How $41,540 in Meta Ads spend became $526,488 in ecommerce revenue.

The account was burning spend on a broken funnel.

When we audited the account in February 2026, Meta was delivering clicks. The site was losing them in the first scroll. The cart recovery flow sent one email, 24 hours later, with no segmentation. We pulled 90 days of data and found the leak in three places.

Leak 1: 68% of paid traffic bounced from the product page in under 15 seconds. The page had been designed by a developer, not a copywriter. Headline was the product name. Hero image was a flat product shot.

Leak 2: The cart recovery flow sent one email. a 20% discount. 24 hours after abandonment. No segmentation. No send-time optimisation. 71% of these emails landed at 3 a.m. local time in the recipient's market. Open rate: 4.2%.

Leak 3: Server-side tracking wasn't deployed. The Meta algorithm was optimising against a signal it could only see 38% of the time. It was guessing.

Rebuild the mechanism before scaling the spend.

We paused scaling for two weeks. Not because we wanted to. because the math demanded it. Scaling a leaking funnel is the most expensive mistake in paid media.

The rebuild order, in sequence:

- Product page: new copy hierarchy, social proof surfaced above the fold, urgency elements without being sleazy - Four-email cart recovery flow: browse → cart → checkout-start → purchase, with send-time optimisation per market - Creative testing matrix: six concepts live at any time, rotated weekly, killed at statistical significance not gut feel - Server-side tracking via Meta CAPI + browser pixel + enhanced matching. algorithm finally saw 92% of conversions

Only after all four were live did we turn the spend back up. The instinct was to scale sooner. The math said no.

Test like a portfolio, not a lottery.

Six creatives live at any time. Each had a clear hypothesis: UGC with founder, lifestyle product, before/after, social proof, offer-led, educational. Rotated weekly. Killed when statistical significance said one was outperforming. not when the founder's gut said so.

Audience layering: coldest prospecting looked different from warm retargeting, which looked different from the recent-purchaser nurture. Each piece knew what the other was doing. The creative that worked at the top of the funnel didn't get served at the bottom. because the message needed to be different.

Email flows segmented by browse, cart, and checkout-start. Each with its own cadence and copy. The flow that recovered a browser was very different from the flow that recovered a checkout-start abandoner.

$526,488 in tracked revenue. 12.87X return.

Six months. Three markets. One accountable team. The brand now runs profitably on paid acquisition. something it couldn't do before. and the email flow alone contributes 18% of monthly revenue without a single additional dollar of ad spend.

The numbers, for the record:

- Total ad spend: $41,540 over six months - Revenue attributed to paid: $526,488 - Blended ROAS: 12.87X - Email-attributable revenue: 18% of monthly total - Average order value: up 22% - Repeat purchase rate: up from 11% to 28%

That's the difference between a funnel and a mechanism. The funnel moved traffic. The mechanism compounds.