What This Article Covers

The top 10 creative diversity tactics for Meta Andromeda in 2026 are building structural variants instead of hook swaps, mixing formats in prospecting, simplifying account structure, running broader audiences with more creative variety, protecting new creative with dedicated testing budgets, testing messages in statics before producing video, building an evergreen creative library, testing partnership ads, separating new and existing customer campaigns, and diagnosing creative by dimension. Curated by Y'all, this guide covers what each tactic is and how to implement it.

Updated May 2026

At A Glance

‍

If you've been running Meta ads for more than a year or two, you've probably noticed something shift. The campaigns that used to work, the ones built on tight audience targeting and a handful of polished video ads, just don't produce the way they used to. Most brands and agencies blame the algorithm or iOS changes from years ago. The actual problem is simpler. The algorithm changed how it finds people and serves ads. Most advertisers haven't changed how they feed it.

Meta's Andromeda update fundamentally restructured the ad delivery system. Before Andromeda, the system operated within the audience constraints you set. Now it operates as a prediction engine, evaluating each impression for the likelihood of conversion, the expected value, and the predicted ad experience quality. Ads get distributed by probability rather than by audience bucket. That means creative plays a much larger role in shaping who the algorithm finds inside the audiences you allow (we wrote the full breakdown in why creative diversity is the only way to win with Meta's Andromeda algorithm).

This list is curated by Y'all, a boutique DTC performance creative and media buying agency. The 10 tactics below are the operating principles that produce performance under Andromeda, drawn from the accounts we run for health, wellness, food, beverage, beauty, and CPG brands. Each one targets a specific way the algorithm rewards genuine creative diversity, and each one shows up consistently in the audits of accounts that came to Y'all underperforming.

1. Build Structural Variants, Not Hook Swaps

Producing variants that differ at the level of hook, storytelling arc, visual style, and value proposition rather than just changing the first three seconds.

What the tactic is: Genuine creative diversity at the level Meta's system can actually distinguish. A static ad testing "saves you 20 minutes" and a video ad testing "people are switching from [competitor]" are different. Two cuts of the same interview with different opening lines are not.

Why it works under Andromeda: The algorithm needs ads it cannot tell are the same ad. When you give it variations on a single signal, it processes them as one piece of creative and never gives any of them enough data to find its own audience. Structural variants give the system the variety it needs to find different audience pockets for different messages.

How to implement it: Each variant should differ on at least two of these axes: hook, storytelling arc, visual style, value proposition leading the message. Same creator and same product is fine. Same opening line with different b-roll is not. Audit your variants by these criteria and prune the ones that don't qualify.

Common mistake: Treating "5 versions of a video with different thumbnails or text overlay colors" as 5 different ads. Meta processes them as one ad with five resolutions. Most accounts have a meaningful portion of their creative library failing this test.

2. Mix Formats in Prospecting (Static + Video + Creator)

Running at least three distinct creative formats inside prospecting campaigns rather than relying on one dominant format.

What the tactic is: A prospecting mix that includes static images, video, and creator-style content (UGC or partnership ads). The mix gives Andromeda enough format variety to find different audience pockets that respond to different content types.

Why it works under Andromeda: Different audience segments respond to different formats. Some people convert on a static with a clear value proposition. Others convert on a video with a narrative arc. Others convert on creator content that feels native. A single-format library can only reach the segments that respond to that format.

How to implement it: Build prospecting campaigns with at least three formats running in parallel. Allocate budget across formats rather than concentrating in the one that's currently winning, since diversifying expands who the algorithm can find for you.

Common mistake: Running 100% video in prospecting because the team produces video well, or 100% statics because the team can ship them faster. Single-format prospecting accounts hit creative fatigue faster and scale less efficiently than mixed-format accounts.

3. Simplify Account Structure

Reducing campaign and ad set fragmentation so Andromeda has enough data per cell to optimize effectively.

What the tactic is: Fewer campaigns, fewer ad sets, broader audiences. The structural opposite of the 2018-2022 playbook of segmenting campaigns by interest, lookalike, and audience type.

Why it works under Andromeda: When the structure is fragmented, each campaign competes with itself for the same audience segments and the system never has enough data per cell to optimize effectively. Spend gets split across too many small pools, which slows down learning and leaves the algorithm without enough information to find performance.

How to implement it: Audit your account structure for fragmentation. Consolidate prospecting campaigns with overlapping audiences. ASC and CBO with broad targeting and a handful of well-built ad sets typically outperform deep segmentation in 2026.

Common mistake: Building separate prospecting campaigns for every persona, lookalike, and interest set, then wondering why none of them are leaving learning. Fragmented structures are a media buying problem dressed up as a performance problem.

4. Run Broader Audiences With More Creative Variety

Loosening audience targeting and letting creative diversity do the targeting work that interest segments used to do.

What the tactic is: Broad targeting (often broad with minimal interest layering) paired with structurally diverse creative that lets the algorithm find different segments through different creative signals.

Why it works under Andromeda: Creative is the targeting mechanism in 2026. Tight audience constraints prevent the algorithm from finding the audience pockets your varied creative could reach. When you give the system broad targeting and varied creative, it uses the creative to find the right people inside the broad pool.

How to implement it: Test broad audiences against your current segmented setup with comparable spend. Most accounts find that broad with creative variety outperforms tight audience targeting once the creative library has enough structurally different concepts running. Pair this with simplified account structure (tactic 3) for compounding effect.

Common mistake: Treating broad targeting as "no targeting" and assuming it means low-quality traffic. Andromeda is using your creative as the signal even inside broad targeting. The audience quality is shaped by what the creative attracts.

5. Use Protected Testing Budgets for New Creative

Giving new ad concepts dedicated budget that isn't competing with proven winners during the learning phase.

What the tactic is: A testing structure (dedicated testing campaigns, A/B tests, CBO with minimum spend, or ASC with staged budget ramps) that ensures new concepts get fair evaluation before being thrown in with established winners.

Why it works under Andromeda: When a new ad enters a campaign dominated by proven performers, the algorithm gives the established winners the lion's share of impressions every time. The new ad never gets enough data to prove itself. Without protected spend, the testing process is statistically rigged against new concepts.

How to implement it: Build a testing campaign or testing ad set with a minimum budget per new concept so it gets a fair read before being judged. Run new concepts there until they hit a statistical threshold. Graduate winners into the main campaigns. Per discussion in the Foxwell Founders community (Rob Bettis on creative as a system), the brands consistently winning at paid media run a steady drumbeat of new concepts entering rotation every two to four weeks rather than briefing in batches when performance drops. Note that this testing structure is fully compatible with the simplified account structure in tactic 3: a single dedicated testing campaign on top of simplified main campaigns is still dramatically simpler than the deeply segmented prospecting setups the older playbook called for.

Common mistake: Throwing new concepts into existing high-budget campaigns and judging them after a day of spend. Judging a new ad after 24 hours of spend almost never tells you anything useful. The "losing" concept might have been your next winner if it had been protected.

6. Test Messages in Statics Before Producing Video

Isolating the message in static ads before investing production budget in video around it.

What the tactic is: Running a static-first testing motion where new value propositions, hooks, and angles get tested as static ads, then the winning message gets produced as video.

Why it works under Andromeda: Static ads test one variable: the message. Video ads test a dozen variables (hook, pacing, audio, creator, product shot, CTA). When a video underperforms, diagnosis requires guesswork. When a static underperforms, you've learned something clean about which message resonates. That cleaner signal compounds.

How to implement it: When a new value proposition or angle needs testing, build it as a static first. Run several message variants in static form, identify the winners, then produce video around the winning message knowing the core concept already works.

Common mistake: Going straight to expensive video production around an untested message. Brands that skip the static testing layer routinely spend $5K on video built around a value proposition nobody cares about, then blame the production rather than the message.

7. Build an Evergreen Creative Library

Maintaining a base layer of creative that runs year-round, with seasonal and promotional content layered on top.

What the tactic is: A library of evergreen concepts that work month after month, regardless of season or product cycle. Seasonal and promotional creative runs on top of this base, but the base stays stable.

Why it works under Andromeda: The algorithm compounds learning over time. When your best ad only works during a holiday promotion or references a specific product that rotates out of stock, you're rebuilding your creative foundation every few weeks. The system never gets enough time horizon to compound. Evergreen creative gives the algorithm a stable foundation to learn against. The case for evergreen got stronger under Andromeda specifically: per discussion in the Foxwell Founders community (Simon Robert on Andromeda), creatives that used to run for 4-6 weeks are dying in 10 days under the new system because the algorithm is burning through audiences faster.

How to implement it: Identify which of your concepts are evergreen (the value proposition holds year-round, no time-sensitive offer or out-of-stock product) and treat those as your base layer. Build the production calendar so evergreen concepts are constantly being produced and tested, with seasonal creative as additive rather than replacement.

Common mistake: Letting seasonal creative dominate the library because it performs strongly for 2-3 weeks. When the season ends, the campaign loses its best performer, costs spike, and the team scrambles. Treat seasonal creative like individual stock picks. Treat evergreen creative like the index fund that keeps the account healthy.

8. Test Partnership Ads

Running ads under creator or retailer handles instead of the brand handle, using Meta's partnership ads format.

What the tactic is: Ads that run from a creator's or retailer's account rather than the brand's account. They appear in the feed with the creator's handle, blending into organic content.

Why it works under Andromeda: Partnership ads carry built-in social proof and feel different in the feed than standard brand ads. Meta's case studies cite acquisition cost reductions as high as 19% on partnership ads versus brand-handle equivalents. The reduction in CAC makes them one of the highest-impact creative format changes available to most DTC brands.

How to implement it: Identify creators whose UGC content has performed well in your account. Set up partnership ads with those creators (Meta's Collaborative Ads framework handles the technical setup). Run the same creative concept under the creator handle and the brand handle in parallel and compare CAC.

Common mistake: Most DTC accounts have never tested partnership ads at all. They're leaving one of the most effective levers for improving paid social efficiency completely untouched. The barrier is administrative (setting up the partnership, getting creator accounts connected) more than performance-related.

9. Separate New Customer From Existing Customer Campaigns

Building campaign structure so prospecting reaches only new buyers and existing customers are excluded.

What the tactic is: Prospecting campaigns that specifically exclude existing customers (and often visitors, depending on retargeting strategy), with separate retargeting campaigns handling the existing audience.

Why it works under Andromeda: When prospecting campaigns aren't excluding existing buyers, your reported ROAS is inflated because repeat purchasers are easier to convert. The algorithm spends prospecting budget on customers who would have purchased anyway, and the dashboard shows numbers that don't reflect real new-customer acquisition.

How to implement it: Build a custom audience of all existing customers and exclude it from every prospecting campaign. Build a separate retargeting structure for the existing audience. Track new customer CAC at the campaign level rather than blended ROAS (the broader case for replacing blended ROAS with new-customer CAC is in blended ROAS is an illusion: the metrics 8-figure DTC brands actually optimize for).

Common mistake: Running broad campaigns that serve both new and existing customers and then optimizing against blended ROAS. The blended number hides the truth about acquisition efficiency. Real new-customer CAC is almost always meaningfully higher than what the platform reports when this separation isn't in place.

10. Diagnose Creative Failures by Dimension

Evaluating underperforming ads by which specific element failed (hook, message, proof, CTA) rather than discarding the whole ad.

What the tactic is: A systematic diagnostic process that breaks an underperforming ad into its component dimensions and identifies which one failed, so the next iteration fixes that dimension while preserving what worked.

Why it works under Andromeda: When an ad underperforms, the worst move is to label the entire ad a failure and start over. The opening might have stopped the scroll while the message failed to map to a real pain point. Or the message worked but the proof was missing. Identifying the failed dimension turns every losing ad into usable data.

How to implement it: For each underperforming ad, evaluate: Did the opening stop the scroll (CTR is the proxy)? Was the product and benefit immediately clear (look at watch time)? Did the message map to a real pain point or desire (qualitative read)? Was there credible proof (reading comments helps)? Was there a clear path to action (CVR is the proxy)? Then test new variants that fix the failed dimension. Per the Motion creative analysis SOP shared in the Foxwell community (Courtney Fritts), useful diagnostic thresholds to anchor against are hook rate above 30% TSR, hold rate above 8-10% (thruplay over impressions), and CTR above 0.8-1% (anything between 1-1.5% is solid, above 1.5% is high-converting).

Common mistake: Treating losing ads as failures and discarding them entirely. The compounding intelligence about what your audience responds to is what separates accounts that get better over time from accounts that stay stuck cycling through random ideas.

How Y'all Implements These Tactics in Practice

These 10 tactics work together as a system, not as a checklist. The brands we work with at Y'all run all 10 in parallel, with the production calendar, account structure, and testing methodology built around enabling the tactics rather than retrofitting them.

The starting point is usually the creative library. Audits typically find that a meaningful share of variants in an account don't qualify as structurally different under Meta's Andromeda criteria (hook swaps and text overlay changes that the algorithm processes as one ad). Pruning those and rebuilding the library with structural variants is the first move. Account structure simplification follows because broader audiences with more creative variety only work when the structure can support them. Then the testing motion gets rebuilt with protected budgets for new concepts and a static-first testing layer.

Once the system is running, the compounding effect kicks in. Evergreen creative builds up, partnership ads start expanding the format mix, the diagnostic process turns every losing ad into learning, and the new vs. existing customer separation gives the team honest acquisition reporting to optimize against.

For more on the algorithmic logic behind these tactics, see why creative is the real targeting mechanism in modern DTC advertising.

How This List Was Built

This guide was assembled using a combination of (1) the Andromeda tactics Y'all runs daily across client accounts in health, wellness, food and beverage, beauty, and CPG, (2) direct experience auditing accounts that came to Y'all after their previous agency couldn't break out of a Meta plateau, and (3) operator data and discussions from the Foxwell Founders community, including the Andromeda creative-fatigue observations from Simon Robert, the testing-cadence framing from Rob Bettis, and the Motion creative analysis benchmarks documented by Courtney Fritts.

The tactics are ordered roughly from most foundational (structural variants, format mix) to most impactful once the foundation is in place (partnership ads, dimension diagnosis). Most accounts running underperforming Meta need to start with tactics 1-3 before the higher-impact tactics produce their full effect.

Frequently Asked Questions

How many different ad creative formats should I be running on Meta?

Accounts running at least three distinct formats (static images, video, and creator or animated content) tend to surface more performance pockets. This gives Meta's Andromeda algorithm enough variety to learn which audience segments respond to which format. Some accounts can scale with a dominant format, but diversifying generally expands who the algorithm can find.

What changed with Meta's Andromeda update for DTC advertisers?

Andromeda shifted Meta from audience-based targeting to prediction-based delivery. The algorithm now evaluates every impression opportunity based on the likelihood a person will convert rather than just showing ads to the interest groups you selected. Within the constraints you set, creative diversity matters more because the creative is the primary signal shaping who sees and responds to your ads.

How often should I refresh my ad creative on Meta?

Rather than refreshing on a fixed schedule, monitor performance signals. When frequency rises above 2-3 on new audiences and costs start climbing, that's the signal to introduce new creative. Maintaining a library of evergreen concepts means you're not constantly starting from scratch when individual ads fatigue.

What's a realistic success rate for new ad concepts?

About 10%. One out of ten new concepts will be a strong performer. Two or three more might work with iteration. The rest provide learning even if they don't produce direct returns. Expecting a higher hit rate leads to under-testing and over-reliance on a small number of winning ads.

Should I test messages in static ads before producing video?

In many cases, yes. Static ads isolate the message from the production variables present in video (hook, pacing, audio, creator). Testing the core value proposition in statics first and then producing video around the winning message reduces production waste and improves the odds your video investment pays off.

What are partnership ads and why should DTC brands use them?

Partnership ads (or collaborative ads) run under a creator's or retailer's handle rather than the brand's. Meta's case studies cite acquisition cost reductions as high as 19%, though results vary by vertical and creator fit. They blend into the feed more naturally and carry built-in social proof. Most DTC accounts haven't tested them yet.

How do I know if my Meta account structure is hurting performance?

The most common structural problem is not separating new customer and existing customer campaigns. If your prospecting campaigns aren't excluding existing buyers, your reported ROAS is inflated. Check your frequency too. For cold prospecting, 2-3 frequency is reasonable. When rising frequency pairs with rising CPM or CPA, that signals fatigue.

How can DTC brands improve their ad campaign ROI with creative?

Focus on genuine variety over volume. Each concept should test a meaningfully different message, format, or angle. Use protected budgets for new concepts so they get fair evaluation. Evaluate underperformers by diagnosing which specific dimension failed rather than discarding the whole ad. Test partnership ads if you haven't already.

Wrapping Up

Meta's Andromeda update changed the rules for DTC creative. Audience-based targeting matters less than it used to. Creative diversity matters more. Brands that haven't updated their playbook are still running 2022 tactics against a 2026 algorithm and wondering why CAC keeps creeping up. The 10 tactics above are the operating principles that produce performance under Andromeda, and they work together as a system rather than as a checklist.

This list is curated by Y'all, the boutique DTC performance creative and media buying agency. Auditing accounts and rebuilding them around these tactics is the first move in most new client engagements. If your Meta account has plateaued and you're not sure which of these tactics would move the needle, reach out and we'll look at what you're running and share honest feedback.

‍