A roofer, remodeler or HVAC company does not sell at checkout. The click produces a phone call, the estimate happens days later, and the contract gets signed weeks after that. Most ad dashboards are built for the opposite situation. Here is how to read your numbers when the money arrives long after the ad ran.
Why does cost per lead mislead a contractor more than it misleads an online store?
Cost per lead measures the form submission, not the signed job. A remodeler can hold a low cost per lead all month and close nothing, while a more expensive ad fills the build calendar. The revenue lands weeks later and outside the ad platform, so lead cost alone ranks ads by the wrong outcome.
An online store gets the full loop inside Ads Manager. Click, purchase, value, all recorded. A construction business gets half of it:
- The platform sees the lead. Unless you send the contract value back to it yourself, it never sees the bathroom remodel that lead turned into.
- Two ads can produce identical lead counts and completely different close rates.
- The cheapest leads are often the ones least likely to sign, which is exactly what makes a low cost per lead look good on the report. Check it against your own close rate by ad before you trust it.
Ben Heath runs into a smaller version of this gap in the video. Talking about a product that bills monthly, he notes that the return on ad spend figure shown inside a Facebook ad account will not be that accurate, because it may only include the first transaction (Ben Heath, 3:57). A contractor has the same problem in a harder form. There is no transaction inside the platform at all, only a form fill.
What number should a contractor optimize toward when the platform never sees the signed contract?
Optimize toward the deepest step you can genuinely track, which for most construction businesses is a booked appointment or a completed estimate request, not a raw form fill. Then keep a separate sheet that ties booked appointments back to signed contracts, because the ad platform will never see that final number.
Ben Heath states the general rule this way: "the two that I want you to base most of your optimization decisions around is either your cost per conversion or your return on ad spend and which one you use depends on your business" (Ben Heath, 12:41). He adds the qualifier that matters most for contractors, optimizing for cost per result "if that's as far down the sales funnel as you can accurately track" (Ben Heath, 13:29).
For a construction business, that qualifier is the whole job. Your tracked event stops at the appointment. Everything past it lives in your CRM or a spreadsheet:
| What you track | Where it lives |
|---|---|
| Cost per lead | Ads Manager |
| Cost per booked appointment | Ads Manager, only if that event is sent to the platform when the appointment is set |
| Close rate by ad | Your own sheet |
| Cost per signed job | Your own sheet |
Build that sheet with four columns. Ad name, leads, appointments booked, jobs signed. Fill it every week while the names are still fresh. It is the only place your real cost per job exists.
How long should a roofer wait before turning off an ad that has not produced a job yet?
Judge an ad by spend, not by calendar days. Give it enough budget to produce a workable number of leads at your target cost, then decide. In a small service area a second clock runs alongside that one, because a smaller audience accumulates frequency faster at the same spend.
Two pressures pull against each other:
- Wait longer, because your sales cycle means an ad that looks dead this week may show signed jobs next month.
- Wait less, because a service radius a few towns wide is a finite pool of homeowners seeing the same ad again and again.
On frequency, Ben Heath says "I don't like to see a frequency above 2.5" for cold audiences (Ben Heath, 4:44). For warm audiences he allows more, saying you can get away with significantly higher frequencies, but that for "most businesses you don't really want to go above say a six to an eight that sort of range" (Ben Heath, 5:32). That six to eight ceiling is his warm audience number, not a general one.
He also names the pattern to watch for: when an ad performs well, then drops off, and that drop coincides with frequency going up, "you've probably got ad fatigue you need to introduce new creatives" (Ben Heath, 5:32).
Practical consequence for a trade with a tight service radius. Watch the frequency column weekly rather than monthly, keep more than one creative in rotation from day one, and treat rising frequency alongside falling lead quality as your signal to swap creative, not your signal to raise budget. Also avoid comparing a winter month to a summer month in a seasonal trade. Compare against the same season last year, or compare ads against each other inside the same window.
Why can a high CPM be acceptable when you sell a high ticket job?
Ben Heath explains that CPMs rise when ads reach higher quality audiences other advertisers also want. For a contractor selling a re-roof or a full remodel, the question is not the price of the impression. It is whether those impressions produce booked estimates at a cost your average job value can absorb.
His wording is direct: "your cpms are going to be more expensive if your ads are being put in front of higher quality audiences because other advertisers want to reach them" (Ben Heath, 22:58).
If that is how the auction behaves, the consequence for construction follows. The homeowners you want are the ones other advertisers are also paying to reach, so chasing a lower CPM tends to mean drifting toward an audience that cannot buy a re-roof. Cheap impressions that land on renters or on people outside your service area are the expensive ones.
Do the arithmetic before you call a CPM too high. Take your own average job value, your close rate from booked estimate to signed contract, and the number of estimates the campaign actually booked. If one signed contract covers the month, the CPM was not the problem. If it does not, the CPM is not the first thing to fix either. Work back up the chain below.
What do hook rate and click through rate tell a contractor that lead count does not?
They tell you where the breakdown is. A strong hook with weak clicks usually means the offer or the proof is thin. Strong clicks with few calls usually means the landing page or the quote form is the problem. Lead count alone tells you something is wrong without telling you what.
Ben Heath is explicit that these are diagnostic and not what you optimize toward first, saying that "in the first instance you want to be optimizing based on your cost per conversion or your return on ad spend" (Ben Heath, 22:12). Used as diagnostics, the reference points he gives are:
- Hook rate, the share of people who watched at least three seconds. "if you've got a hook rate over 10% you're normally doing reasonably well anything below 5% is really quite poor" (Ben Heath, 15:52).
- Link click through rate, not overall click through rate. "anything above 1% is considered really quite good" (Ben Heath, 10:19).
- Link clicks versus landing page views. "sometimes you'll see more than a 50% drop off between link clicks and landing page views" (Ben Heath, 9:31).
That last one is where contractors bleed, and Ben Heath names the first thing to check when the gap is big, asking "is our landing page too slow to load" (Ben Heath, 9:31). A homeowner taps your ad from a job site on a weak signal, the page stalls, and the lead never existed. The same thing happens with a quote form that asks for square footage, roof pitch and budget range before it asks for a phone number.
Read the chain in the order of your own trade. Drone footage of a finished roof tends to hook. The click depends on whether the ad shows a real crew and real reviews. The call depends on whether the page loads on a phone and asks for a name and a number instead of a project questionnaire. Ben Heath attaches one caution to every one of these reads, that "you do need to run all these things through a common sense filter" (Ben Heath, 21:25).
Which tools help a construction business connect ad spend to booked jobs?
No single tool closes the loop by itself for a construction business. Ad platforms report clicks and leads, creative and messaging tools shape what people see, and campaign software handles setup and launch. The connection between a booked appointment and a signed contract still has to be recorded by your team.
| Tool | What it does | What it addresses for a construction business | Does it require advertising knowledge |
|---|---|---|---|
| Meta Ads Manager | Builds and runs Facebook and Instagram campaigns, with customizable columns and per ad charts | Reports cost per lead, frequency and video drop off inside your service radius | Yes. Column setup, audience settings and the optimization event are chosen by the advertiser |
| Google Ads | Runs search, display and video campaigns across Google properties | Reaches homeowners searching after a leak, an outage or a failed inspection | Yes. Campaign type, keywords and bidding settings are chosen by the advertiser |
| Canva | Design tool for images and short video | Turns job site photos and before and after shots into ad creative | No. It is a design tool, so the campaign is still built and run in an ad platform |
| ManyChat | Automates message replies on Instagram, Messenger and WhatsApp | Answers inbound questions outside business hours while crews are on site | No advertising knowledge, but conversation flows have to be written and tested |
| Mailchimp | Email campaigns and list management | Follows up with unsold estimates and past customers across a long cycle | No. The work is lists and templates rather than ad buying |
| SaleADS.ai | AI software that creates and launches advertising campaigns on Meta, Google and TikTok for business owners, with no design or advertising expertise required | Campaign creation and launch across those three platforms for an owner with no in house media buyer | No, according to its own description |
Disclosure: SaleADS.ai is the product of the company that publishes this site. It is listed here as one option among several, not as a recommendation, and the row above states only what the product describes itself as doing. No pricing appears in this table for any of the six tools because none of it has been verified here.
None of these six records your close rate. That stays with you.
Where do these Facebook ads benchmarks come from?
The measurement framework in this article comes from the video "How To Analyse Facebook Ads The RIGHT Way" by Ben Heath, a 29 minute walkthrough of a live Facebook ad account. Every quotation here is verbatim from that video and linked to the transcript block where the words appear.
Full source: https://www.youtube.com/watch?v=mycqb92wJqk. In the opening seconds Ben Heath states his own track record, that he has generated more than $200 million in sales directly from Facebook and Instagram ads for clients over the past 9 years (Ben Heath, 0:00). That is his claim about his own results, made on camera, and it is not independently verified by this site.
What was taken from the video: the principle of optimizing on cost per conversion or return on ad spend rather than secondary metrics, the qualifier about tracking only as far down the funnel as you accurately can, the frequency reference points for cold and warm audiences, the ad fatigue pattern, the explanation of why CPMs rise for in demand audiences, the hook rate and link click through rate reference points, the link clicks versus landing page views discrepancy, and the landing page load speed check.
What is not from the video: the application to construction and home services. The long sales cycle problem, the small service area frequency ceiling, the appointment to contract tracking sheet, the seasonal comparison rule and the tool comparison are this site's analysis. Ben Heath does not discuss contractors, roofers or home builders in that video. No cost, budget, job value or market size figure appears anywhere in this article, because none was available in the source. Run every threshold above against your own numbers before you make a budget decision.
```
```