Paid ads
How much should a small business spend on ads?
Do not budget a percentage of revenue. Work backwards from one job: take what a closed customer is worth to you, multiply by the share of quotes you win, and that is the most you can pay for a lead. Multiply by your leads-per-month target and you have a ceiling. Then check it against the floor — a search campaign needs roughly 30 conversions a month before its bidding can learn anything, and under about $1,000 a month most local campaigns never get there.
The short version
- Two numbers set the whole budget: what a customer is worth, and what share of quotes you close.
- There is a floor as well as a ceiling. Spread too thin across too many keywords and nothing works.
- Track booked jobs, not clicks or form fills. Optimising to the wrong event is how accounts quietly fail.
- Ad platform automation needs conversion volume. Below it, the automation makes things worse.
How to do it
Find what one customer is actually worth
Average job value times gross margin, then times how many times a typical customer comes back. A furnace tune-up worth $120 once is a different business from one that becomes a $9,000 replacement in three years.
Find your close rate on quoted work
Of the people you actually quote, what share hire you? One in three is common in the trades. This number is the difference between a profitable account and an expensive one, and most owners guess it wrong on the high side.
Multiply to get your maximum cost per lead
Customer value times close rate equals what a lead is worth. Decide what share of that you are willing to pay — half is a reasonable starting point — and that is your target cost per lead.
Check the floor before you commit
Target cost per lead times leads needed per month gives your budget. If that number is under about $1,000, narrow the campaign until the budget is concentrated enough to produce steady conversions rather than a trickle.
Make sure the platform is counting the right thing
If a phone call that books a job and a form fill from a job-seeker both count as one conversion, the bidding will happily buy more job-seekers. Feed back the events that represent money.
Why the percentage-of-revenue rule fails small businesses
The usual advice is to spend some share of revenue on marketing. It is a useful benchmark for a company with an established mix and a finance department, and it is close to useless for a five-person shop deciding whether to turn on Google Ads next month.
It fails because it anchors on what you currently make rather than on what an additional customer is worth. Two businesses with identical revenue can have wildly different correct budgets — one sells a $200 one-off, the other a $14,000 project with a referral tail. The percentage rule gives them the same answer.
Work from unit economics. It gives a number you can defend, and it tells you immediately whether the channel can work at all.
The floor nobody mentions
Every conversation about ad budget is about the maximum. The more common failure is the minimum.
Modern search bidding is automated, and automation needs examples. Roughly thirty conversions a month is the point where the system has enough signal to distinguish a good click from a bad one. Below that it is guessing, and it will spend your money learning things it never quite learns.
A small budget spread across forty keywords produces a handful of clicks on each and conversions on none. The same budget aimed at four terms where someone is ready to hire produces enough volume to optimise. Narrow beats broad, and it is not close.
Optimise to money, not to activity
The single most common way a local ads account fails is that it works perfectly toward the wrong goal.
If every form submission counts the same, the platform will find you the cheapest form submissions — which are the ones from people who were never going to buy. Cost per lead falls, the report looks great, and the phone stops ringing with real work.
The fix is to send back what actually happened. A call over a minute long. A form from the service area with a real job description. A booked appointment. When the platform is fed outcomes instead of actions, it starts buying the traffic that produces them.
Be careful with the automation upsells
Ad platforms release features that promise more conversions for the same money, and the sales pitch always arrives with a vendor-published number attached.
Google claims about a 7% conversion lift from the full AI Max for Search suite. Independent testing found 84% of advertisers saw neutral or negative account-level outcomes. Both can be true: features that need volume to work will help large accounts and quietly hurt small ones.
The rule for a small local account: do not enable anything that widens targeting until your conversion tracking is provably correct and you are consistently above the volume floor. Broad automation on a thin account spends the budget on discovery you cannot afford.
| Input | Example | Where yours comes from |
|---|---|---|
| Average job value | $4,200 | Last 50 invoices |
| Gross margin | 45% | Job costing, not revenue |
| Repeat / referral multiple | 1.4× | How often a customer returns or refers |
| Customer value | $2,646 | job × margin × repeat |
| Close rate on quotes | 30% | Quotes sent vs jobs won |
| Value of one lead | $794 | customer value × close rate |
| Target cost per lead (50%) | $397 | Your call on how much to keep |
| Leads needed per month | 8 | Capacity, not ambition |
| Monthly ad budget | $3,176 | cost per lead × leads needed |
Illustrative arithmetic using example inputs, not a Revio benchmark or a promise of results. Substitute your own numbers — the method is the point, and if the answer comes out below roughly $1,000 the campaign needs narrowing before it needs funding.
Common questions
- Is $500 a month enough for Google Ads?
- Only if it is aimed at a very small number of high-intent searches in a tight area. At that level you cannot cover a category, so pick the two or three searches that mean someone is ready to hire today and own those. Spread across a broad keyword set, $500 buys data and no jobs.
- Should I run Meta ads or Google ads?
- Google captures people already looking for what you sell, so it is usually first for services bought at a moment of need. Meta creates demand for things people do not search for, so it is stronger for new offers, events and visual products. If you have budget for one and the phone needs to ring this month, start with search.
- How long before ads pay off?
- Leads inside a week if the tracking and targeting are right. A reliable cost per booked job takes about 60 to 90 days, because that is how long it takes to accumulate enough conversions to separate signal from noise. Judging an account at three weeks is judging noise.
- Who should own the ad account?
- You should. If an agency runs ads in an account they own, you lose the conversion history the day you leave, and that history is most of what makes the account efficient. Give an agency admin access to your account instead.
Where these numbers come from
- Google reports roughly +7% conversions from the full AI Max for Search suite; independent testing found 84% of advertisers saw neutral or negative account-level outcomes.
- — Google product claims vs independent advertiser testing, 2026
- With an AI Overview present, paid click-through rate fell about 68%.
- — Seer Interactive, Jun 2024–Sep 2025
Last reviewed .
Read next
Want this done for your business?
Tell us where you are trying to get to. If the honest answer is that you do not need us, we will say so.