In our Google Ads test selling a telemedicine platform to European clinics, reported conversions cost €288.87 against a €15,000 average deal — under 2% of deal value. Narrowing to explicit telemedicine intent brought that to €237.12, while broader digital-health wording cost €366.49.
Case study
What paid acquisition actually cost us: four tests, four markets, real numbers
Benchmarks · 2025–2026 · Meta and Google Ads. Every figure below comes from a validation test we ran — not a modelled estimate, not an industry average.
The numbers
| Test | Channel | Geo | Model | Spend | Leads | Cost per lead |
|---|---|---|---|---|---|---|
| Fractional property investing app | Meta | UK | B2C | £63.71 | 33 | £1.93 |
| Startup accelerator, founder intake | Meta | USA & EU | B2B | $583.97 | 74 | $7.89 |
| VR therapy app, practitioner intake | Meta | USA | B2B | $765.77 | 46 | $16.65 |
| Telemedicine platform, clinic intake | EU | B2B | €5,777 | 20 | €288.87 |
Currencies are left as spent. Converting them would imply a precision the tests do not have, since each ran in its own market at its own time.
Segment and creative splits
Where a broader and a narrower segment ran side by side:
| Test | Narrower segment | Broader segment | Difference |
|---|---|---|---|
| Startup accelerator | Seed-accelerator intent — $6.48 | Broad startup intent — $10.09 | narrow is 36% cheaper |
| Telemedicine platform | Telemedicine intent — €237.12 | Digital-health wording — €366.49 | narrow is 35% cheaper |
Where creative format decided the result:
| Test | What won | What lost |
|---|---|---|
| Fractional property, UK B2C | Static only — one visual carried 21 of 33 leads at £1.60 | Video was not the winning route |
| VR therapy, USA B2B | Video — best set at $10.12 | Static feed sets produced zero leads |
Methodology
What a lead means here. In the three Meta tests, a lead is a completed website lead form. In the Google test the platform-reported conversion is used; the qualification stage behind it is not disclosed, so €288.87 should be read as cost per reported conversion, not cost per qualified opportunity.
Sample sizes are small on purpose. Spend ranges from £63.71 to €5,777. These were validation budgets, not scaling budgets — the goal was to find out whether a channel, audience, and message could produce responses at a cost the economics allowed, and to find it out before committing real money. A test that answers that question on £64 has done its job.
Every test ran 2025–2026. Platform costs move. Treat these as a reference for the segment shapes, not as current rate cards.
Clients are anonymised. Category, channel, geography, and business model are disclosed; client names are not, under NDA.
What the numbers actually show
1. Cost per lead tracks what the lead is worth, not how good the marketing was. £1.93 for a UK consumer signing up to an app with a £100 entry point and €288.87 for a European clinic evaluating a €15,000 platform are not different levels of performance. They are different products. Against that €15,000 average deal, €288.87 is under 2% of deal value.
2. Narrowing intent cut cost by about a third — in both tests where it was measured. Seed-accelerator intent came in 36% below broad startup intent. Telemedicine intent came in 35% below general digital-health wording. Two tests, two channels, two markets, nearly the same number. The cheapest available lever in both cases was not creative or bidding — it was refusing the broader audience.
3. "Video wins" is not a rule. In the UK consumer test, static did the work. In the US practitioner test, static produced zero leads across the whole run and every one of the 46 leads came from video. The same format decision went opposite ways in the same year on the same platform. Format follows the buying situation, not best practice.
4. Inside a winning format, one creative usually carries the result. In the property test, a single static visual produced 21 of 33 leads at £1.60 — below the £1.93 blended figure. The blended number describes the test. It does not describe what scaling would have cost.
What these numbers do not prove
- They are not lead quality. Except where noted, these are form completions. Only the telemedicine test carried a known deal value behind it.
- They do not predict your cost. Different offer, geography, season, and account history all move cost per lead more than anything on this page.
- They do not survive scale. Every figure comes from a small budget. Cost per lead typically rises as spend increases and the cheapest audience is exhausted.
- They are not a channel recommendation. Meta appears three times and Google once because of where the buyers were, not because one platform is better.
FAQ
Frequently asked questions
In our UK test for a fractional property investing app with a £100 entry point, leads cost £1.93 at a 4.69% all-click rate, using static creative only. The best single visual produced leads at £1.60.
Less than most teams assume. These four tests ran on £63.71, $583.97, $765.77, and €5,777. Each one answered whether the audience, message, and format could produce responses at a viable cost. Scaling budgets come after that answer, not before it.
No. In our US test selling a VR therapy tool to practitioners, static feed creative produced zero leads and all 46 leads came from video. In our UK consumer property test, static was the winner. The format that works follows the buying situation.
Narrow the audience to explicit intent. In the two tests where broad and narrow segments ran side by side, the narrow segment came in 36% and 35% cheaper — a bigger move than any creative change in the same tests.
Prefer a conversation first?
If you would rather talk it through before sending a brief, book a short routing call and we will point you to the right next step.