Skip to content
Schedule a free consultationSee results
← All case studies
PPCLead Generation4 months · April – August 2026

Lead Gen Recovery

UK premium vehicle services · Google Ads (Search + Performance Max)

An account that had collapsed to four enquiries a month at £145 each. Three months later it produced 37 at £29 — an 80% cut in cost per lead, with 86% more spend delivering 825% more enquiries.

-80%
Cost per lead
£145.19 → £29.22
// Snapshot

By the numbers

9.3x
Monthly enquiries
4 → 37
£21.62
Best cost per lead since
part-month, following August
15.88
Wasted-spend ratio at audit
well-run accounts sit under 2.5
10 of 13
Primary conversion actions junk
bidding trained on noise
42
Duplicate keywords removed
bidding against each other
83%
Impressions lost to budget
on the best campaign in the account
// The Problem

Where we started

The account had once worked well: the previous November and December produced 41 and 43 leads at around £21 and £23 each. Then it broke. January collapsed to eight leads at £138.26. February and March partially recovered but never returned to form. April was the worst month on record — four leads at £145.19 each.

The business was spending roughly the same money and getting a tenth of the enquiries. Demand had not disappeared; the structure and the conversion signal had turned against it.

The audit found 13 primary conversion actions with 10 of them junk — map directions, page engagements, smart-campaign leftovers — so Google's bidding was being trained on noise rather than enquiries. £936 of the previous 90 days' spend, 41% of it, had gone to search terms that never converted. And the single best campaign in the account, converting at £37 a lead, was capped at £4 a day while losing 83% of its available impressions to budget.

// What We Did

The approach

01

Clean the signal first

We demoted the ten junk conversion actions, kept the three real lead actions — quote requests, calls, email clicks — and set realistic values on each so the bidding could tell a £100 quote request from a £50 email click. Until this was done, every other change would have been optimised against the wrong target. The account had been 'optimised' that way for months.

02

Stop the waste

Negative keyword lists built against the £936 of non-converting search-term spend, plus cross-campaign negatives to fix £297 of clicks landing in the wrong campaign. Three dedicated shared negative lists were added later to route each service line and the brand traffic to the correct campaign instead of letting them cannibalise each other.

03

Fix the structure

The 42 duplicate keywords were consolidated. Four oversized ad groups — 16 to 26 keywords each, and the source of every Quality Score problem in the account — were split. A −30% bid adjustment went on desktop, which was converting at less than half mobile's rate.

04

Rebuild Performance Max, then fund the winners

The existing Performance Max campaign carried a page-view conversion goal, so it was optimising towards people looking at a page rather than enquiring. We paused it, launched a rebuilt campaign with the correct goal, and topped the asset library up to its 20-image cap. Only once the signal was clean did budget move — towards the best-converting service campaign and the brand campaign, the latter converting at £2 a lead.

// Results

What happened

  • April (worst month, and where we start): £581 spend, 4 leads, £145.19 each. July: £1,081 spend, 37 leads, £29.22 each.
  • Read the two ends together — 86% more spend produced 825% more enquiries. This was not a case of buying volume; efficiency improved at the same time as scale.
  • Comparing the four broken months (50 leads at £67.12) against the two clean months after the work landed (73 leads at £33.28) gives the same picture without cherry-picking a single month: 46% more leads at half the cost each.
  • The first thirteen days of August ran at £21.62 a lead — the account's best rate in nine months, on a pace that would pass its previous best month for volume. That is a part-month figure and should be read as a trend indicator rather than a result.
  • The account has now recovered past the efficiency it had before it broke, rather than merely back to it.
// Next step

Want results like these?

Schedule a free consultation and we'll run the same diagnostic on your account.

All figures pulled directly from the client's Google Ads account at account level. Client name and location anonymised pending permission; real data available on request.