Cost per lead down 58%.
A medical aesthetics device manufacturer was paying nine times more per lead than it used to. Here's what I found, what I changed, and what's still left to do.
The numbers so far.
Figures from the ad account, leads counted on click date. Client and product names removed. Recent days may still move as late leads arrive.
A slow slide that kept speeding up.
- Cost per lead rose almost ninefoldOver the same stretch, spend grew more than 80-fold.
- The slide started earlyCosts were already climbing months before a new brand campaign launched, then accelerated once it took nearly 60% of spend.
- The cheapest source became the most expensiveOne product campaign went from the best source of enquiries to the worst.
- Targets only ever went upEvery bid target had been raised, so the platform kept getting room to bid higher.
Three causes, not one bad campaign.
1. No commercial targets
Every bid target sat above the client's real cost per lead, from 1.4x to 4.7x higher, and none had ever been lowered.
2. Automation left unsupervised
Auto-applied recommendations added broad keywords, removed protective negatives and, on one campaign, wiped its target entirely.
3. Budget not following lead quality
Spend went to the campaigns that converted worst. Half the spend bought 79% of the leads; the other half bought 21%.
The finding that changed the plan
Obviously wrong searches were only 2.2% of spend. The real waste was spread thinly across more than 3,000 search terms. Longer blocklists can't fix that. Match types, campaign structure and sensible targets can, so the plan started there.
Three phases, plus a daily routine.
- Phase 1
Stop the bleeding
Brand keywords to exact match, broad match moved to phrase across the account, a cap on brand click prices at under half the old ceiling, and daily budgets reset in stages.
- Phase 2
Fix the structure
Gave three products their own campaigns instead of sitting in brand, split the main product by search intent, and removed negatives that were blocking campaigns' own best terms.
- Phase 3
Targets to the client's numbers
Measured every campaign against the client's cost per lead, stepped targets down in stages, and moved budget to the client's priority order.
- Ongoing
Governance
Full reviews twice a week, a daily spend check against each campaign's normal, early-warning limits below the weekly cap, and every change logged with a reason and a read date.
What isn't fixed yet.
The trend is right, but the job isn't finished.
- The target hasn't been reachedCost per lead is still above the client's target, and every product campaign is over its own number.
- Lead quality isn't measured yetSales take weeks to months to close, and sales data by campaign hasn't been shared. Cheaper leads aren't automatically better leads.
- Landing pages are unchangedThey depend on the client's web team and are next in line.
- Volume dipped brieflyWhen spend was cut by about 40%, leads dropped for a short time. The client chose lower costs over volume.
Habits, not hacks.
Judge by what the business can afford
One campaign looked healthy against its old target but cost 1.7x what the client could pay per lead.
Step down, never cliff
Targets and budgets came down in stages, with written stop rules, so delivery didn't collapse.
One change at a time
Budget and target changes never land on the same day, and each gets a 3-day and 7-day read.
Check "done" against the account
Every "already fixed" claim is checked against a fresh export before it's reported.
Work inside the client's limits
When a pause wasn't approved, budget and target cuts did the job instead.
What's next
Lead values on each form, remarketing to past visitors, and budgets weighted towards weekdays.
Want this kind of look at your account?
A free ad-to-customer audit, with me, on your accounts. You get a short video and a one-page list of where your leads go cold.