The account
A multi-brand dealership in Lima selling European cars, with list prices from US$18,000 to US$113,000. Spend managed: US$29,200 over three months, April to June 2026, across native lead forms and website conversions.
The problem
The portfolio was judged by cost per lead. At almost the same CPL, the cost of a lead the sales team could actually work varied 1.8× from one brand to another. The cheapest leads were not the useful ones.
What I did
- May: built lookalike audiences per brand, seeded with the quality leads from the client’s CRM instead of every form fill.
- June: moved budget from website conversion campaigns to native lead forms, where cost per quality lead was lower.
The result
Cost per quality lead went from US$387 in April to US$213 in June. Monthly spend stayed within a 10% range, so the improvement didn’t come from spending more or less.
How to read this case
- Quality lead is a lead the sales team marked as reachable and worked: in negotiation, qualified, open or won. It’s not a sale. The client’s CRM doesn’t isolate closed sales, and this case doesn’t claim them.
- Why not ROAS. The buying cycle takes months and the deal closes in person at the dealership. A platform ROAS here measures form fills, not revenue.
- Why not CPL. The brand with the cheapest CPL (US$2.82) had the worst quality rate (1.10%). The one with the most expensive CPL (US$8.32) had the best (2.19%). Neither was the most efficient on the combined metric. Budget was allocated by cost per quality lead.
- What’s excluded. The case covers the three brands with the most spend. Two smaller ones are left out: one without CRM integration and one with four quality leads in three months. Including every brand with spend, the figure is US$273.68 per quality lead, against US$273.92 for the three case brands on the same basis. The cut doesn’t change the result.
- How many leads. 108 quality leads out of 6,307 recorded in the CRM: 1 in 58 reaches a sales negotiation. The case doesn’t claim that rate is high. It shows that the cost of getting each one fell 45% without touching the budget.
- Why the series starts in April. A CRM migration in February reassigned lead statuses in bulk, so January to March isn’t comparable and is left out of every quality reading.
Monthly evolution, Apr–Jun 2026
See the full table
| Month | Spend | Leads | Quality leads | Cost per quality lead |
|---|---|---|---|---|
| April | US$10,448.57 | 2,210 | 27 | US$386.98 |
| May | US$9,074.09 | 2,029 | 35 | US$259.26 |
| June | US$9,780.66 | 2,156 | 46 | US$212.62 |
| Total | US$29,199.64 | 6,393 | 108 | US$270.37 |
Each change comes before the month it improves: lookalikes went live in May, the budget shift in June.
Breakdown by brand, Apr–Jun 2026
See the full table
| Brand | Spend | Leads | CPL | Quality leads | Cost per quality lead | Quality rate |
|---|---|---|---|---|---|---|
| A | US$12,509.76 | 3,285 | US$3.81 | 59 | US$212.03 | 1.81% |
| B | US$4,697.22 | 1,667 | US$2.82 | 18 | US$260.96 | 1.10% |
| C | US$11,992.66 | 1,441 | US$8.32 | 31 | US$386.86 | 2.19% |
| Total | US$29,199.64 | 6,393 | US$4.57 | 108 | US$270.37 | 1.71% |
Brand B has the cheapest lead in the portfolio and the weakest sales performance. Optimizing for CPL would have moved budget in the wrong direction.
Source and scope
Spend and platform leads from Meta Ads Manager (USD), Apr 1 – Jun 30, 2026. Lead quality from the client’s CRM cross-reference file, same period. Monthly rows come from the ad-level export and differ from the campaign view by 0.35% (campaigns outside the nine selected). Platform and CRM lead counts differ by 0.4–2.4% per month; the quality rate uses the CRM base (6,307 leads). No direct CRM access: the data stops at the lead classified by the sales team, not the sale. Price range from the dealership’s public list prices, not transaction prices. Client anonymized.