Blog

Marketing agency for startups: what to buy at pre-seed and what to buy at seed

By · - A pre-seed cheque and a seed round buy different kinds of marketing help. What each stage can afford, the Canadian programs that won't pay for it, and when an agency is the wrong call.

A small terracotta pot with a single sprouting seedling beside a larger pot holding a leafy young plant on a linen tablecloth, with a brass trowel and a plain kraft seed packet

The short answer

At pre-seed, the founders should be doing most of the selling themselves, and an agency's job, if you hire one at all, is narrow: make the material that founder-led sales runs on, and build outbound that the founders still close. At seed, you have enough money and enough evidence to hand one channel to someone else and run it every week, usually before you hire a head of marketing. The mistake at both stages is the same one. Buying a broad retainer before you know which channel works.

What each stage can actually spend

The two rounds are now different in size by more than an order of magnitude, and that alone decides most of this.

Carta, the cap-table platform, counted $3.19 billion raised across more than 11,500 pre-seed SAFEs and convertible notes in the second quarter of 2026, with the average instrument at a record $276,000. Pre-seed is often several of those stacked together, but plenty of companies start on one. At seed, Carta's benchmarks put the median round at $4.5 million on a $23.9 million post-money valuation for the same quarter.

Now put a marketing line against each. A $2,000 monthly retainer for twelve months is $24,000. Against a $276,000 cheque that's 8.7 percent of all the money you have, money that would otherwise pay for months of a founder's salary or an engineer. Against a $4.5 million seed it's about half a percent. Carta's figures are US dollars and our fees are Canadian, which only makes our share smaller, but the point holds either way: at pre-seed every marketing dollar competes with runway, and at seed it mostly competes with other marketing dollars.

Pre-seed vs seed, side by side

Pre-seedSeed
What's usually trueA product, a handful of users or pilots, no repeatable way to find morePaying customers, a sales motion that works when the founder runs it
Who sellsThe founders, by handStill mostly the founders, with help on the top of the funnel
What to buySales material, a site that explains the product, outbound infrastructureOne channel run every week, with one number everyone agrees on
What not to buy yetPaid ads at scale, a rebrand, a full-service retainerFour channels at once, or a senior marketing hire before a channel works
A sensible marketing spendProject fees, or one line at the low endOne or two monthly lines, plus ad budget if ads are the channel

The table is a pattern, not a law. A pre-seed company selling to consumers through ads may need paid acquisition before a B2B company would. But if your startup sells to businesses, the left column is close to universal.

Pre-seed: the founders are the marketing

Paul Graham's essay on this is still the clearest thing written about it: "The most common unscalable thing founders have to do at the start is to recruit users manually." The Stripe founders didn't wait for beta sign-ups; they installed the product on people's laptops on the spot. That isn't a job you can hand to an agency, and an agency that offers to take it over at this stage is selling you the wrong thing.

What an outside team can usefully do at pre-seed is make the founders' hours go further:

  • The material a sales call needs. A one-page explainer, a short product video, a deck for the customer rather than the investor, two or three answers to the objections you hear every week
  • A website that says what the product does in the first sentence. Most pre-seed sites are written for investors and read like it
  • Outbound infrastructure. A separate sending domain, warmed up properly, a list of the companies you should be talking to, and emails a person reads before they go. The founder still takes the meetings

All three can be bought as projects. None of them needs a twelve-month contract.

Seed: hand over one channel, not the whole function

By seed, something works when the founder does it: a LinkedIn post that brings demos, a webinar that converts, a cold email that gets replies. The job now is to make that thing happen every week without the founder's calendar being the bottleneck.

This is where an agency usually beats a first marketing hire. A senior marketer at seed will spend their first quarter learning the product and their second deciding which channel to bet on. An agency running the channel that already works can start the week it's briefed, and if it doesn't move the number in a quarter, you stop paying. Hire the head of marketing once you know which channel they'll be scaling. If you need senior judgment in between, a fractional CMO is the usual answer, and it's a different purchase from an agency.

Our work with VisaFlo is the version of this we can show. VisaFlo, in its third year and based in Burnaby, sells software to Canadian immigration consultants. When we started, more than 80 percent of its customers stayed and 40 to 50 percent of demos turned into customers. Selling worked. Content was the bottleneck. We ran one bet: two kinds of content, two to three posts a week, judged by one number, demos that came from content. Over ten weeks, according to the case study, weekly sales calls went from 3 to as many as 16 and monthly recurring revenue doubled. VisaFlo's team did the selling, and a new landing page and outreach ran alongside, so we don't claim the whole result. The order of operations is the lesson: the product and the sales motion came first, the weekly channel second.

The Canadian money that won't pay for it

Canadian founders often assume government programs will offset marketing. Mostly they won't, and it's better to know before you budget.

SR&ED doesn't cover marketing. The definition in subsection 248(1) of the Income Tax Act, as the CRA reproduces it, excludes "market research or sales promotion" outright. Your engineering work may qualify; the go-to-market around it doesn't.

CanExport SMEs is narrower than it sounds. The program pays up to 50 percent of eligible costs on international expansion projects, with requests of $10,000 to $50,000. But the 2026-27 applicant's guide requires at least three full-time employees and $300,000 in revenue declared in Canada, which rules out most pre-seed companies. And the marketing it will fund is mostly translation, pitch decks and trade-show material for the target market. Online and social advertising, search optimisation, blogs, webinars, case studies, email campaigns, branding and CRM software are all listed as ineligible.

Neither of those is a reason not to apply for what you qualify for. They're a reason not to plan your marketing around a refund that isn't coming.

What it costs with us

We sell lines one at a time, with prices on the pricing page. For a startup, two of them carry most of the weight.

  • B2B Sales, $2,000 setup, then $1,000/mo. 250 companies picked and contacted each month by email and a LinkedIn connection. AI drafts each email and a person reads every one before it sends; replies come back sorted, and your founders take the meetings. Nothing sends in month one, because the sending domain needs three weeks to warm up. Year one is $14,000, about 5 percent of an average pre-seed cheque
  • Social, $1,000/mo. Nine pieces a month, each approved by you before it posts. This is the line that fits a founder whose LinkedIn already brings demos and who has stopped having time to post
  • Search & GEO and Paid Growth, $2,000/mo each. Usually seed-stage purchases. Search takes about 90 days to move. Ads need a budget of their own, paid from your account, and below about $1,000 a month a campaign never learns

Two honest limits. If you're pre-seed and haven't yet sold the product by hand to ten customers, don't hire us or anyone like us; spend the money on the founders' time, and come back when you know what a good customer looks like. And if what you need is a single launch video or a pitch deck, a good freelancer will beat a monthly line on price. That's a real answer. The agency, freelancer and DIY comparison lays out the trade.

Sources

Bring the stage you're at

Book a free consultation and tell us what's working in your sales today; we'll say which one line we'd run first, or that it's too early for any. What a marketing agency does all month and the questions to ask before you sign are the two closest reads.