Key takeaways
- There is no industry number worth using. Your floor is the leads it takes you to book one job, times what a lead costs you, times enough repeats that the result is not luck.
- Measuring whether ads caused a booked job is harder than it looks. According to Liu, Bettaney and Chamberlain (2018), because "Facebook lift studies differ from standard A/B tests, the online experimentation literature does not describe how to calculate parameters such as power and minimum sample size," and their paper exists to supply those calculations.
- Remodeling closes slowly. Money spent in March books work in April or May, so judging a budget on a calendar month reads the result before it exists.
Most pages that answer this question give you a monthly budget. Those numbers usually come from an agency describing its own clients, which makes each one a number for somebody else's business in somebody else's market.
Kenneth Media is a DFW marketing agency that runs paid ads, builds websites and wires up follow-up for contractors and home-service businesses. The budget is the wrong place to start. Start with two numbers you already have, and the budget falls out of them.
How much should a remodeler spend on Facebook ads? Start with your own numbers
Start with your close rate. Count how many new leads it took to book your last job. If ten conversations produced one signed kitchen, ten is your number. Then estimate what a lead costs to buy. Multiply the two and you have the cost of one booked job. That is your floor for a single job, and a floor is not a budget.
Here is the arithmetic, in order:
- Leads per booked job. Go back through the last ten jobs you signed and count how many new inquiries each one took. Use the real number, not the flattering one.
- Cost per lead. If you have run ads before, you already know this. If you have not, the first month of spending is how you find out, and that is the honest purpose of a first budget.
- Cost per booked job. Multiply the first two.
- Multiply by three. One booked job proves nothing. Three gives you a pattern you can act on without fooling yourself.
- Check it against the job. If the cost of buying one booked job is a meaningful share of what that job earns you, ads are not your problem. Your close rate or your average job size is.
That last step is the one people skip, and it is the one that decides whether to spend at all.
Want help running these numbers on your business? On a free 30-minute strategy call we will do exactly this with your figures, whether or not you hire us.
How long before the number means anything?
Longer than a two week test. An ad account needs enough conversions for the platform to find the people worth showing your ad to, and you need enough booked jobs for the result to be a pattern rather than a run of luck. Both take real volume, and volume takes both budget and time.
This is not a marketing opinion. It is a measurement problem that researchers have written about directly. According to Liu, Bettaney and Chamberlain (2018), Facebook lift studies differ from standard A/B tests in ways the existing experimentation literature did not cover, and the authors set out to provide "the statistical power and required sample size calculation" for them. The practical translation for a remodeler: whether your ads caused a booked job is a question with a required sample size, and a handful of leads is well below it.
So when someone tells you to "test it for two weeks and see," what they are really proposing is to collect too little data and then make a large decision with it. If the budget you can commit only buys a handful of leads, the honest move is to wait until you can fund a real test rather than run a small one and conclude that Facebook does not work.
Leads or booked jobs: what should I actually count?
Booked jobs, and it is not close. Leads are the cheapest number to move and the easiest to fake. An ad set optimized for form fills will happily deliver people who fill out forms and never answer the phone, and the report will look great while your calendar stays empty.
| What you can count | What it tells you | Why it misleads | |---|---|---| | Impressions and reach | How many times the ad was served | Reach is not interest | | Clicks | That the creative got attention | Curiosity clicks cost the same as real ones | | Leads and form fills | That the offer was worth an email address | A form fill is not a conversation | | Booked appointments | That somebody let you into their house | The first number that correlates with revenue | | Signed jobs | The only number that pays anyone | Lags the spend by weeks |
The fix is to send outcomes back into the account. When you record which leads became appointments and which appointments became signed work, you can judge the ads on the last two rows instead of the first three. That is a CRM job more than an ads job, and it is covered on our CRM systems page.
What should I fix before I raise the budget?
Speed of reply, before anything else. Ads that bought leads nobody reached in time did not fail; the follow-up did. A lead that fills out a form at nine at night and hears back at noon the next day has probably already talked to two other companies.
Three things to have in place before more money goes in:
- Someone answers, or something does. Missed calls during a job are normal for a remodeler. A missed call with no callback is a lost job. A voice AI agent or a missed-call text back closes that gap.
- The landing page answers the question the ad raised. Google's own guidance on helpful content says its systems look for signs of experience, expertise, authoritativeness and trust, and that "Of these aspects, trust is most important." A homeowner clicking an ad applies the same test in about five seconds.
- Every lead lands in one place. If leads live in a phone, an inbox and a notepad, you cannot tell which ones stalled, and you cannot tell anyone which ads worked.
Ads pour water into the bucket. Fixing the holes first is cheaper than buying more water.
What is different about running remodeling ads in DFW?
Three things. The service radius is real, the competition is uneven town to town, and the buying cycle is long. A remodeler in Plano reaching Allen, Frisco and McKinney is bidding against different companies in each of those towns, and a single campaign covering the whole metro averages three very different markets into one confusing number.
The long cycle matters more than most people expect. A homeowner planning a kitchen talks to family, collects bids and waits for a season that suits them. The ad that started the conversation ran weeks before the contract was signed, which is why a month of spend judged at the end of that month usually looks worse than it was.
So judge remodeling ads in this market over a full buying cycle, counted in booked jobs, with tracking on every lead so you can tell which ads produced work rather than which ones produced forms. The arithmetic is the same on Google and on Facebook: what a lead costs, how many leads a booked job takes, and enough of both to see a pattern. Six weeks is a short window for that. Six days is noise.
Run your numbers before you run the ads
On a free 30-minute strategy call we will look at where your leads come from today, what a booked job costs you, and where follow-up slips, then work out your actual floor together. You will leave with a clear next step whether or not you hire us. If ads are the right move, paid advertising is what we do all day. Prefer the phone? Call (214) 425-1670.