Supercharging revenue by 46% with smarter ads

SofaBay sells sofas online across the UK, with strong products and a catalogue that customers were already buying from. But both ad accounts had been left to run on autopilot, with no clear structure and almost no testing.
‍

On Meta, everything ran through one main prospecting campaign and a single remarketing campaign. Advantage+ creative enhancements were left on, so the catalogue appeared on every ad and new visuals could never be tested properly. Winning ads were neither protected nor optimised. Most orders came through the catalogue, yet there was no dedicated catalogue campaign. The ads relied on one line of visuals and never communicated SofaBay’s key selling points: fast delivery, pet-friendly fabrics and bestsellers. Interests, audiences, locations, age groups and ad angles had never been tested.
‍

On Google, the budget was spread across many Performance Max campaigns, one per product group, so purchases were split between them and none collected enough data to optimise. Location targeting was wrong, so SofaBay was paying for traffic from outside the country it delivers to. The ROAS targets were set too high for Google to reach, so campaigns underspent and brought little volume. The Merchant Center feed was in poor shape.
‍

The challenge was bigger than adjusting a few campaigns. SofaBay needed a clear account structure on both platforms, a proper testing framework, and a clean product feed: a foundation that could take more budget and turn it into more sales.

‍

We approached SofaBay as a restructuring project, not a quick account tweak. The goal was to give every campaign one clear job, start testing systematically, and build a setup that could take more budget without losing efficiency.
‍

A Meta structure built for testing and sales
We replaced the single catch-all setup with three campaigns, each with its own role. ABO prospecting campaigns reach new audiences and test new angles, visuals and messages. A separate catalogue campaign is where most sales happen: new customers first see the prospecting ads, then the catalogue follows them. Keeping it separate also lets us run tests on the catalogue itself. A retargeting campaign reaches people who visited the site in the last 30 days, pushes the bestsellers, and is now testing FAQ-led ads to handle objections.
‍

Creative that speaks to different buyers
We turned off the creative enhancements that put the catalogue on every ad, so each new visual could be judged on its own. Then we started testing different angles for different audiences: fast delivery, pet-friendly sofas and bestsellers, alongside tests on interests, audiences, locations and age groups. Winning ads are now protected and scaled, and weak ones are paused.
‍

A consolidated Google Ads account
We merged the scattered Performance Max campaigns so purchases collect in one place. The main campaign is now a prospecting PMax that looks for new, cold audiences and doesn’t take credit for brand searches. The brand is protected by separate campaigns, and Search is split into brand and generic searches. We fixed location targeting, so SofaBay now only pays for traffic from the country it delivers to. On bidding, Search moved from Maximize Clicks to Target Impression Share, so SofaBay appears in the top position for practically every brand search.
‍

A Merchant Center feed rebuilt from scratch
We removed duplicate products, so clicks now concentrate on the main products instead of being split between competing copies. We corrected titles and images by hand, product by product, and added the missing attributes: material, normalised colours, gender, condition and product highlights. Around half the catalogue (the products we actively advertise) got new descriptions, and over 80% of those got exact dimensions. All of this went live without a single new error or disapproved product.

‍