A bad lead vendor decision does not cost you a setup fee. It costs you a quarter. You fund ad spend out of your own account, you pay per lead on top of it, your reps burn six weeks on leads that will not convert, and by the time the pattern is undeniable you have handed away the best selling months of the season. In a trade averaging $12,000 a ticket, three wasted months at a normal appointment volume is six figures of missed revenue.
The thing that is supposed to prevent that mistake is the case study. And most contractor lead generation case studies are worthless — not because they are lies, but because they omit exactly the numbers that would let you check the claim.
Here is what a real one contains, how to run a vendor's numbers through your own math, and the questions to ask before you sign.
Why do most contractor lead generation case studies prove nothing?
Because they show one number in isolation. There are three standard versions of empty proof.
The screenshot. An ads manager panel showing $8.14 per result. No date range, no trade, no market, no indication of what a "result" was. A $8 result can be a video view, a landing page click, or a form fill from someone who typed "asdf" into the name field. It tells you nothing about whether a homeowner ever answered the phone.
The testimonial. "These guys changed my business." That is a sentence, not evidence. It has no volume, no cost, no time frame, and no way to tell whether the contractor is still a client or churned out eight months ago.
The vanity number. "Over 40,000 leads generated." Across how many contractors, how many years, how many trades? A number that large with no denominator is designed to be impressive rather than checkable.
The common failure in all three is the missing denominator. Leads without spend. Percentages without a base. Revenue without a cost. When you see a lead generation claim, your first instinct should be to ask what number it was divided by. If the vendor cannot answer immediately, they either never ran the math or do not want you running it. Real proof looks like the anonymized studies on our case studies page, where the trade, the state, the model, and the raw counts are all stated together.
What is a lead generation case study actually supposed to prove?
Four claims, and only four. Everything else is decoration.
- Volume. Can this vendor produce enough leads in a market like yours to matter?
- Exclusivity. Did each of those leads go to one contractor, or was the same homeowner sold three times?
- Quality. Did the leads turn into appointments and contracts, or just into phone numbers?
- Durability. Did the account still work in month six, or was it a launch spike that decayed?
Most case studies address the first one and skip the other three.
| Claim | Evidence that proves it | Evidence that does not |
|---|---|---|
| ClaimVolume | Evidence that proves itLead count with a date range and market | Evidence that does not"Thousands of leads generated" |
| ClaimExclusivity | Evidence that proves itWritten policy plus whose ad account ran the spend | Evidence that does not"High-intent leads" |
| ClaimQuality | Evidence that proves itBooked appointment rate and close rate | Evidence that does notCost per lead alone |
| ClaimDurability | Evidence that proves itMonth-by-month counts across 6+ months | Evidence that does notA single best month |
Which seven numbers should a vendor put in front of you?
Use this as a checklist on the call. Ask for all seven. Note which ones they can answer without going to look.
- Trade and market. Roofing in a Texas metro is not the same as fencing in a rural county. If the proof is in a different trade and a different market density, it is directional at best.
- Date range and tenure. "Six months" means something. "Recently" does not. Ask how long the account has been running and whether it is still active today.
- Total leads and last-90-day leads. Two numbers, not one. Lifetime total tells you scale. The last 90 days tells you whether it still works.
- Ad spend across that window, and whose account it ran through. If the vendor cannot tell you the spend, they cannot tell you the true cost per lead. If the spend ran through the vendor's account instead of the contractor's, you never see the real numbers and the brand equity is theirs, not yours.
- All-in cost per lead. Ad spend plus per-lead fees plus amortized setup, divided by leads. Not the ad-platform number.
- Booked appointment rate. What percentage of leads turned into a sit? This is the single most predictive number in the whole document.
- Close rate and average ticket. Which gives you cost per acquired job, which is the only figure your P&L reacts to.
How do you turn a vendor's numbers into your own math?
Do not accept their conclusion. Run their inputs through your own formulas.
All-in CPL = (ad spend + lead fees + amortized setup) ÷ leads delivered
CPA = all-in CPL ÷ your close rate
Max CPL = (average job value × gross margin × close rate) ÷ 3
Work an example. Say a vendor shows 400 leads over six months on $24,000 of ad spend, with per-lead fees of $95. All-in cost is $24,000 + $38,000 = $62,000, or $155 per lead. At a 22 percent close rate that is a $705 cost per acquired job. On a $12,000 average roofing ticket at 40 percent gross margin, that job produces $4,800 of gross profit. A $705 CPA against $4,800 of gross profit is a business that works.
Now run the same math on a cheaper channel. A $45 shared lead closing at 6 percent is a $750 CPA — worse than the $155 exclusive lead, on a smaller average ticket, after your reps chased four other contractors' phone calls. That inversion is the whole argument, and it is laid out in detail in exclusive vs. shared leads.
| Channel | Typical cost per lead | Typical close rate | Cost per acquired job |
|---|---|---|---|
| ChannelReferrals | Typical cost per lead$0 | Typical close rate40-60% | Cost per acquired jobNear $0 |
| ChannelOrganic / Google Business Profile | Typical cost per lead$0-50 | Typical close rate20-40% | Cost per acquired job$100-250 |
| ChannelShared marketplace leads | Typical cost per lead$25-75 | Typical close rate2-8% | Cost per acquired job$900-3,000+ |
| ChannelMeta advertising (self-run) | Typical cost per lead$25-110 | Typical close rate10-20% | Cost per acquired job$250-1,100 |
| ChannelExclusive leads (performance partner) | Typical cost per lead$50-200 | Typical close rate15-25% | Cost per acquired job$200-1,000 |
| ChannelExclusive booked appointments | Typical cost per lead$200-400 | Typical close rate25-40% | Cost per acquired job$200-1,200 |
If you want the arithmetic done for you against your own ticket and close rate, run it through the lead ROI calculator before the vendor call, not after. Walking in with your maximum cost per lead already calculated changes the entire conversation.
What does an honest anonymized case study look like?
Anonymization is not a red flag. Most contractors do not want a competitor two markets over knowing who runs their advertising, and any vendor who publishes client names without permission is telling you something about how they treat client data.
What matters is whether the anonymized version is still specific. "A roofing company saw amazing growth" is not specific. These are:
- Roofing, Wisconsin, across roofing, siding, windows and decking: 676 leads and 488 booked appointments — 72 percent of leads booked.
- Roofing, Pennsylvania and Florida on one system: 675 leads and 462 booked appointments — 68 percent.
- Windows and siding, Nebraska, pay-per-lead: 468 leads since late February, with 225 homeowners self-booking their appointment online and the client's own team booking the rest by phone.
Each one carries a trade, a state, a model, a raw count, and a rate you can divide yourself. That is the standard. Ask any vendor to hand you their numbers in that shape. If they can only produce a percentage, ask for the denominator, and watch what happens.
What questions should you ask on the vetting call?
Write these down and work through them in order.
- Whose ad account does the spend run through — mine or yours?
- Who owns the pixel, the audiences and the retargeting lists when we part ways?
- Are the leads exclusive to me, in writing, and how do you enforce that?
- What is the dispute process for a bad lead, and what percentage of disputes get approved?
- What was your worst-performing account in the last twelve months, and what went wrong?
- What is your client retention past six months?
- Can I see the delivery platform live on a screen share instead of a PDF?
- Can I speak to two current clients in my trade?
- What do you need from me to make this work, and what happens if I do not do it?
- What is the contract length, and what does exit look like?
Question five is the most useful one on the list. Every vendor has a failed account. A vendor who claims otherwise is either new or not telling the truth, and a vendor who can explain the failure in operational terms — the contractor took four hours to call leads back, the service area was too small to spend into, the ticket could not support the CPA — is showing you how they think.
The broader framework for evaluating any marketing relationship, including retainer agencies, is in our guide on how to hire a marketing partner.
What are the red flags that mean walk away?
| Red flag | What it usually means |
|---|---|
| Red flagGuaranteed close rate or guaranteed ROI | What it usually meansThey do not control your sales process and are promising something they cannot deliver |
| Red flagRefuses to run through your ad account | What it usually meansYou never see the real spend, and the brand equity accrues to them |
| Red flagNo written exclusivity policy | What it usually meansThe lead is being sold more than once |
| Red flagNo dispute process | What it usually meansBad leads are your problem and their revenue |
| Red flagLead counts with no date range | What it usually meansThe number is padded across years or accounts |
| Red flagContract longer than the proof cycle | What it usually meansThey need 12 months of your money to survive a 90-day failure |
| Red flagCase studies only in unrelated trades | What it usually meansUntested economics in your ticket range |
| Red flagWill not connect you with a client | What it usually meansThe clients are not happy, or do not exist |
One more that gets missed: a vendor who never asks about your capacity. If nobody asks how many crews you run, how fast you answer the phone, or what your sales team looks like, they are selling volume rather than outcomes. Lead flow you cannot work is worse than no lead flow, because you pay for it either way.
How do you protect yourself in the first 90 days?
Agree on the scorecard before launch, in writing, so nobody argues about the definition of success in week ten.
| Window | What you measure | What good looks like |
|---|---|---|
| WindowDays 1-14 | What you measureLaunch speed, creative approval, lead delivery working | What good looks likeCampaigns live, leads landing in your CRM in real time |
| WindowDays 15-45 | What you measureLead volume vs. forecast, contact rate, dispute rate | What good looks likeVolume within range, disputes handled without a fight |
| WindowDays 46-90 | What you measureBooked appointment rate, close rate, CPA vs. max CPL | What good looks likeCPA under your calculated ceiling, appointments sitting |
Hold up your end too. Speed to lead is on you, not the vendor. A five-minute callback converts dramatically better than a thirty-minute one, and "bad leads" are very often just leads that sat for two hours while somebody finished a roof.
What if a vendor has nothing in your trade?
Smaller verticals genuinely have thinner proof. That is not automatically a no. But raise your requirements to compensate:
- Full numbers from the closest adjacent trade, with ticket and close rate attached.
- A written 90-day scorecard agreed before launch.
- A defined dispute process with a stated turnaround.
- No contract term that extends past the point you would know it is failing.
- Campaigns in your own ad account, so if you leave, the data leaves with you.
The short version
- A case study without a date range, a spend figure and a close rate is a screenshot.
- Ask for the denominator on every number you are shown.
- Cost per acquired job is the verdict. Cost per lead is a talking point.
- Anonymized is fine. Vague is not.
- The best vetting question is "what was your worst account last year, and why?"
- Agree on the 90-day scorecard before launch, not during the argument.
If you want to see live numbers instead of a deck — real accounts, real dates, real cost per lead in the trades we run — book a call and we will walk the platform on screen. Bring your average ticket, your close rate and your target CPA, and we will run the math together in front of you.
Frequently asked questions
What should a contractor lead generation case study include?
At minimum: trade, state or market type, a date range, total leads delivered, ad spend over that window, whose ad account it ran through, booked appointment rate, close rate, and average ticket. Without a date range and a spend figure you cannot calculate cost per acquired job, which is the only number that matters.
Are anonymized case studies a red flag?
No. Most contractors do not want competitors in neighboring markets knowing who runs their advertising, so vertical-plus-state anonymization is normal and honest. What matters is whether the numbers behind the anonymized study are specific and whether the vendor will put you on a reference call with an actual client.
How do I verify a lead vendor's numbers are real?
Ask for a live screen share of the delivery platform instead of a PDF, ask whose ad account the spend ran through, ask for two reference contractors in your trade and actually call them, and ask what the worst-performing account of the last year looked like and why.
What if a vendor has no case studies in my trade?
That is not automatically disqualifying, especially in smaller verticals. Require adjacent-trade proof with full numbers, a written 90-day scorecard you both agree to before launch, a defined lead dispute process, and no long-term contract locking you in past the point you would know it is failing.
Should I trust a guaranteed close rate or guaranteed ROI?
No. A vendor controls lead volume and lead quality; you control speed to lead, your sales process, your pricing, and your ticket. Any vendor guaranteeing a close rate or a revenue figure is either guessing or selling you something they cannot deliver.
Frequently asked questions
What should a contractor lead generation case study include?
At minimum: trade, state or market type, a date range, total leads delivered, ad spend over that window, whose ad account it ran through, booked appointment rate, close rate, and average ticket. Without a date range and a spend figure you cannot calculate cost per acquired job, which is the only number that matters.
Are anonymized case studies a red flag?
No. Most contractors do not want competitors in neighboring markets knowing who runs their advertising, so vertical-plus-state anonymization is normal and honest. What matters is whether the numbers behind the anonymized study are specific and whether the vendor will put you on a reference call with an actual client.
How do I verify a lead vendor's numbers are real?
Ask for a live screen share of the delivery platform instead of a PDF, ask whose ad account the spend ran through, ask for two reference contractors in your trade and actually call them, and ask what the worst-performing account of the last year looked like and why.
What if a vendor has no case studies in my trade?
That is not automatically disqualifying, especially in smaller verticals. Require adjacent-trade proof with full numbers, a written 90-day scorecard you both agree to before launch, a defined lead dispute process, and no long-term contract locking you in past the point you would know it is failing.
Should I trust a guaranteed close rate or guaranteed ROI?
No. A vendor controls lead volume and lead quality; you control speed to lead, your sales process, your pricing, and your ticket. Any vendor guaranteeing a close rate or a revenue figure is either guessing or selling you something they cannot deliver.