A cheap lead is not an eligible one.
A homebuilder in Brazil’s subsidized affordable-housing program, with developments across six regions. The real filter sits outside the ad platform: income inside the program band, clean credit and mortgage approval. Media has to deliver model-home visits rather than form fills.
The business context
The product is the same, the buyer is not. Each region has its own sales velocity, inventory and buying behavior, and media tracks that variation week by week alongside the sales team.
The funnel starts before media does
Income inside the program band, time in the federal housing fund and credit standing decide who can actually buy. Optimizing for lead volume ignores that filter.
A development sells out and comes off air
In the Northeast the towers sell faster. Budget has to move in days rather than monthly cycles, so the ads never promote what is already gone.
The decision happens at the model home
People who visit buy four times more often than people who only talk on the phone. The visit is the event the campaign chases.
Qualification feedback to the platform
The CRM sends back leads qualified on income and documents. Campaigns learn from that event, and the reported KPI becomes cost per qualified lead.
An income question inside the form itself
One income-band question cuts raw volume and lifts the visit rate. CPL goes up, cost per visit goes down. The plan takes that trade deliberately.
Media plan
Goal for the month: 12,400 leads at up to R$50, producing 1,295 model-home visits and 170 attributed sales. The media target is the visit; the lead is how you get there.
| Region | Inventory status | Spend | Share | Leads | CPL |
|---|---|---|---|---|---|
| São Paulo and ABC | High inventory | R$198,400 | 32% | 3,421 | R$58.00 |
| Rio de Janeiro | Medium inventory | R$111,600 | 18% | 2,067 | R$54.00 |
| Salvador metro | Final units | R$99,200 | 16% | 2,420 | R$41.00 |
| Recife metro | Final units | R$86,800 | 14% | 2,226 | R$39.00 |
| Goiânia | High inventory | R$74,400 | 12% | 1,617 | R$46.00 |
| Porto Alegre | Medium inventory | R$49,600 | 8% | 800 | R$62.00 |
| Month total | R$620,000 | 100% | 12,551 | R$49.40 | |
| Scenario | Assumption that changes | Leads | Visits | Sales | Cost per sale |
|---|---|---|---|---|---|
| Conservative | Mortgage approval drops to 33% under the bank’s current criteria | 12,551 | 1,310 | 138 | R$4,493 |
| Base case | Six-month historical rates, with qualification fed back to the algorithm | 12,551 | 1,310 | 172 | R$3,605 |
| Optimistic | The income question in the form lifts the visit rate to 22% | 11,400 | 1,454 | 191 | R$3,246 |
The optimistic scenario has fewer leads and more sales. That is the numeric case for why form volume is the wrong target in this operation.
Creative brief
The communication axis is trading rent for a mortgage payment on your own apartment. Every asset carries a verifiable offer and comes down when that offer changes.
- Audience
- Ages 25 to 45, within 8 km of the development, declared income inside the program band
- Format
- 9:16 video, 15s · Reels and Stories
- Specs
- 1080×1920, payment figure in an editable field, no rebuild required
- Pull rule
- Pulled when fewer than 15 units remain in the advertised tower
What you pay in rent already covers the payment on your own apartment.
Two bedrooms, an installment down payment and up to R$55K in subsidy. Check whether you qualify.
- Audience
- People who watched 50% of the upper-funnel video in the last 14 days
- Format
- Native lead form with an income-band question
- Specs
- Three fields, income band required, direct CRM integration
- Pull rule
- Cost per qualified lead above R$96 for 3 consecutive days
Three questions tell you how much subsidy you qualify for.
No credit check at this stage. You get an answer within 10 minutes.
- Audience
- Qualified leads who have not booked a visit within 7 days
- Format
- 4:5 static and Stories with a dynamic sales-office address
- Specs
- 1080×1350, real photography of that region’s model home, no generic render
- Pull rule
- Cost per booked visit above R$520
Seeing it in person changes the conversation.
Sales office open Monday to Sunday, 9am to 6pm, next to the construction site.
- Audience
- Lead base in regions with fewer than 30 units left
- Format
- 1:1 static with a unit counter
- Specs
- 1080×1080, unit count pulled from the sales spreadsheet, reviewed every Monday
- Pull rule
- The asset comes down the day the tower sells out
18 apartments left in tower B.
Same floor plan, same terms. Once they are gone, the next tower delivers in 2028.
Delivery dashboard
Day 18 of 30. The dashboard joins media and CRM so the conversation with sales opens on visits and sales rather than on form cost.
| Region | Spend | Leads | CPL | Visits | Sales | Cost per sale |
|---|---|---|---|---|---|---|
| São Paulo and ABC | R$119,000 | 2,222 | R$53.56 | 195 | 24 | R$4,958 |
| Rio de Janeiro | R$67,000 | 1,271 | R$52.71 | 128 | 15 | R$4,467 |
| Salvador metro | R$59,500 | 1,512 | R$39.35 | 168 | 24 | R$2,479 |
| Recife metro | R$52,100 | 1,389 | R$37.51 | 151 | 22 | R$2,368 |
| Goiânia | R$44,600 | 1,008 | R$44.25 | 104 | 14 | R$3,186 |
| Porto Alegre | R$29,800 | 512 | R$58.20 | 46 | 5 | R$5,960 |
| Running total through day 18 | R$372,000 | 7,914 | R$47.01 | 792 | 104 | R$3,577 |
The read this week
- The Northeast carries the group result. Salvador and Recife take 18% of the budget and produce 44% of sales, at less than half São Paulo’s cost per sale. The constraint there is inventory rather than demand.
- Porto Alegre does not add up. R$5,960 per sale, 65% above average, with the lowest visit rate in the operation. The hypothesis is distance to the model home, and the test is an ad offering weekend transport.
- The income question in the form worked. CPL rose R$4 and the visit rate rose 1.8 points, which cut cost per visit from R$512 to R$470.
- Decision taken to the client: hold Salvador and Recife at current budget until the new tower releases, and move R$18K from Porto Alegre to Goiânia, which has both inventory and a healthy visit rate.