How much should a small business spend on marketing?

A percentage of revenue answers the wrong question

January, 9 p.m., at the kitchen table. You're planning the year and the same answer comes up everywhere: put 5 to 10% of your revenue into marketing. On $900,000 in sales, that's $45,000 to $90,000. Simple, comforting, and upside down.

A percentage of revenue looks backwards. It tells a business that's doing well to spend more, and one that's short on customers to spend less, at the very moment it needs them. And it ignores the only two things that matter: how many new customers you need, and what one customer actually leaves you.

Look at where the rule comes from, too. BDC presents the 2-to-5% range (B2B) and 5-to-10% range (B2C) as a common rule of thumb, with no study behind it. A US Small Business Administration blog post from 2019 quotes a 2018 agency survey: 7.9% of revenue on average. And a figure you'll often see quoted, 7.7% in 2025 from Gartner's CMO Spend Survey, comes from 402 marketing leaders, the vast majority at companies with more than $1 billion in annual revenue. An average is a snapshot of your neighbour, not a plan.

Our position: a small business should build its marketing budget backwards. Start with the customers you need to win, multiply by what it costs to win one, check that the margin on the first job can carry it, then add a test reserve and your own hours. The percentage is, at best, a sanity check afterwards.

The backwards budget, in four lines

Four questions, in this order.

1. Customers to win. How many new customers do you need over 12 months, beyond the ones who come back and the ones referrals already bring? Start from your capacity: a crew to keep busy, empty appointment slots.
2. Cost per customer. Cost per enquiry ÷ the share of enquiries that become customers. With no history, write down a cautious assumption, in pencil.
3. The ceiling. The gross margin on the first job: the price minus direct costs (materials, labour, subcontractors). If the cost per customer goes over it, either the channel or the goal doesn't work. Our guide to customer acquisition cost walks through this test.
4. The budget. Customers × cost per customer, plus a test reserve of 15%, plus your hours at an hourly rate.

The test reserve is our house rule, not an industry standard: money set aside up front so you can try an idea without raiding what already works. Your hours count just like the ads: three hours a week at $50 is 3 × $50 × 52 = $7,800 a year, invoice or not.

One last constraint: fund one channel enough to learn before you open a second. For search advertising like Google Ads, BDC's Jessica Horvath recommends planning for at least $1,000 a month. Four channels at $250 each are four samples too small to decide anything. If that amount is out of reach, see what your first $500 on Google Ads can still prove.

A worked example: a renovation company in Laval

Take a bathroom renovation company in Laval, just north of Montreal: six people, $900,000 in revenue. All the numbers below are assumptions for the example.

• Customers to win: 20 new jobs a year, on top of referrals.
• Cost per customer: through Google Ads, $120 per enquiry and one enquiry in five signing, so $120 ÷ 0.20 = $600.
• Ceiling: an average job of $18,000 at a 30% gross margin, so $5,400. The cost per customer takes 11% of it ($600 ÷ $5,400): the channel passes easily.
• Budget: 20 × $600 = $12,000 in ads, plus a 15% test reserve ($1,800), for $13,800 in cash. On top of that, the owner's three hours a week: $7,800 in time.

The $900,000 in revenue never enters the calculation. The 5-to-10% rule would have suggested three to six and a half times what the goal actually requires. And judge the result on margin, not sales: "$13,800 for $360,000 in jobs" (20 × $18,000) flatters everything. What the business keeps is 20 × $5,400 = $108,000 in gross margin, before your hours and overhead.

Sometimes the math pushes the other way. A young physiotherapy clinic in Gatineau with $180,000 in revenue wants to fill a second therapist's schedule: 200 new patients. Assumptions: $40 per enquiry, one enquiry in two booking, so $80 per new patient, against a $300 gross margin on a first episode of care. Budget: 200 × $80 = $16,000, plus a $2,400 reserve, for $18,400.

Example	5-to-10% rule	Backwards budget (cash)
Renovation, Laval, $900,000 in revenue	$45,000 to $90,000	$13,800 (1.5% of revenue)
Clinic, Gatineau, $180,000 in revenue	$9,000 to $18,000	$18,400 (10.2% of revenue)

The rule tells the established business to overspend and the growing one to underspend. The backwards budget follows the goal.

Quebec changes the calendar, not the formula

Don't divide the budget by 12: follow the months when your customers decide, not the months when you do the work. In renovation, people often ask for quotes months before the job, and Quebec's construction holiday hollows out part of the summer. In Laval, $12,000 over eight active months is $1,500 a month. Google Ads works with an average daily budget: a campaign can spend up to twice that amount on a given day, but no more than 30.4 times it over the month. So set $1,500 ÷ 30.4, or about $49 a day.

If you serve customers in English, plan the French first. According to Quebec's language office, the Office québécois de la langue française (page in French), commercial advertising must be in French, and another language can appear alongside it as long as French is clearly predominant. In your budget, English is an extra line (copywriting, visuals, sometimes a separate campaign), never a substitute.

On the tax side, advertising is generally deductible, with exceptions: the Canada Revenue Agency only lets you deduct 50% of an ad in some periodicals, and nothing for advertising aimed mainly at the Canadian market that runs with a foreign broadcaster. For Google or Meta in your situation, ask your accountant. Deductible means cheaper, not free.

What AI shifts in your budget

BDC sees it coming: with AI becoming more common, marketing spend will probably decrease or be reallocated. For a small business, the line that moves most is your hours. A first draft of a post, ad variations, a reply to a review: two hours saved a week are worth 2 × 52 × $50 = $5,200 a year at the example's rate. Put them where they pay: calling enquiries back faster often beats one more post.

The other shift is in how people find you. More and more customers put their question to ChatGPT or Google and read an answer before they click. What feeds those answers is mostly what you already own: your Google Business Profile, your reviews, pages that answer real questions. This week, ask two assistants your customer's exact question ("best bathroom renovator in Laval") and note whether you show up.

When the math isn't enough

• A new business with no data: your cost per customer is a guess. Make the first three months a measuring period, with a single channel funded at the threshold, then redo the math with real numbers. If the total is more than you can afford, lower the number of customers you're aiming for, not the cost per customer you're hoping for.
• Cash flow: the math tells you what a goal costs, not whether you have the money. Paying for ads in March for jobs you get paid for in June is financing. Talk to your accountant or your banker.
• What doesn't count in customers: a sign, a sponsored minor hockey team. Give them a small fixed line, without pretending to measure them.

Build your backwards budget in one evening

1. List your new customers from the last 12 months and where each one came from.
2. Work out the gross margin on a typical first job, with your accountant if needed.
3. Set the number of customers to win based on your capacity, not your wishes.
4. Write down a cost per customer: your own data if you have it, a cautious assumption if not.
5. Run the math: customers × cost, plus a 15% reserve, plus your hours. Compare it with the ceiling.
6. Pick one channel and set its daily budget (monthly amount ÷ 30.4) for the months when your customers decide.
7. Book a review every quarter, as BDC recommends: real cost per customer, real jobs, new assumption.

If your total lands far from 5 to 10%, good: your budget is finally about your business.