THE LIBRARY · A PLAIN GUIDE
How much should a small business spend on marketing?
Ask around about a marketing budget, and someone will have a number ready for you: 7 to 8 percent of revenue. It usually comes with "the SBA says" in front of it, and it's said with the kind of confidence that ends the conversation. So let's start by checking what the SBA actually says.
Its own page on the subject says "there's no hard and fast answer to how much your marketing budget should be." Then it quotes other people's averages from 2018:
- 7.9 percent of revenue on marketing, from one survey
- 1.08 percent of revenue on advertising, from another
- 11.8 percent for consumer services and 6.9 percent for business services
The page is from 2019, and it never recommends a number of its own. Somewhere along the way, those borrowed averages turned into a rule.
If you're setting next year's budget, or wondering whether this year's was too much, the rule won't get you far. By the end of this article, you'll have a budget built from three numbers you already have, and a way to tell by the end of the quarter whether it was right.
Where do the percentages come from?
Two big surveys land in the same range as the rule, and both are of companies nothing like yours.
Gartner's 2026 CMO Spend Survey found marketing budgets at "7.8% of company revenue in 2026 from 7.7% in 2025." It surveyed 401 marketing leaders, "with the vast majority of respondents reporting annual revenue of over $1 billion."
The CMO Survey, run by Duke's business school with Deloitte and the American Marketing Association, went to marketing leaders at U.S. companies in January 2026. Of the 308 who responded, about half answered the question of what share of revenue goes to marketing. The mean was 8.96 percent. The median was 5. About two-thirds of the companies in the survey had revenue above $100 million, and 13 percent were under $10 million.
So what do you take from that? Two things.
First, those figures describe companies with marketing departments, agencies of record and a CFO who asks about the number in a budget meeting.
Second, and more useful, the median is far below the mean. Half of those companies spend 5 percent or less. A handful spend a lot more (the highest answer was 43 percent), and they pull the average up. So when someone tells you to spend 8 percent, they're quoting the average of companies dozens of times your size, often a hundred or more.
Put the three sources side by side and look at the last column, because that's the part the rule always leaves out:
- Median: 5% of revenue
- Mean: 9% of revenue
The scale runs to 43%, the highest answer.
| Source | What it measured | Figure | Who answered |
|---|---|---|---|
| SBA blog, 2019 | Other surveys' 2018 averages | 7.9% of revenue on marketing; 1.08% on advertising | Not stated; secondhand |
| Gartner, 2026 | Marketing budget as a share of revenue | 7.8% | 401 marketing leaders, mostly at companies over $1 billion |
| The CMO Survey, 2026 | Marketing expenses as a share of revenue | Mean 8.96%, median 5% | 308 senior marketers at U.S. companies, two-thirds over $100 million (154 answered this question) |
SBA blog, 2019
- What it measured
- Other surveys' 2018 averages
- Figure
- 7.9% of revenue on marketing; 1.08% on advertising
- Who answered
- Not stated; secondhand
Gartner, 2026
- What it measured
- Marketing budget as a share of revenue
- Figure
- 7.8%
- Who answered
- 401 marketing leaders, mostly at companies over $1 billion
The CMO Survey, 2026
- What it measured
- Marketing expenses as a share of revenue
- Figure
- Mean 8.96%, median 5%
- Who answered
- 308 senior marketers at U.S. companies, two-thirds over $100 million (154 answered this question)
So the percentages are real. They're just real for someone else. For a business at $800,000 or $3 million, 8 percent of revenue is a number with no reasoning behind it. It might be right. It would be right by accident.
What's wrong with a percentage of revenue?
It looks backward. Last year's revenue is the result of last year's marketing (and the weather, and a competitor closing, and a referral from a customer's brother-in-law). Setting next year's spend as a slice of it means a bad year buys less marketing, which is the opposite of what a bad year usually needs.
It also ignores what the money is for. Two businesses with identical revenue can need very different budgets: one turns work away every week and would waste most of what it spent, and the other has a crew idle two days out of five. The percentage can't tell them apart. The jobs can.
And it hides the two kinds of money inside every marketing budget: the fee you pay people to do the work, and the spend that goes to Google or Meta to buy attention. A percentage lumps them together. So you can't see that the fee stayed flat while the ad spend doubled, or the reverse.
What should you use instead?
Three numbers, all from your own books, and about ten minutes.
What a new customer is worth. Not the first ticket; the first year. A plumber's average job might be a few hundred dollars, but the customer who calls twice a year and sends a neighbor is worth several of those. A restaurant's new regular is worth a year of Tuesdays. Write down your real figure for the first twelve months of a typical new customer, after the cost of serving them.
What you can afford to pay for one. A share of that first-year value that leaves the job worth doing. If a customer is worth $1,200 in the first year at a decent margin, paying $150 to $250 to get one is a good trade and paying $800 is not. This is your maximum cost per new customer, and it's the number that tells you whether any given spend is working, long before a dashboard does.
How many more you could take. Marketing that works brings calls. If the calls can't be answered, the jobs can't be scheduled or the tables are already full, the money is wasted at the moment it succeeds. Count the customers you could actually serve next quarter beyond the ones you already have. If the answer is "none until we hire," the budget goes to hiring first.
Now multiply what you can afford to pay for a customer by how many more you could take. That's your marketing budget for the quarter, with a reason behind every dollar. Twenty more customers you could serve, at $200 each, is $4,000 for the quarter, fee and ad spend together. If a percentage of revenue gives you a different number, you now know which one to trust.
Your budget for the quarter
- $200 × 20 more customers = $4,000
That figure is a ceiling, though, so start below it. The first months of any new marketing tend to cost the most per customer, while the ads gather data and the pages are new. If the work is right, that cost comes down. If it isn't, it goes up, and you'll want to catch that early. We'll get to how.
What counts as marketing spend?
Before that number can mean anything, decide what goes in it. The percentages above don't say, and it's easy to count it differently from one year to the next. Here are the four lines we'd use:
- Fees. What you pay people to do the work: an agency, a freelancer, a part-time marketer's wages. The published ranges for the agency line are in our piece on what an agency does in a month.
- Spend. What goes to Google, Meta, a directory or a printer to buy attention. Keep this on its own line, always, so you can see it move on its own.
- Tools. Email software, the website's hosting and upkeep, call tracking, a scheduling link. Each one's small, but together they can run a few hundred a month.
- The website. A build is a one-time cost, and it's marketing, because every other line sends people to it. Spread it over the years it will last rather than counting it all in one.
Then there's a fifth line nobody writes down: your own time. Three evenings a week on posts and ad settings comes to about 40 hours a month. At $75 an hour, that's $3,000 of your time, which is more than some agencies charge to do the work. Count it at whatever an hour of yours is worth, and doing it yourself stops looking free.
| This quarter | Next quarter | |
|---|---|---|
| Fees | ||
| Spend | ||
| Tools | ||
| The website (per year) | ||
| Your time |
What should the budget buy first?
Not ads. Ads are the easiest thing to spend on, and the fastest way to pay twice, because they send people to whatever's already there. If the website doesn't turn visitors into calls, every click is a cost with no customer at the end of it. Our piece on whether your website needs a redesign is where to check.
If you're starting from not much, here's the order we'd spend it in:
- The free things, done properly. A Google Business Profile that's current, a review request after every good job, and a short email to past customers every month or so. These cost time, not money. How to keep the Profile current and how to ask for the review are each their own piece.
- The website that converts. The pages that answer what customers search for, with the phone number where a thumb can reach it. A one-time cost, mostly.
- Then paid traffic, sized by the most you can afford to pay for a customer, and tested for a fixed stretch before anyone commits to a year of it. Our piece on what Google Ads management costs walks through sizing the spend, including Google's rule that lets a busy day run up to double your daily budget.
- Then help, in whatever form the gap takes: an agency for the work, or a senior person for the direction. Our pieces on choosing an agency and what a fractional CMO costs are the two halves of that decision.
If the budget is small, spend it on the free things and the website, and nothing else until they're done. Five hundred dollars a month in ads, pointed at a site that loses people, buys a lot of visits and very few customers.
How do you know the number was right?
This is the early check we mentioned. At the end of the quarter, look at what each new customer cost you, and hold it up against the most you said you'd pay. That's the whole test. Not impressions, not traffic, not followers, and not the percentage of revenue, which will look fine whatever happens.
If the cost is under your maximum and you can take more work, spend more. The budget was too low.
If it's over, stop and find out why before you spend another month. Check the site, the phone and the offer before you blame the budget.
And if you can't work out the cost per customer at all, because nobody's tracking which calls came from where, fix that first. It's cheaper than anything else on this page.
Then do it all again next quarter, from the same three numbers, and let the budget move. Remember the SBA page from the top, the one that said there's no hard and fast answer? It was right. The answer changes every quarter, and a budget rebuilt from what a customer is worth and what one cost you last time keeps up with it. A percentage set in January doesn't.
Where we fit
If you'd rather not work out those numbers alone, that's what the Marketing Plan is for. It takes two weeks and costs $1,500. You get a written plan built from how work reaches you today: three ranked moves, each with its cost, the number it has to hit and the point at which you'd stop. (That last part is the end-of-quarter check above, written down before anything starts.)
The plan is yours either way, and it's written so you can carry it out yourself. If you go on to Marketing Monthly, Fractional CMO or an Ads Launch with us within 60 days, the $1,500 comes off.
Not sure the Plan is the right first step? The fit call is 15 minutes. We'll tell you whether it is, or whether you don't need us at all.
Not sure where to start? That’s what the call is for.
Fifteen minutes about your business and how work reaches you today. We’ll tell you whether the Marketing Plan is the right first step, or whether you don’t need us at all.
