Anonymised and sector-framed. Figures are the client’s own, reported the same way each month, and shown here with their permission.
A retailer of considered, high-value goods was running a lead campaign that looked, on paper, like it was working, while the numbers underneath it were fiction. Faulty tracking was firing on every conversion rather than on genuine enquiries, so the reported figures could not be trusted and spend was being optimised against noise. Behind that noise sat a real problem: a punishing cost per lead and a conversion rate under half a percent, at 0.34%. Before anything could be improved, the measurement had to be made honest.
Fix the data first, and every decision after it can finally be trusted.
Fix the data before touching the spend
We started where the problem actually was, not where the reporting said it was. The tracking was counting every conversion indiscriminately, inflating performance and feeding the ad platform false signals to optimise against. We rebuilt the conversion tracking so that a recorded lead meant a real enquiry, deduplicating events and aligning what the platform saw with what the business valued. Only once the measurement was trustworthy could any decision that followed be trusted too.
Re-baseline against honest numbers
With clean tracking in place, the true picture emerged, and it was harder than the dashboard had suggested. A conversion rate of 0.34% and a high cost per lead were now measured facts rather than estimates. We treated this corrected baseline as the real starting line, so that every subsequent change could be judged against reality instead of against a flattering error.
Fix the structure, not just the settings
Faulty data had been masking structural weakness in how the campaign was built. We restructured the account so that budget followed genuine intent: tightening how audiences and search intent mapped to ad groups, removing spend that the corrected tracking now exposed as wasted, and concentrating investment where real enquiries were actually coming from. For a considered, high-value purchase, the job is to reach the small number of serious buyers efficiently, not to chase volume for its own sake.
Align the message with a considered purchase
A premium product with a long deliberation cycle needs the campaign and the landing experience to speak to that mindset. We tightened the match between the promise in the ad and what the visitor met on arrival, reducing the friction that quietly kills conversion on high-value goods. The aim was fewer wasted clicks and a cleaner path for the people genuinely in market.
Optimise on signal, not noise
With honest conversion data flowing back to the platform, optimisation could finally do its job. The algorithm was now learning from real enquiries rather than phantom ones, so bidding and targeting sharpened in the right direction over time. We monitored against the corrected baseline and kept pruning what did not earn its place, compounding the gains rather than resetting them.
The turnaround started with the data. Because the platform had been optimising against conversions that were not real, fixing the tracking changed what the campaign was learning from, and that alone reset the trajectory. Once genuine enquiries were the signal, the structural and message work compounded on top of accurate feedback rather than fighting it.
The results reflect that sequence. Cost per lead fell 85% because spend stopped chasing false conversions and concentrated on real intent, while lead volume grew 6.4x as the corrected campaign reached the right buyers more efficiently. The conversion rate moving from 0.34% to 1.19% shows the same story from the site's side: a cleaner match between ad and landing experience turning more of the right visitors into real enquiries.