Scaling paid ads from 2x to 6x ROAS

The problem

MotoAmerica's official merch store was trading well. Organic traffic was converting, the product was strong, and the audience - motorcycle racing fans - was clearly defined. But paid ads were stuck at 2.19x ROAS across six campaigns, and every attempt to scale just pushed the return down further.

The brief was to figure out why the store couldn't scale paid acquisition and fix it. The goal was 4–5x ROAS, a 25% lift in sales, and a 15% increase in average order value by the end of Q4 2023.

The audit (4 weeks)

We ran a structured audit: one week of prep and data access, one week of quantitative collection (Ads Manager, Shopify Analytics, industry benchmarks), one week of qualitative analysis (on-site behavior, customer feedback, competitor stores), and one week of A/B test design.

The diagnosis was a list of missing fundamentals. No Conversions API, so iOS purchases were invisible to Meta. No retargeting audiences, so every dollar was going to cold traffic. No Lookalike audiences built from purchasers. Click-through rates were below benchmark. The site was slow, bounce rates were high on key pages, and there were no upsell or cross-sell blocks anywhere in the flow. Product images were JPGs instead of WebP, which added unnecessary weight to an already slow mobile experience.

What the audit found

• No Conversions API - iOS attribution was missing, so Meta couldn't optimize for mobile purchases.
• No retargeting campaigns - 100% of spend was going to cold traffic with no second touch.
• No Lookalike audiences - campaigns were using interest targeting instead of modeling from actual purchasers.
• CTR below benchmark - creative and copy weren't stopping the scroll.
• Slow site speed and high bounce rates on product pages.
• No upsell or bundle offers anywhere in the cart or product page flow.
• JPG product images instead of WebP, adding load time on mobile.

The changes

We implemented Conversions API so Meta could see iOS purchases. Retargeting campaigns went live targeting site visitors, add-to-carts, and product viewers. Lookalike audiences were built from the purchaser list and segmented by value. Creative testing focused on stopping power - product in context, not just product on white.

On the site side, we optimized product pages for speed, fixed high-bounce landing pages, converted images to WebP, and built upsell blocks into the cart. The campaign structure consolidated from six cold campaigns to three: one retargeting track, one Lookalike, and one Advantage+ campaign that let Meta optimize placement and creative dynamically.

December campaign breakdown

• Outerwear Sales - 11.93x ROAS. Seasonal push on jackets and hoodies, retargeting product viewers.
• Retarget Products - 4.80x ROAS. Broad retargeting of site visitors and cart abandoners.
• Advantage+ All Products - 4.27x ROAS. Dynamic campaign letting Meta optimize creative and placement.

The honest read

December is peak season for apparel, and this store sells motorcycle merch - a category that moves in Q4. The 270% sales lift is real, but it happened in a month that was always going to outperform October. To prove the changes stuck, you'd want to see January, February, or a year-over-year comparison.

The AOV goal was +15%. The result was +7.6% ($27.01 to $29.05). That's progress, but it's not the target. The upsell and bundle work started late in the quarter; a proper post-purchase flow and a product recommendation engine would close that gap.

What I'd do next

The AOV goal wasn't met. A post-purchase upsell flow and dynamic product recommendations on the cart page would move that number without needing more traffic.

The data is clean now, but there's room to scale. Testing a 2%–5% Lookalike audience expansion and a dedicated prospecting campaign for new customer acquisition would show whether the current efficiency holds at higher spend.

Email flows. There's no cart abandonment sequence, no post-purchase follow-up, and no win-back campaign. Klaviyo integration with a three-email cart recovery flow would recapture revenue that's currently walking away.