They asked for Google Ads.
Meta became their best channel.
Capital Car Finance came to us for help with Google Ads. The audit opened a second question: could Meta add a useful source of enquiries alongside search? They were sceptical, agreed to a controlled test on Meta, and nine months later it is the channel they want to put more money into.
The highest return of any paid channel
In September, every pound spent on Meta came back as four pounds of revenue, measured on closed deals in the client's own records rather than platform-reported conversions. No other paid channel CCF runs returned more that month.
What the brief actually was
- Originally: get more out of Google Ads.
- After the audit: test an additional source of paid demand alongside search.
- Prove Meta could produce leads that the finance team could actually fund, not just cheap form fills.
- Do it on a test budget small enough that a sceptical client could say yes.
That last point is why it happened at all. Nobody was asked to take Meta on faith. They were asked to fund a test with a clear pass or fail.
The audit changed the question
Capital Car Finance approached us looking for Google Ads support. That is what we audited. The account was not the problem we expected it to be.
Their Meta ad account had not spent a single pound in more than two years.
Search captures people actively looking for car finance. With Meta inactive, we wanted to test whether paid social could add suitable enquiries alongside that existing activity. The test needed to be assessed on quality as well as lead cost.
Why they were sceptical
Car finance lives or dies on lead quality. A lead who cannot pass an eligibility check costs the business a phone call and returns nothing, and the common view is that social platforms produce exactly that kind of lead.
It is a fair concern. Meta will happily sell you cheap enquiries from people who were never going to qualify.
What we agreed to measure
Not cost per lead on its own. The test would be judged on the same outcomes the business already used: how many leads went on to complete a PCP eligibility check, how many were quoted, and how many turned into funded deals.
If Meta could not hold up on those numbers, it would be switched off.
Test, scale, refine
Meta went live in January 2026 on a deliberately modest budget. Each stage only started once the one before it had earned it.
Launch lean, with two jobs
A broad prospecting campaign to find new drivers who were not yet searching, and a small remarketing campaign to bring back people who had already visited. Both were optimised to the client's PCP lead conversion rather than to clicks or traffic.
Let the algorithm find the audience
Rather than guessing at interests, we kept targeting broad and let the conversion signal do the work. There was no gender targeting at all, yet more than four in five leads ended up coming from men. By March the cost per lead was a third of what it had been in the first month, which was the evidence the client needed to go further.
Scale on a second campaign
In April we added an expanded audience campaign using Advantage+ audiences, and spend nearly tripled across the second quarter. The original campaign kept running, so new budget was tested alongside a proven baseline instead of replacing it. One rate-led ad carried almost all of it: 97% of that campaign's leads came from a single creative.
Assess funded deals alongside lead costs
From June 2026, Meta results were reviewed against the client's own lead tracker, alongside every other channel, on eligibility checks, quotes and sales. That is what moved the conversation from "is Meta safe?" to "how much more can it take?"
Nine months, three phases
Every figure here is relative to the first three months, when Meta was running on test budget. The pattern is the useful part: spend nearly tripled, efficiency dipped while the new campaign bedded in, then came back to test-phase levels at full scale.
| Phase | Monthly spend | Monthly leads | Cost per lead |
|---|---|---|---|
| Test Jan to Mar | Baseline | Baseline | Baseline |
| Scale Apr to Jun | 2.8× | 2.4× | +15% |
| Refine Jul to Sep | 2.7× | 2.8× | −3% |
Source: Meta Ads Manager, PCP lead conversions, January to September 2026, all campaigns.
What changed inside the account
The headline cost per lead hides three shifts that explain why it fell, and why it held once the budget went up.
PCP leads relative to link clicks
In September, recorded PCP lead conversions were approximately one third of link clicks, compared with fewer than one fifth in February. This is an aggregate account ratio, not a count of unique people followed from click to lead.
Media got dearer, leads got cheaper
Between February and September the cost of reaching a thousand people rose by a third. Over the same months the cost per lead fell by a tenth. Lead costs improved despite a higher CPM. This comparison does not isolate the effects of creative, targeting, conversion changes or shifts in the audience.
A launch-month problem, caught early
In month one, remarketing was reaching a small pool of past visitors more than 20 times each, at more than double the cost per lead of prospecting. From February its budget was cut right back and the spend moved to prospecting, which is where the growth came from.
Who converted, and where
Nobody briefed Meta on who a car finance customer is. The campaigns were set up broad, with no gender targeting, and the pixel did the teaching. What it found is worth knowing, because it is not the audience most people would have guessed at.
These breakdowns describe where recorded leads came from. They do not establish which demographic group produced the most funded deals.
- From men
- 84%
- Aged 25 to 54
- 77%
- From Reels and Stories
- 45%
- Reels and Stories cost per lead vs feed
- −15%
- Served inside the Facebook or Instagram app
- 98%
Source: Meta Ads Manager, PCP lead conversions, January to September 2026, all campaigns. The lowest cost per lead of any age band came from 35 to 44 year olds, and the cheapest major placement was Instagram Reels.
Cheap leads were never the goal. Leads the finance team could fund were.
What happened to lead quality
This was the question the whole test hinged on, and it went the right way. In September, nearly two in three Meta leads went on to complete a PCP eligibility check, up from around two in five the month before.
Further down the funnel, the share of Meta leads that became funded deals rose by more than half between June and September, and the cost of each sale more than halved.
- Revenue per pound of spend, Sep
- 4.0×
- Cost per sale, Jun vs Sep
- −56%
- Lead to sale rate, Jun vs Sep
- +57%
- Revenue per pound of spend, Jun
- 1.9×
- Leads completing an eligibility check, Sep
- 64%
- Same measure, Aug
- 41%
Source: the client's own lead tracker, which records Meta outcomes from June 2026. Revenue and sales are funded deals attributed to Meta leads. Absolute figures withheld. Revenue divided by media spend is not profit after management and other costs. These monthly comparisons do not isolate the effect of individual campaign changes.
What we are not claiming
We are not claiming Meta replaced search. Search still brings in most of CCF's leads, and roughly one pound in five of the paid media budget now goes to Meta. The point was never to swap one channel for another. It was to stop depending on one.
We are also not claiming it was a straight line. When spend nearly tripled in the second quarter, cost per lead went up before it came back down, which is normal when a new campaign is still learning. And September is the first month Meta returned more than every other channel. It was not ahead in July. We think the trend is real, but it is one month, and we would rather say so.
Position at September 2026
A test that the client was sceptical about is now the channel they are asking to grow. The next phase is more budget on Meta, measured the same way.
The creative is already moving. A new set of "check your rate" ads went live in September to challenge the ad that has carried the account since April. Early cost per lead is around a third lower, on a sample still too small to call.
Meta launched in January. The client lead tracker provides the funded-deal and revenue comparisons from June onwards. Any further budget increase needs to be judged against those outcomes as well as lead cost.
Explore our paid social management approach and our work with financial services advertising.
Is all your demand coming from one channel?
We work with businesses spending £10k a month or more on paid media. Start with a free audit and we will tell you where your next customers are likely to come from, even if it is not the channel you asked us about.