2.3→4.6 ROAS, blended

Situation

A home goods brand doing roughly $180K a month had spent two years optimizing everything except the thing that mattered. Media buyers rotated audiences and creatives, agencies came and went, and blended ROAS stayed pinned around 2.3 — profitable enough to keep going, too thin to scale.

The traffic itself was fine. Click-through rates were above category benchmarks, cost per click was sane, and the product had strong reviews. But of every 100 paid visitors, fewer than two bought anything. The account's problem did not live in the ad account.

Diagnosis

The audit shifted attention from campaigns to the click's destination, and the picture was unambiguous.

The landing experience was a generic product grid: paid clicks carrying a specific intent — a promoted bundle, a seasonal offer, a solution to a concrete problem — all landed on the same catalog page and were left to fend for themselves. Mobile, which carried 78% of traffic, loaded in almost five seconds; nearly a third of paid clicks bounced before the page finished rendering. The checkout demanded account creation, and shipping costs appeared only at the final step — the classic two-step cart abandonment machine.

Analytics compounded the damage. The browser-side pixel was losing roughly a third of conversions to tracking prevention, so the platforms were optimizing on partial data and the team was killing campaigns that were actually working.

System built

The rebuild ran the full circuit — page, measurement, then traffic — in that order, because relaunching campaigns onto a broken page would only have paid for the same waste faster.

• Dedicated landing pages for each core traffic intent, designed from the ad backward: the angle in the creative continues in the headline, the objections surface in the page's order, the offer matches the click.
• A mobile-first rebuild engineered for Core Web Vitals: LCP under two seconds, no layout shift, third-party scripts deferred or removed.
• Checkout stripped to its minimum: guest purchase by default, shipping shown early, one screen fewer than before.
• Server-side tracking with event deduplication, restoring the lost third of conversion signal to the platforms and to reporting.
• Campaigns relaunched against the new pages with a clean testing structure, budget concentrated behind the two strongest intents.
• A standing CRO backlog: research, hypothesis, test, document — page improvements shipped monthly instead of at redesigns.

Results

Conversion rate on paid traffic went from 1.4% to 3.1% — more than doubling revenue per click before a single dollar of media efficiency. With the restored conversion signal, the platforms' optimization sharpened; CPA fell 44% blended, and ROAS climbed from 2.3 to 4.6 over five months.

The compounding detail: none of the gain came from spending more. The same traffic, hitting infrastructure built to convert it, produced twice the return — which is what made the next stage of scaling affordable.

Before / After

Case studies are anonymized to protect client confidentiality. Metrics reflect actual account performance over stated periods.