is how tightly integrated they are into Meta’s delivery engine. Partnership ads (a modern evolution of white-listed collabs) now allow brands to dynamically toggle between:

  • Branded or creator-led delivery

  • Shared or solo handles

  • AI-driven creative selection per user

Meta wants to serve the most resonant message from the most relevant messenger. And it’s building its infrastructure to do just that.

According to Vertigo Agency, partnership ads now reduce CAC by up to 19%, especially when paired with performance-tested creator UGC.

2. Creative Variation + Signal Quality = Scaled Testing

Forget “launch 100 ads a week.” The new best practice is tighter, smarter creative inputs with stronger conversion signals.

Thanks to updates in Meta’s Andromeda system, you can now:

  • Assign custom conversion events to specific SKUs or product categories

  • Optimize for hero products instead of generic purchases

  • Feed Meta real business objectives (e.g. “buyers of bundle X” or “new subscription starts”)

These signals fundamentally change how the algorithm targets, bids, and delivers. As Charley put it:

“You’re no longer just optimizing for a purchase — you’re training the machine on what kind of purchase actually matters.”

3. AI-Assisted Creative Is Getting Smarter — But It Needs Guardrails

Meta’s generative tools for copy and visuals are rapidly improving — and they’re starting to nudge creative elements based on performance context.

At the Summit, Meta previewed how AI will increasingly nudge ad creative in real time:

  • Adjusting visuals and copy for different audience segments

  • Selecting the best-performing creator variant based on user profile

  • Matching tone, length, and format based on funnel stage or conversion likelihood

Meaning you define the brand guardrails and the AI personalizes per user, per format, per funnel stage.

In early tests, brands combining static creative + AI-powered personalization for new audience segments saw a 5–10% CAC improvement, with no loss in brand voice (source).

🥡 Takeaway — Signal Is the New Creative Strategy.
Relying on high-volume creative development may be coming to an end.

Instead, build fewer but better ad units and feed them richer inputs: creators, context, and clear value signals. That’s how you train the algorithm and scale efficiently.

🔎 Explore how Meta's creator partnerships and signal-first creative strategies are reshaping paid social in 2025.

Tactics The New Meta Performance Stack

Welcome to Meta 2.0.

The Meta performance stack has been completely overhauled. Here are the emerging 3 core principles to consider in your Meta playbook moving forward:

  1. Tell Meta what you want (real business goals, not just conversions)

  2. Give it better data to work with (custom events, COGS, LTV)

  3. Let the machine optimize toward profit, not proxy metrics

Below are the tools (and tactics) that enable that shift.

1. Value-Based Optimization (VBO) Isn’t Just for Big Brands Anymore

You can now feed Meta not just your AOV, but your actual COGS and operating costs. The goal is for Meta to optimize for profit volume, not just revenue or purchases.

Think: AOV – CPA = GPT (Gross Profit per Transaction)
And now Meta can bid toward maximizing that number (not just your blended ROAS).

This is a massive unlock for brands with wide product spreads.

Got a $30 SKU and a $200 bundle? You can now tell Meta which outcome matters most. It also opens the door for lead gen brands to optimize for lead quality, not just lead volume.

2. Incremental Attribution Is the Hidden Growth Lever

Forget last-click. Meta’s newest optimization setting — “incremental attribution” — isn’t really an attribution model at all.

It’s a delivery mode.

When you toggle it on, Meta actively shows ads to users who are less likely to convert — and tracks how many new customers you actually drive.

This is Meta’s answer to ad fatigue and audience overlap. Instead of fighting for credit, you're telling the machine: “Find me net-new buyers.”

Adobe reported a 35% lift in new customer acquisition just by switching to this model (source).

3. Lift Testing Is Back (and Easier Than Ever)

Meta is pushing monthly Conversion Lift Studies as a best practice. Why?

Because brands aren’t just asking “What performed?” They’re asking, “What changed behavior?”

Power Lift studies now let you:

  • Measure true incremental performance (not just attributed sales)

  • Compare bidding strategies (e.g., cost caps vs value-based)

  • Validate long-term customer impact

Don’t wait for a rep. If you’re spending over $500/day, push Meta to run a lift study or use your third-party attribution tool to simulate it.

4. Scorecards, Match Rate, and the Return of Clean Data

Charley’s 3-step action plan from the Summit recap:

  1. Define your high-value actions with custom conversion events

  2. Audit your data regularly (match rate >7/10 is the goal)

  3. Fix one major weak spot per month — like feed quality, signal depth, or attribution structure

The Opportunity Score and LTVision tools are rolling out fast, and operators who get ahead here will enjoy lower CPMs and smarter delivery before the herd catches up (Herd recap).

🧠 Bonus: AI Agents Are Coming for Your Checkout Flow

Meta didn’t just show off bidding models — they previewed the future of AI-native ad units.

Imagine this: a user sees an ad, taps it, and instead of landing on a static PDP, they start a conversation. An AI agent (trained on your product catalog) responds:

“Yes, that’s in stock in red, size M. Buy 2, get 1 free today. Want me to add them to your cart?”

Meta demoed this exact interaction at the Summit. These AI-powered chat ads will soon handle objections, suggest upsells, and guide users through the funnel without ever leaving the app.

While still in early testing, this signals a seismic shift: your media won’t just persuade, it will negotiate and convert in real time.

🥡 Takeaway — This Isn’t a UI Update. It’s a Strategy Reset.

Meta’s tools are changing and so will the rules of the game.

The brands getting cheaper CPMs, higher new customer counts, and real profit growth are the ones that’ve retrained the machine around real business outcomes.

🎯 For a tactical walkthrough, watch Charley’s recap or read the full breakdown from Herd.

Quick Hits

Reply

or to participate.