A 119% ROAS increase Meta ads case study answers a question every ecommerce founder is asking in 2026: can we make more money without spending more on ads? Yes. We restructured one client’s Meta (Facebook) ad account, rebuilt their creative strategy, and delivered a 119% lift in ROAS (Return on Ad Spend), a 357% revenue increase, and a 42% drop in CPA (Cost Per Acquisition), all with $0 extra ad spend. This post breaks down exactly what we changed inside the account, why it worked in the post-Andromeda algorithm environment, and how you can replicate the framework.
TL;DR
– Result: +119% ROAS, +357% revenue, +303% orders, +255% conversion rate, -42% CPA over 3 months (March 13 to June 13, 2025). – Ad spend change: essentially flat. Sessions only grew 18%. – Website change: none. No offer change. No landing page change. – What changed: account structure collapsed from 20 active campaigns to 4, plus a research-driven creative system producing batch releases every 2 weeks. – Creative pricing model: client paid per winner, not per creative produced.
The Baseline: 20 Campaigns, Messy Targeting, Random Creative
The account we inherited was structurally broken by every 2026 Meta best practice. Between January and March 2025, the brand was running 20 active campaigns simultaneously. Each ad set mixed UGC (User Generated Content) videos, statics, and professional videos inside the same auction, forcing Meta to distribute budget across formats it could not fairly compare.
The five specific problems we documented:
- 20 active campaigns fragmenting the budget and preventing the learning phase from resolving.
- Messy audience targeting with overlapping interest stacks inside a single “conversions add to cart” ad set.
- No clear account structure distinguishing testing from scaling.
- Wrong audience segmentation with no exclusion between engaged audiences, existing customers, and cold prospects.
- Random creative testing mixing videos and statics in one ad set, which biases Meta toward the highest-engagement asset rather than the highest-ROAS asset.
The creative side was equally weak: 5 to 10 ads per month, no research foundation, and “cool looking” content that did not sell.
The New Account Structure: 4 Campaigns, Consolidated Budget
We collapsed 20 campaigns down to a 4-campaign consolidated structure aligned with how Meta’s auction actually distributes budget as of 2026. According to Meta’s own documentation on the learning phase, ad sets need roughly 50 optimization events per week to stabilize. Twenty campaigns starve every one of them.
The new setup:
– 1 testing campaign (standard structure) to prove new concepts. – 1 Advantage+ Shopping campaign per country for scaling proven winners. – 1 strategic retargeting campaign using catalog ads for product-specific engagers.
We also audited placements. Instagram was producing both the highest volume of sales and the lowest CPA, so we concentrated spend on IG placements rather than defaulting to Advantage+ placements everywhere.
Post ID Scaling: Don’t Kill the Testing Ad
When a creative won inside the testing campaign, we duplicated it into the Advantage+ Shopping campaign using the same post ID so social proof (likes, comments, shares) compounded across both placements. Critical detail: we never turned off the winning ad in the testing campaign. Killing it destroys accumulated engagement signal that Meta uses to rank the duplicated post ID.
Targeting was broad, filtered only by gender, with a 60-day purchaser exclusion so the budget acquired new customers instead of paying to reconvert people already reachable through email or SMS.
The Creative System: Research First, Then Batch Production
Better ads did the heavy lifting on the 119% ROAS increase, not budget shifts. Sessions only grew 18%, but conversion rate jumped 255%, which proves the traffic we were sending was dramatically more qualified. That only happens when the message on the ad matches the awareness stage of the person clicking it.
Our creative process:
- Audience research using Reddit threads, Facebook groups, and product reviews to extract real language, pain points, and objections.
- Data analysis of every historical creative to identify which angles, formats, and hooks had traction.
- Awareness-stage mapping across problem-aware, solution-aware, and product-aware buyers.
- Format decision: for this brand, research told us videos would outperform statics at the top of funnel. Statics were assigned to middle and bottom of funnel.
This full-funnel format split is a core principle of the post-Andromeda playbook, because Andromeda rewards accounts that let the algorithm distribute the right creative to the right stage of intent rather than forcing one format everywhere.
Batch Releases Every 2 Weeks
We launched a new creative batch every 14 days. Each batch contained:
– New concepts based on fresh research angles. – Iterations on the previous batch’s winners.
Each video was produced once and then output with 6 text-on-screen variations, each targeting a different awareness stage or pain point. One video, six messaging angles, minimal production cost. We routed all variations through a single ad set, isolated the winner, then moved that winner into the scaling campaign via post ID.
For statics, we produced roughly 50 ads in a couple of weeks across multiple languages. The winners were what we call original concept ads: creative that does not copy competitors, does not lean on generic “benefit bullet” layouts, and introduces a visual or narrative angle the audience has not seen before. Original concept ads consistently generate the majority of scalable winners in our accounts.
Why Sessions Barely Moved but Conversion Rate Jumped 255%
The single most important number in this case study is the 255% conversion rate lift on flat traffic. It tells us the algorithm was suddenly sending pre-qualified buyers instead of curious scrollers. Three inputs drove that shift:
– Consolidated budget let each ad set exit the learning phase and optimize on real purchase data. – Broad targeting with exclusions stopped wasting spend on returning customers and let Meta find high-intent lookalike behavior. – Awareness-stage creative matched the ad’s message to the exact objection the viewer was carrying into the funnel.
When those three inputs align, click quality changes dramatically. That is the mechanic behind a 42% CPA reduction on the same ad spend.
Pay Per Winner: The Pricing Model That Aligned Us With the Client
We charge per winning creative, not per creative produced. The client received 50+ statics and multiple video batches without paying a flat production fee for every deliverable. They only paid for ads that hit our win threshold and got promoted to the scaling campaign.
This matters because it forces our creative team to prioritize research and concept quality over volume. Producing 100 mediocre ads is a losing strategy in 2026. Research from Think with Google on creative effectiveness consistently shows that creative quality is the single largest driver of ad performance variance, and Meta’s Andromeda update amplified that further by concentrating impressions on top-decile creative.
FAQ
How did you increase ROAS by 119% without increasing ad spend?
We consolidated the account from 20 campaigns to 4, cut messy audience overlaps, split video and static formats across the funnel, and rebuilt the creative process around audience research and 2-week batch releases. Sessions grew only 18% but conversion rate improved 255%, which means each dollar of spend acquired dramatically more qualified traffic.
How many campaigns should an ecommerce brand run on Meta in 2026?
For most 6- and 7-figure ecommerce brands, 3 to 5 active campaigns is optimal: one testing campaign, one to two scaling campaigns (Advantage+ Shopping segmented by country if relevant), and one retargeting campaign. Running 20+ campaigns starves each ad set of the optimization events needed to exit the learning phase.
Why did you separate videos and statics into different ad sets?
Meta tends to distribute budget toward the ad with the highest engagement, not necessarily the highest ROAS. Videos naturally generate more engagement than statics, so mixing them in one ad set biases spend toward video regardless of purchase performance. Splitting by format lets each creative type compete against peers.
What is post ID scaling and why does it matter?
Post ID scaling means duplicating a winning ad into a new campaign using its original post ID so all likes, comments, shares, and video views compound across placements. It preserves social proof and gives the duplicated ad a stronger auction signal from day one. Never turn off the original testing ad, or you lose that accumulated engagement.
Should I exclude existing customers from prospecting campaigns?
Yes. We exclude 60-day purchasers from prospecting so ad spend goes toward net-new customer acquisition. Existing customers are far cheaper to reconvert through email and SMS. Paying prospecting CPMs to reach them is inefficient.
How often should I launch new creative batches?
Every 2 weeks works well for most ecommerce brands at scale. It gives the algorithm enough time to identify winners inside each batch without letting creative fatigue build up in the scaling campaign.
About the Author
Antonio Ventre is the founder of Skaleit, a Meta ads agency for ecommerce brands. Skaleit has managed 8-figures in Meta ad spend across supplement, fashion, skincare, jewelry, and DTC categories, with a specialization in post-Andromeda scaling strategy.
Want Skaleit to Build This System for Your Brand?
If you want a Meta ads account restructured for 2026, a creative system that pays per winner instead of per deliverable, and a scaling framework built for the post-Andromeda algorithm, book a free strategy call with the Skaleit team. We only take on brands we’re confident we can scale profitably.