If your Meta account is performing worse than it did eighteen months ago and you have not changed anything, you are experiencing what the platform changed. Two shifts — the Andromeda system rollout and the maturation of Advantage+ campaigns — rewrote how the algorithm selects audiences, prices inventory, and rewards creative quality. This is not a temporary adjustment. It is a structural change to how Meta advertising works, and most ecommerce brands are still running account structures designed for the 2023 system.
This article covers what changed operationally, what it means for a $1M to $10M DTC brand running $30K to $150K per month in Meta spend, and what fundamentals are still working exactly the way they always have.
The Andromeda Shift: What Actually Changed in Meta's Core System
Andromeda replaced Meta's previous manual-targeting-based delivery system with an AI-driven creative-matching engine, completing its global rollout to all advertisers by April 2025. Before Andromeda, the system's job was to find people who matched your audience parameters. After Andromeda, the system's job is to find people who respond to your creative. That is a different input and a different optimization target.
The practical shift: your targeting configuration used to determine who saw your ad. Now your creative does. Andromeda reads visual signals, copy tone, product context, audio, and behavioral matching to identify buyers. A UGC video that shows a 35-year-old woman using your skincare product will reach 35-year-old women with skincare purchase intent — not because you targeted that segment, but because the creative signals it. Interest stacks, demographic restrictions, and behavioral exclusions are increasingly irrelevant inputs because the system is no longer optimizing against them.
According to Meta's Andromeda documentation reported by AdExchanger, the system delivers a 6% improvement in ad recall, 8% higher ad quality, and runs on 10,000 times the model capacity of the prior architecture. These are not incremental improvements. They represent a different class of system making different decisions.
The operational implication for DTC brands is this: the skill that generated ROAS in 2022 was audience configuration. The skill that generates ROAS in 2026 is creative production. Brands that shifted their resource allocation toward creative volume and quality in 2025 are outperforming brands that kept their emphasis on campaign configuration.
What Advantage+ Actually Needs to Perform
Advantage+ Shopping Campaigns are not a performance upgrade you apply to your account. They are a signal-dependent system that performs well when they have sufficient conversion data to optimize against — and significantly underperforms when that data is thin, broken, or inconsistent.
The threshold that matters is 50 purchase events per week at the campaign level, with 100 or more giving reliable optimization. Below that number, the algorithm is guessing. A brand spending $20,000 per month generating 25 weekly purchases should not be running Advantage+ as their primary structure — the system does not have enough signal to learn. A brand at $80,000 per month generating 120 weekly purchases is a strong candidate for Advantage+ because the signal is sufficient for the algorithm to find buyers efficiently.
According to Meta's internal data and third-party benchmark reporting from Enrich Labs, Advantage+ Shopping Campaigns average 4.52x ROAS compared to the 2.79x ecommerce average across all campaign types. Meta reports that brands using Advantage+ correctly see a 22% higher ROAS and 12% lower cost per action versus manual campaign equivalents. These numbers assume the data layer is intact.
The other Advantage+ dependency is tracking quality. Browser-based pixel tracking alone is now approximately 40% accurate due to iOS restrictions and ad blockers. Running Advantage+ on degraded tracking data is one of the most common reasons brands see the feature underperform: the algorithm is optimizing against incomplete signals and rewarding the wrong behavior. Brands seeing Advantage+ underdeliver should audit their Conversions API setup before drawing conclusions about the feature itself.
The Account Structure That Works in 2026
Consolidation is the correct structural principle for Meta in 2026. The logic is simple: the algorithm needs concentrated purchase signal to optimize. Fragmented signal — split across many small campaigns, ad sets, and budgets — is weaker signal, which means slower learning and less efficient delivery.
For a brand spending $30,000 to $100,000 per month, the 2026 account structure typically works best as: one to two Advantage+ Shopping Campaigns for broad prospecting, one manual retargeting campaign for cart abandoners and checkout initiators, and no more than six to eight active ad sets in total. This is a consolidation from what most accounts were running in 2022 and 2023, when 14 or 20 ad sets across four or five campaigns was a common setup.
A DTC apparel brand that consolidated from 16 ad sets across five campaigns down to two Advantage+ campaigns in mid-2025 cut CPA by 23% over the following six weeks with no change to creative. The campaign structure change gave the algorithm enough pooled purchase signal to exit the learning phase reliably and optimize delivery at scale. The ads did not change. The signal quality did.
One rule that has not changed: every campaign still needs 50 purchase events per week to exit the learning phase. This is a function of budget, not structure. If consolidation means some campaigns are generating fewer than 50 weekly conversions, the fix is to combine those campaigns, not to add budget.
Creative Is Now the Targeting Mechanism
Because Andromeda uses creative signals to identify and match buyers, the creative brief has changed. Ads are no longer just persuasion tools — they are targeting inputs. A creative that clearly signals who the product is for will be shown to people who match that signal. A generic creative with broad appeal will be distributed broadly and inefficiently.
The formats performing best in 2026 are UGC-style video between 15 and 30 seconds for cold prospecting, carousel ads for consideration-stage audiences comparing options, and static images for retargeting where intent is already established. Research from wetracked.io on post-Andromeda creative performance confirms that short-form video under 15 seconds with authentic UGC production outperforms polished brand creative for cold audience prospecting at current CPM rates.
Creative volume requirements have also increased. The practical benchmark: three to five genuinely new creative concepts per month for every $50,000 in monthly Meta spend. This means different hooks, different angles, different formats — not five versions of the same ad with different fonts. A supplement brand spending $60,000 per month needs six to ten new creative concepts monthly to give Andromeda sufficient variation to test and rotate before fatigue sets in.
The signal to watch for creative fatigue under Andromeda is faster than the pre-2025 standard. CTR dropping below 1.5%, CPM rising more than 25% week-over-week, or ROAS declining for three or more consecutive days on a previously stable creative are all indicators to rotate before the decline compounds. Andromeda's faster delivery cycle means creative exhausts at higher speed than it did under the previous system.
The hook rate metric — the percentage of video viewers who watch past three seconds — is the leading indicator for creative performance under Andromeda. A hook rate below 25% means the creative is not stopping the scroll, and no amount of downstream optimization will improve a creative that does not earn a view.
How to Measure Performance Correctly in 2026
Platform ROAS reported in Meta Ads Manager is an unreliable standalone metric for account decision-making in 2026. It under-credits Meta's influence on branded search and direct traffic, overstates last-click performance in multi-touch journeys, and produces inconsistent numbers depending on attribution window settings. Brands that scale or cut based solely on platform ROAS are making decisions on incomplete data.
Marketing Efficiency Ratio (MER) is the more reliable companion metric. The formula: total revenue divided by total marketing spend across all channels. A brand generating $600,000 in monthly revenue while spending $100,000 on Meta and $80,000 on other channels has an MER of 3.3x ($600K / $180K). If the same brand's Meta Ads Manager reports a 4.8x platform ROAS, the MER is telling a different story — it accounts for all spend, not just Meta's view of its own credit.
The practical use of MER is for scaling decisions. If MER is stable or improving as Meta spend increases, the channel is generating incremental revenue and the scale is justified. If MER declines as Meta spend increases, the marginal spend is not generating proportional revenue — which means the account may have a funnel problem, a saturation problem, or a margin problem, not a targeting problem. Knowing when to scale versus when to fix requires looking at MER, not just platform ROAS, to make that call correctly.
Track both metrics together: Meta platform ROAS for creative and campaign-level diagnostics, and MER at the account level for scaling and budget allocation decisions.
What Still Works Exactly the Same
Two years of platform changes have made ecommerce brands anxious about which fundamentals still apply. Most of them do.
The three-layer full funnel works. Cold prospecting to reach new buyers, warm engagement for consideration, hot retargeting for high-intent users — this architecture still reflects how buyers make decisions and how Meta's inventory is priced. What changed is that Advantage+ automates more of the audience identification within each layer, not that the layers themselves are obsolete.
Dynamic Product Ads for cart abandonment retargeting still deliver the highest ROAS of any campaign type in most ecommerce accounts. Showing someone the exact product they browsed or added to cart is as effective as it ever was. This campaign type benefits from Andromeda's faster learning cycles, not disrupted by them.
Offer quality still determines whether clicks convert. No creative optimization, account structure change, or algorithmic improvement fixes a weak offer. If cold traffic CVR is below 0.8% on a product priced above $80, and creative is generating healthy CTR, the constraint is the offer. Diagnosing a ROAS drop before touching the ad account still starts upstream of campaign settings.
The learning phase threshold has not changed: 50 purchase events per week at the campaign level remains the minimum for Meta's algorithm to optimize reliably. This was true before Andromeda and remains true after it.
Key Numbers for the 2026 Meta Landscape
Figures every DTC operator running Meta Ads should know heading into 2026.
ROAS Lift from Advantage+
Brands using Advantage+ Shopping Campaigns with clean Conversions API tracking and adequate purchase volume report 22% higher ROAS versus equivalent manual campaign structures, according to Meta's internal benchmarks.
Average Advantage+ ROAS
Advantage+ Shopping Campaigns average 4.52x ROAS across ecommerce, compared to the 2.79x average across all campaign types — a gap that only materializes for accounts meeting the 50 purchase events per week threshold.
Browser Pixel Accuracy in 2026
Browser-based pixel tracking now captures approximately 40% of actual conversions due to iOS restrictions and ad blockers. Running Advantage+ on pixel-only data means the algorithm optimizes against incomplete signals.
Frequently Asked Questions
Is Advantage+ better than manual campaigns for ecommerce in 2026?
Advantage+ Shopping Campaigns outperform manual campaigns for brands generating at least 50 purchase events per week with clean Conversions API tracking in place. For brands below that threshold, manual campaigns with consolidated ad set structure give the algorithm more stable signal. The correct answer depends on your current purchase volume and tracking quality, not a blanket preference for either format.
How much spend do I need before Advantage+ performs reliably?
The relevant threshold is conversion volume, not spend. Advantage+ needs 50 purchase events per week at the campaign level to exit the learning phase. At a 2% CVR on paid traffic, that implies roughly $25,000 to $40,000 in monthly spend depending on your CPA. Brands below that volume typically see more stable performance from manual campaigns with broad targeting.
What does "creative is the new targeting" mean in practice?
Under Andromeda, the creative signals you feed the system determine who sees your ads. A video that shows a real customer using your product in a recognizable setting signals buyer intent to the algorithm. Your creative brief now functions as targeting specification: be explicit about who the product is for, what problem it solves, and who the ideal customer looks like in the creative itself.
Should I still use interest targeting after Andromeda?
Interest targeting is largely redundant for most ecommerce campaigns. Andromeda identifies buyers through creative signals and conversion history. Running broad targeting with Advantage+ and letting creative do the audience work outperforms layered interest stacks for most DTC brands. Niche products where buyer intent is hard to infer from creative alone may still benefit from light interest guidance in early campaign delivery.
What is MER and how is it different from platform ROAS?
Marketing Efficiency Ratio (MER) is total revenue divided by total marketing spend across all channels. Platform ROAS is revenue Meta attributes to its ads divided by Meta spend only. MER captures Meta's halo effect on other channels and avoids attribution window distortions. For scaling decisions, MER gives a more accurate picture of whether additional Meta spend is generating incremental business revenue.
How often should I refresh creative in 2026?
Most creatives exhaust within two to four weeks under Andromeda's faster delivery cycles. The practical signals to watch: CTR below 1.5%, CPM rising more than 25% week-over-week, or ROAS dropping for three or more consecutive days. For brands spending $50,000 or more per month, three to five new creative concepts monthly is a reasonable minimum to sustain performance.
If your Meta account is running on a 2023 structure in a 2026 system, the gap between what you are spending and what you should be returning is mostly structural — not creative, not targeting, not budget. The Andromeda and Advantage+ changes are recoverable with the right account architecture and creative cadence. Mapping where your account sits against the 2026 baseline is exactly what we do in the Growth Diagnostic Sprint.