We audit a lot of accounts. The same seven formats keep showing up as top performers.

Everything else either consolidates into one of these buckets or quietly dies inside the 30-day fatigue cliff. Most brands don't notice. They keep paying to refresh ads that the algorithm already wrote off two weeks earlier.

What's interesting isn't that the list is short. It's that it's the same list regardless of category. Food and bev, apparel, supplements, CPG, home goods. The product changes. The format library doesn't.

Which means the question facing every brand right now isn't what to ship next. It's how many of these seven you're actually running today. And how many you've left as pure white space inside your account.

Here's the shortlist. The one format on it almost no agency will pitch you. And the brand currently running all seven inside the same library.

Every format below has held alive past the 30-day mark inside accounts spending six and seven figures monthly. That's the only working definition of "working" in 2026.

1. Founder based. Most brands run one version and assume they've covered the format. They haven't. The category contains at least five distinct sub-styles. Ugly founder shot on a phone. Founder story with B-roll. Podcast-setup founder. Warehouse-shot founder. Founder on a call with a customer. Five different entity IDs hiding inside one bucket.

2. Cultural relevance ads. The hook isn't your product. It's whatever's trending across culture right now.

GLP-1s. The seed oil debate. The dopamine-detox movement. The hook is the cultural conversation, and the product is positioned as the answer to whatever downside that conversation creates. Trend-jacking that just references the moment dies fast. Trend-jacking that contrasts the product against the cultural problem scales. The brands winning here aren't borrowing relevance from the moment, they're solving for it.

3. The Love Letter. Low-effort, high-leverage. B-roll in the background, on-screen text doing all the talking, no creator on camera, no voiceover required. The format works because of how Meta reads it.

Long on-screen text forces longer hold times. Readers replay to catch what they missed. Both signals tell Meta this is content, not an ad, and the algorithm scales it accordingly. The lift to produce is almost nothing, a phone camera, some product B-roll, a Notes-app script. The output is one of the highest-scaling formats currently running because you're feeding the system the exact engagement signals it's calibrated to reward.

4. Yapper ads. Fast-talking, high-energy, unscripted creator monologues. The rawness is the product. Scripts kill the format the moment the creator slows down to remember a line. Brief talking points, one take, minimal editing. The energy is what makes the viewer stay.

5. The "Math" framework. Ads that lead directly into the price objection. Comment-replies to "why is it so expensive." Breakdowns of cost-of-goods. DIY-versus-retail math. Counterintuitive, undersupplied, and the most interesting format on this list. More on this below.

6. The Contrarian. Direct contrast ads that position the product against real-world substitutes. Not "we're better than the other guys in our category." Better than the alternatives your audience is actually using when they're not buying you.

For Huel that means contrasting against the rotisserie chicken, the protein bar grabbed at the gas station, the lunch sandwich from the deli. Same protein, half the convenience, none of the micronutrients. The contrast is what makes the buying decision feel obvious. The brands running this well aren't comparing themselves to competitors, they're comparing themselves to the daily compromise their audience already makes.

7. Warehouse POV. Behind-the-scenes content shot from inside operations. Packing orders. Pulling ingredients off the shelf. Showing the production line. Walking the warehouse floor. The format works because it does something polished ads structurally can't, it shows the brand is real. Real building, real people, real product moving.

For brands manufacturing or fulfilling at scale, this format is sitting unused on most accounts because nobody thinks to point the camera at it. It costs nothing to produce. It generates a distinct entity ID inside Andromeda. And it's one of the only formats that builds operational trust without requiring a creator at all.

Here's the part the industry won't say out loud.

Most performance creative agencies pitch "format diversification" like this. One creator shoot day. Four to six variations of the same UGC concept. Delivered as a "creative package." Different headlines, different B-roll, sometimes a different opening line.

Same entity ID inside Andromeda. Same audience pocket. Same auction slot.

That's not diversification. That's volume on a single format, dressed up in a deliverable that looks productive on a creative calendar.

It's also the most common form of waste sitting inside DTC ad accounts right now. Paid for monthly. On retainer. By brands who've never been told the difference.

The seven formats above are what actual diversification looks like. Not because they're trendy. Because each one registers as a structurally distinct fingerprint to Meta's computer vision, and unlocks a separate audience pocket the brand currently isn't competing in.

Seven formats means seven independent learning phases. Seven delivery graphs. Seven auction segments. One format in six outfits means one of all of the above.

If your agency can't tell you how many of these seven formats are currently live in your account this week, you don't have a creative partner. You have a production vendor. Those are different things. They cost the same.

This is the framework we run across our portfolio. Bespoke Post, Barry's Bootcamp, Olyra foods, and more.

1. Map your current library against the seven. Pull every active concept. Tag each one against the format list. Most brands find they're running two formats. Sometimes three. The other four are pure white space. That's where the next dollar goes. Not into more variations of what's already running.

2. Budget by entity ID, not by ad count. Each format registers as its own entity ID inside Andromeda. Seven structurally different concepts give you seven learning phases. Five variations of one founder UGC ad give you one. At $50-100 minimum per ID to exit learning, a seven-format library running two concepts each is a $700-1,400 weekly testing floor before scaling.

3. Use creators as a format multiplier, not a category. One creator roster can shoot founder UGC, yapper ads, contrarian comparison content, and Love Letter narratives in a single day. Brief for format diversity, not hook diversity. The math favors creators every time when the brief is structured this way.

4. Ship "We're Not Cheap" before your competitors do. This format has the lowest current competitive density in almost every DTC category. Because most brands won't produce it. If you're willing to put real cost math on camera, you scale on a format that's structurally uncrowded inside the auction.

5. Run the system on a 30-day refresh cycle. Each entity ID needs roughly 30 days to fully prove or disprove itself inside Andromeda. Build the calendar around format coverage and refresh rhythm. Not around individual ad output. The system is the asset. The ads are just what falls out of it.

The seven formats above aren't a forecast. They're an inventory.

What's currently surviving, inside accounts spending real money, in May 2026.

Brands building production systems against the full list are already pulling away. Competitors still iterating on the same two formats they ran in 2024 are still iterating. The gap compounds monthly.

Huel figured this out early. The spread isn't a creative accident. It's the architecture of a brand that understood the new auction before most of the category did.

If you read this and wondered which of the seven your own account is actually running, that's the right instinct.

We're doing a limited number of free creative format audits this month. For brands spending $50K or more monthly on Meta.

We'll pull your library. Tag every active concept against the seven formats. Identify exactly which entity ID pockets you're not showing up in. Send you a one-page production priority list of what to ship next.

Three audit slots are open, first come, first served.

If this issue was useful, forward it to an operator who needs it.

Until next time
✌️, Conrad