Competitor Research · Market screen
Which Markets Give Us Cheap Leads — and Why
Competitor ad feed vs our own cost per lead in 18 metros we have actually run · Ad Library sweep August 22, 2026
The one-line rule. A metro gives us a cheap lead when its Facebook feed already carries price-forward replacement ads (mini-split $/mo, install price up front, lead-gen shops) — 18% or more of HVAC ads — and the market is thick enough to matter — 40+ HVAC advertisers. The rule calls 14 of 18 metros correctly on CPL. It says nothing about booking rate or lead intent (Phoenix, Lubbock) — those are setter- and client-side.
1What the screen measures
- Price-forward share — % of a metro's HVAC ads that are mini-split monthly-payment ads, full-install-price ads, or ads run by lead-gen shops. High = the audience is already trained to respond to a replacement offer on Facebook. We do not copy those offers; we keep ours.
- Advertiser count — distinct pages running HVAC ads. Under 40 = nobody has proven the channel there (Charleston, Baton Rouge, Vancouver, Salt Lake).
- Lead-gen shop share — shown separately because Phoenix (10%) produced $34 leads with zero bookings: trained to click, not trained to buy.
- Crowdedness does not hurt us: Lubbock has ~10× Dallas's HVAC ads per capita and the same CPL as Atlanta.
2Live client metros
| Metro | Our account | Our CPL | Book % | HVAC ads | Advertisers | Price-forward | Lead-gen shops | Screen |
|---|---|---|---|---|---|---|---|---|
| Dallas–Fort Worth, TX | Boost Air | $23 | 22.0% | 296 | 97 | 12% | 5% | COLD |
| Cleveland, OH | State Heating | $34 | 35.8% | 107 | 70 | 21% | 4% | PASS |
| Lubbock, TX | Call on the Double | $35 | 15.8% | 132 | 78 | 19% | 5% | PASS |
| Atlanta, GA | Weather Masters | $35 | 37.7% | 214 | 97 | 23% | 14% | PASS |
| Baton Rouge, LA | Cold AC | $40 | 23.9% | 18 | 9 | 22% | 17% | THIN |
| Baltimore / Harford Co., MD | Hughes | $41 | 45.5% | 126 | 84 | 21% | 4% | PASS |
| Central New Jersey | Sam's Air Control | $42 | 42.3% | 175 | 73 | 23% | 10% | PASS |
| Charlotte, NC | QPS N. Charlotte | $46 | 35.1% | 109 | 64 | 20% | 6% | PASS |
| Tucson, AZ | Heating & Air Paramedics | $49 | 35.9% | 159 | 67 | 9% | 2% | COLD |
| St. Louis, MO | AirBear STL | $57 | 28.0% | 75 | 48 | 4% | 1% | COLD |
| Charleston, SC | QPS Charleston | $64 | 32.7% | 21 | 11 | 0% | 0% | THIN |
| Austin, TX | Act Right | $74 | 24.5% | 249 | 89 | 14% | 3% | COLD |
| Dayton, OH | Paradise Heating | $95 | 36.7% | 168 | 67 | 11% | 3% | COLD |
| Nashville, TN | Hendersonville HVAC | 115 | 72 | 19% | 4% | PASS |
Good CPL = $46 or under (90-day KPI, May 24 – Aug 22). Exceptions the rule misses: Dallas (12% price-forward, best CPL in the book — Boost Air's scale and 1.43% CTR carry it); Baton Rouge (thin but $40 CPL — CPA still poor).
3Churned-client metros — the out-of-sample test
| Metro | Our account | Our CPL | Book % | HVAC ads | Advertisers | Price-forward | Lead-gen shops | Screen |
|---|---|---|---|---|---|---|---|---|
| San Antonio, TX | American Air Designs (cash-strapped) | $24 | 197 | 48 | 14% | 11% | COLD | |
| Phoenix, AZ | Turn Key (0 booked; highest wrong-intent kill rate — Sean 08-22) | $34 | 0.0% | 486 | 255 | 28% | 10% | PASS |
| Las Vegas, NV | Go Ecco (ghosted) | $49 | 123 | 38 | 18% | 1% | THIN | |
| Vancouver, WA | QPS Vancouver (cancelled) | $91 | 14 | 9 | 21% | 0% | THIN | |
| Tampa, FL | Air Mavericks (tough market) | $110 | 469 | 264 | 13% | 2% | COLD | |
| Houston, TX | Synergy HVAC (never launched) | 414 | 192 | 15% | 4% | COLD | ||
| Salt Lake City, UT | Home Efficient (volume complaints) | 28 | 12 | 39% | 14% | THIN |
Predictions were written before this data landed: Phoenix high, Tampa low. Both held on CPL. Vancouver passes the index but fails density (9 advertisers) — the screen's second half catches it. San Antonio's $24 is 8 leads over one month; treat as noise. Kokomo and Providence were pulled but excluded: the keyword search returned national ads (0–1% mention the metro) — a small-town pull that hits the 900-ad cap is not a local read.
4What this is not
- Not a booking predictor. Phoenix: cheapest-looking market on the board, 0 booked, highest wrong-intent kill rate. Lubbock: $35 CPL, 15.8% booking. CPA splits into CPL (market + creative) × booking (client + setters); this screen covers only the first.
- Not "match their offers." We never run tune-up or repair ads. The index reads demand-readiness, not a creative brief.
- Small sample. 18 metros with an outcome. Rank correlations are 0.4–0.6. Good enough to screen a prospect's metro in 60 seconds; not good enough to overrule a signed client.
5How to use it
- Prospect call prep: pull the prospect's metro (one Apify run, ~$0.50, 2 min), read price-forward % and advertiser count. PASS = expect $35–46 CPL. THIN or COLD = price the deal for $60–95 CPL or pass.
- Existing clients: COLD/THIN metros with a live client are where creative angle matters most (Tucson, Dayton, Austin) — the feed isn't doing the warm-up for us.
- Next data join (not built): KILL SWITCH reason mix by metro, to find what predicts intent the way this predicts cost.