Halved cost-per-conversion, doubled revenue, two markets scaled.

The starting account had multiple product lines competing for the same budget without clear prioritisation logic. We rebuilt it so every dollar's destination was visible, and every campaign's performance was measurable in isolation.

• Separate Search campaigns per major product line, with controlled budget caps
• Shopping prioritised on the highest-margin SKUs, with explicit product priority levels
• Brand defence campaign ring-fenced to keep brand traffic measurement clean
• Geo-targeting split AU + NZ into separate sub-campaigns for distinct optimisation

The restructure made the account legible. Suddenly we could see which products were converting, which audiences were responding, and which campaigns were carrying inefficient spend. That visibility was the prerequisite for every subsequent optimisation gain.

Sensory product buyers come from three very different starting points, and treating them as one audience meant nobody saw an ad that spoke to them. We split the audience and let each speak its own language.

• Healthcare professionals: copy + landing pages anchored in clinical context, evidence, and therapeutic outcomes
• Parents researching for children: copy focused on calming effect, school suitability, durability
• Adults seeking solutions: copy focused on personal sensory experience, dignity, everyday wearability
• Each segment got its own ad set, its own ad copy variants, and its own landing-page flow

Segmentation lifted both conversion rate and Quality Score, the two metrics that compound against cost-per-conversion. As Quality Score improved, the same bid bought better positions; as conversion rate lifted, each click was worth more. Both directions of compounding.

The 50% reduction in cost-per-conversion didn't come from a single big move. It came from twelve weeks, then twenty-four, then a hundred, of small continuous adjustments: ad copy variants, bidding strategies, negative-keyword expansion, audience filter refinement.

• Weekly review of conversion data, with hypothesis-driven test queue
• Continuous A/B on headlines, descriptions, ad extensions
• Bidding strategy adjustments as Search and Shopping performance diverged across products
• Negative-keyword discipline preventing budget bleed into low-intent queries

Over the engagement, monthly conversions climbed from ~380 (Dec 2022 low) to 815 (Sep 2024 peak), while cost-per-conversion settled into a sustainable range. That's the compounding: small disciplined optimisation, repeated for years, producing engines that competitors can't catch.