CPA cut from $135 to $47. 2.1× more conversions at the same total spend.
A D2C pet supplement brand at $95K/mo Shopify revenue was burning $84K/year on misallocated Meta spend. Audience overlap, no purchaser exclusions, retargeting at 35% of total budget, and a Pixel firing wrong on checkout. We surfaced the leak in week 1 and fixed it across 9 months. CM2 doubled. Same spend, double the customers.
"Our CPA keeps creeping up. We don't know why."
This pet supplement brand had grown to $95K/mo Shopify revenue mostly through Meta. The previous 6 months had been a slow CPA creep — $98 → $112 → $128 → $135. Every month, slightly worse. Every month, slightly less profitable. The brand couldn't tell whether it was creative fatigue, audience saturation, attribution issues, or something else.
Our CPA used to be $98. Now it's $135. Nothing about our offer or product has changed. Something is wrong but I don't know what. — Founder, audit call
The reconciliation against Shopify Reports showed the gap: Pixel was double-firing on checkout (counting a single purchase as 1.7 conversions on average). Meta thought the brand was getting more conversions than reality, so the algorithm was learning on inflated signal. That alone explained 25% of the CPA creep.
The other 75% was structural: 35% of total Meta spend going to retargeting, no purchaser exclusions on prospecting, and audience overlap pushing CPMs 18% above benchmark.
Four leaks compounding to $7K/mo. $84K a year.
Each one fixable in days. None of them invisible to a senior operator who knew where to look.
Pixel double-firing on checkout
Standard Pixel + a manually-fired purchase event from a third-party app were both running on the order-status page. Every single purchase counted as 1.7 conversions on average. Meta's algorithm was learning on completely wrong signal.
Recovery: data accuracy +70%No purchaser exclusions on prospecting
Meta prospecting was running broad with zero purchaser exclusions. The algorithm was finding repeat buyers and counting them as new acquisitions. ~28% of "new customer" CPA was actually re-acquiring existing buyers.
Recovery: CPA −22% on prospectingRetargeting at 35% of total spend
Retargeting CBO was getting 35% of Meta budget — reporting an 8.4× ROAS. Reconciled against Shopify, real incremental contribution was 1.6×. Most of it was reaching cart abandoners who would have converted anyway.
Recovery: $11K/mo redirectedCreative monoculture · pet lifestyle only
Single creative format (lifestyle photos of dogs) running across 11 different audiences. CTR decay clear by month 3. Missing UGC, product demo, before/after testimonial formats. CPC climbing 30% over 6 months.
Recovery: 4 new winning formatsPixel fix on day 1. Architecture in week 2. Compound across 9 months.
Some leaks are 1-day fixes that reset the entire system. The Pixel double-fire was that. Then 8 more months of compounding architecture work.
Pixel double-fire fixed
Identified the third-party app firing duplicate purchase events. Disabled it. Standard Pixel + Conversions API only. Within 7 days, Meta's reported CPA "increased" $135 → $172 — but that was the right number. Algorithm finally had clean signal.
Purchaser exclusions + retargeting cap
90-day purchaser exclusions across all prospecting. Retargeting reduced from 35% to 12% of total spend, capped going forward. Audience separation — no overlap between cold prospecting, ASC, retargeting, DPA.
Conversions API + Enhanced Match
Server-side CAPI deployed. Customer email + phone hashed for Enhanced Match. iOS attribution recovery improved 25%. Algorithm signal cleaner than the brand had ever had.
Creative library expansion
3–5 new concepts every 2 weeks. UGC reviews from real pet owners (not stock dogs) became the #1 format by month 2. Before/after product demos for joint supplements specifically. Founder origin story for cold audiences. Lifestyle creative downscaled.
Selective scale on validated winners
Spend reallocated within Meta — prospecting CBO grew while retargeting stayed capped at 12%. Total Meta spend held flat at $30K/mo. CPA dropped from $172 (real number after Pixel fix) to $89 by month 6.
Compound — CPA halves again
By month 9, CPA at $47. Same total spend ($30K/mo). Conversion volume up 2.1× — from 222/mo to 468/mo. CM2 climbed from 11% to 21%, doubling the brand's true profit margin per order.
Same spend. 2.1× the customers. CM2 doubled.
The most expensive part of the leak wasn't the dollars wasted — it was the customers we weren't acquiring at the wasted spend levels. Fixing structure unlocked both.
CM2 doubling on flat spend is the kind of compound that gets noticed by founders who run real P&Ls. 468 customers/mo at $47 CPA > 222 customers/mo at $135.
Pixel/CAPI accuracy is the foundation. Most accounts don't have it.
The single most-impactful change in this engagement was a 30-minute fix in week 1: identifying and disabling the third-party app that was double-firing purchase events. Every Meta optimization decision for 6 months prior had been made on inflated conversion data. The algorithm was learning on bad signal.
This is more common than people realize. Almost every D2C account I audit has at least one tracking issue — Pixel firing wrong, CAPI not deployed, Enhanced Conversions broken, GA4 missing checkout events, or some third-party app interfering.
Why this fix unlocked everything else
Once Meta had clean conversion signal, the algorithm started finding actual buyers instead of optimizing toward inflated conversion counts. CPA "went up" in week 2 (from $135 to $172) because the platform was finally reporting real numbers. Then it started dropping — fast — as the architectural fixes (purchaser exclusions, retargeting cap, audience hygiene) compounded.
By month 9, CPA at $47 — a third of the starting number, on cleaner data than the brand had ever had. That's not aggressive optimization. That's removing waste.
$84K of annual waste eliminated without touching topline. We literally double our customer count at the same Meta budget. Compounding into year 2 — the difference is starting to compound on top of compound. Best money I've spent on this brand.
— Founder, D2C Pet Supplements · 9 months inDifferent shapes. Same methodology.
Suspect your tracking might be lying?
The pet supplements brand above started with the $499 Google Ads Audit — and the Pixel double-fire was found in the first 30 minutes. Live walkthrough of the account, written 22-point plan, real number on the leak. Credited to month-1 if you sign within 14 days.