For most small service businesses, the floor that produces usable results is about $35 a day on Meta, roughly $1,050 a month, or about $100 a day on Google, roughly $3,000, on one platform. I start clients at $50 a day or more. Anything much under that usually buys noise instead of data. The percentage-of-revenue rules you read online, spend 5 to 10 percent of revenue, fall apart for a one or two truck operation, because a small slice of a small number is a budget too thin to learn anything. The better way to set your number is to work backward from what one job is worth to you and what you can afford to pay to book it. Spend below the data threshold and you're paying for the platform's education while learning nothing yourself. This piece walks through the minimum spend that produces decision-grade data, how to size a budget off your average job value, what to fix before you spend a dollar, and when to hold off on ads entirely.
Why does the percentage-of-revenue rule mislead small operators?
The percentage rule breaks down at small scale because a percentage of a small revenue number is too little money to buy clear data. A shop doing $300k a year at 8 percent gets about $2,000 a month, which can work. A newer operator doing $80k gets around $530 a month, roughly $17 a day, and that sits under the threshold where a platform can figure out who to show your ads to. The rule was built for bigger companies with predictable baselines and a marketing department to feed. For a small operator, the slice of revenue is the wrong thing to size, so ask whether you can put enough behind ads to get a real read, and whether each job pays back enough to justify the spend in the first place. Job value drives the answer, and a flat percentage ignores it.
What is the minimum ad spend that produces useful data?
The minimum is the point where the platform collects enough leads each week to optimize, which for most local service businesses comes to around $35 a day on Meta and closer to $100 a day on Google, because search clicks cost more and the account needs more of them to learn. Ten dollars a day feels careful, and on Meta or Google it is mostly a donation to the platform. At that level you might get a click or two a day, a handful of leads across a whole month, and no version of your ad ever gets enough traffic to prove itself against another. The algorithm needs a stretch of consistent spend and a few hundred dollars of leads before it learns who to chase. Starve it and it never gets smart, so you pay full price for guesses that never improve. Spend enough to pull in maybe 20 to 50 leads in a month and now you can see which ads book work and which ones burn cash.
How do I set an ad budget from my average job value?
Work backward from what a booked job is worth and what you can afford to pay to win one. Take your average job value, multiply by your profit margin, and that is the most you can spend to land a customer and still come out ahead. Divide that ceiling by how many leads it takes to book one job, and you have your target cost per lead. From there the daily budget just has to clear the data threshold and give the math room to breathe.
A detailing shop makes a good example. Say a paint correction and ceramic coating job runs $1,500 at a 50 percent margin. That shop can afford up to $750 to book the customer and still profit. If one lead in four turns into a booked job, the target sits around $187 per lead. I ran the full version of this math for detailing shops here. A $150 mobile wash has a far tighter ceiling and needs cheaper leads or repeat visits to make ads pay. The formula is the same, and the number comes out very different.
What should I have in place before I spend a dollar on ads?
Before you run ads you need a page built to turn a click into a booking and a way to answer new leads fast. Ads are the tap. With no bucket underneath, most of the water hits the floor. Two things catch it. First, a page with one job, turning the click into a call or a booked appointment, not your homepage with a menu of twelve links pulling people in every direction. Second, an automated way to text and email a new lead back within minutes, around the clock. The research on lead response is blunt: reach someone in the first five minutes and you are many times more likely to book them than if you wait an hour. Most owners lose leads they already paid for by calling back the next afternoon, once the person has already hired whoever answered first. Fix the page and the follow-up before you spend, or you will blame the ads for a leak that sits downstream of them.
When should you not run ads yet?
Hold off on ads when your average job is worth too little to absorb the cost of winning a customer, or when you cannot answer the phone. If a job is worth $80 and you keep $30 of it, no amount of clever targeting makes paid ads math out at current lead prices. If you are already booked solid and cannot take on more work, ads just generate leads you will ghost, which wastes money and dents your reputation with people who might have hired you later. And if there is no dedicated page and nothing catching leads fast, ads are premature. Referrals and word of mouth might be carrying you fine for now, and that is worth protecting. Ads are for when your jobs carry real margin and you have room to take the work. Add a page and follow-up that catch what comes in, and then you are ready to spend.
Is the first month of ads supposed to be profitable?
Usually no, and treating month one like it should be is how owners quit right before it starts working. The first month or two is a testing investment. You are paying for the platform to learn, for your ads to sort winners from losers, and for you to find your real cost per booked job. Early leads cost more because nothing is dialed in yet and the account is still guessing. By month two or three the picture settles: the losing ads get cut, budget flows to what books work, fresh creative goes in before the winners wear out, and the cost per job usually drops and holds. Give ads a 90-day window before you judge them, because two weeks is too short to tell. The operators who win are the ones who spent enough to get a usable read and gave it long enough to harden into an efficient machine.
How I would size your number
If I were setting this with you, I would start from one booked job's worth to you, back into a cost per lead you can live with, and set a daily budget that clears the data threshold on one platform before adding a second. Most small operators do better funding one platform properly than spreading a thin budget across three and starving all of them. The full process I run, from that first number to a settled cost per booked job, lives on my method page.
If you want a straight read on what your budget should be for your specific job value and margin, book a 30-minute call and I will walk the math through with you.
