How much should a startup spend on marketing? The most common answer is a percentage of revenue, and at Seed it's close to useless. Ten percent of very little is very little, and it won't pay for anyone's time. So I'm going to refuse that rule outright and build the answer the way an engineer estimates anything: from what the budget has to cover.
You'll also notice there's no industry spend benchmark in this post. The ones that circulate rarely publish their sample or their period, and a figure I can't check isn't one I'll hand you to plan runway on.
A marketing budget covers three things
Every marketing budget, at any stage, is paying for some mix of these:
Someone's time. Yours, a fractional marketer's, an agency's or a hire's. At Seed this is almost always the biggest line.
Tools. A CRM, an email platform, analytics, a site builder or CMS, maybe an SEO tool. Individually small, collectively sneaky, because they renew whether anyone logs in or not.
Media. Paid ads, sponsorships, events. The one line that can go to zero this month and come back next month.
If it helps, map it onto your engineering costs. Time is your payroll, tools are your SaaS subscriptions, and media is your cloud bill: elastic, useful, and the easiest to overspend on when nobody's watching the dashboard.
Start with time, because it's the line you can't fake
The real decision at Seed isn't how much. It's whose hours. There are four realistic answers.
The founder. Cheapest in cash, most expensive in focus. Fine until it starts eating the hours you need for sales or product.
A fractional marketer or a small agency. Senior hours without a salary, usually month to month.
A full-time hire. The right call once there's a repeatable motion for someone to own.
Nobody, yet. Also a choice, and an honest one while you're still finding product fit.
I've compared the costs of the middle two options in the fractional, agency or hire breakdown, with Canadian salary and payroll figures, so I won't repeat them here. If you're still deciding whether outside help makes sense at all, start with when a startup should hire a marketing agency.
Tools: keep the stack smaller than the team
A rule of thumb from someone who used to build these systems: buy the tool when a process breaks without it, not before. A CRM earns its seat the week a lead falls through the cracks. An email platform earns it when you have a list worth writing to, and the consent to write to it.
That second part matters in Canada. Under Canada's Anti-Spam Legislation (CASL), sending a commercial email means getting consent, identifying who sent it, and including a working unsubscribe (sourced from the CRTC). That's general guidance rather than legal advice, so check anything consequential with counsel. Budget the tool and the consent process together, because the consent process is where startups get caught.
Media: the dial, not the foundation
Paid media is the only line you can turn up or down weekly. Google Ads, for one, has no minimum spending commitment (sourced from Google Ads Help), which makes it tempting to start there. Don't, unless two things are already true: the page you'd send clicks to converts, and someone has time to watch the campaign. Media spent before those two exist mostly buys data nobody has time to read.
A worked example: what each band of spend buys
Here's how the bands look using our own published startup offers, because they're the figures I can stand behind.
One-time, to find out what to do. The Diagnostic is a fixed-scope engagement at $1,500, one time: a written diagnosis and a sequenced 90-day plan. Start a retainer within 30 days and the whole $1,500 comes off month one.
Monthly, to keep one channel working. The Operator retainer is $2,500 a month: positioning and messaging kept current, technical SEO and the content engine, and two researched articles a month.
Monthly, to run marketing as part of the team. The Embedded retainer is $4,500 a month and adds fractional product marketing, managed paid acquisition on LinkedIn and Google, and one end-to-end launch per quarter. Ad spend sits on top, in your own ad account.
Both retainers run month to month, with two months' notice and no annual commitment. For pre-seed companies incorporated in Canada there's also a Founding Marketer arrangement, paid in equity rather than cash, with a six-month minimum and limited places. All of it is laid out on the startup pricing page.
The point isn't that you should buy any of these. It's that a real budget is a scope with a price on it, not a percentage.
A simple way to set your first number
Name the one or two outcomes marketing must produce in the next two quarters. Demos booked, trials started, a launch landed.
Decide whose time will produce them, and price that time honestly, including your own.
Add only the tools those outcomes need.
Hold media at zero until the page and the person exist, then test small, with a stop date.
Check the total against runway, not revenue. If it shortens runway past your next raise or milestone, cut scope rather than quality.
So what do your next two quarters of marketing need to prove, and whose hours are going to prove it?