The order is the strategy
Startups rarely fail at marketing because they picked the wrong channel. They fail because they ran a channel that only works at a later stage, concluded it does not work, and moved on.
Paid acquisition needs a funnel that converts, which needs a product that retains, which needs users you got some other way. SEO compounds over quarters and pays nothing in month one. Neither is wrong; both are wrong first.
The sequence that survives contact with reality: earn a small number of users by hand, learn what they actually say about the product, turn that language into positioning, then put that positioning in front of people at scale. Most teams skip to step four.
What works at what stage
| Stage | What actually works | Why |
|---|---|---|
| Pre-product-market fit | Direct outreach, communities, hand-won users | You are buying information. Scale at this stage only obscures the signal. |
| Early traction | Founder-led content, launches, design partners | You now have something specific to say. The founder is the most credible channel you have. |
| Repeatable sales | Amplified launches, content that compounds, partnerships | Positioning is proven; the job becomes reach. |
| Scaling | Paid, SEO at volume, lifecycle, category building | The funnel converts well enough that buying traffic is rational. |
The first ninety days
A startup marketing plan that fits on one page beats a document nobody opens twice. Ninety days is enough to run the sequence above once, properly.
- Weeks 1 to 4. Thirty conversations with people who look like buyers, before any deck or campaign. Write down the sentence each one uses to describe the problem they have.
- Weeks 5 to 8. Rewrite the site and the one-liner in that borrowed language. Ship one piece of founder-written content a week, on the single platform where those thirty people actually are.
- Weeks 9 to 12. Pick a moment worth amplifying: a release, a raise, a number you can prove. Put reach behind it deliberately rather than posting it and hoping.
At the end you know where customers came from, what they call the product, and whether reach or retention is the thing holding you back. That is a marketing strategy for a startup. Every channel decision follows from those three answers.
What to ignore early
- A content calendar. Publishing on schedule before you know what resonates produces volume without producing distribution.
- Brand guidelines. A tone-of-voice document is a solution to a consistency problem you do not have at five people.
- Paid, before retention. Buying traffic into a product people leave is the fastest way to spend a seed round on a lesson you could have had free.
- Being on every platform. Effort split five ways produces a slope nothing notices. One channel where your buyer actually is beats five where they might be.
Measuring without volume
Conversion rates are meaningless at low numbers, and most startup dashboards are theatre. Three things are worth tracking early:
- Where the last ten customers came from, specifically. Not the channel, the actual post, person, or conversation.
- What they say the product does, in their words. That sentence is your positioning; you are just transcribing it.
- Whether week-two retention is moving. Nothing else matters if it is not.
Once those three are stable, reach becomes the constraint, and that is a solvable problem rather than a discovery one.
Agency, in-house, or the founder
The hiring question usually arrives a month after a raise, and for most seed-stage companies the honest answer is none of the three yet. A first marketing hire inherits whatever positioning exists, so if that positioning is still a guess, you have hired someone to scale a guess.
- The founder, before product-market fit. The conversations that produce positioning cannot be delegated without losing the only useful output of the stage.
- An agency, for a motion with a date on it. A launch, a funding announcement, a category push. Agencies are bought and measured on an outcome, which makes them a good fit for a deadline and a poor one for open-ended discovery.
- In-house, once the motion repeats. The second time you run the same play, the knowledge is worth keeping in the building.
Marketing agencies for startups sell four fairly different things under one label, and the distance between what a founder thinks they are buying and what the agency does is where most of the disappointment lives.
What launch agencies actually do covers the four, and how to choose a marketing agency has the questions that separate an operator from a reseller.
Questions we get asked
What should a startup do first in marketing?
Win a small number of users by hand through direct outreach and communities, then learn how those users describe the product in their own words. That language becomes the positioning. Scaling a channel before positioning is proven amplifies a message you have not yet validated.
When should a startup start paid acquisition?
Once the product retains and the funnel converts. Paid acquisition multiplies whatever your funnel already does, so buying traffic into a product people leave converts a budget into a lesson you could have learned for free.
How many marketing channels should an early startup run?
One, done properly, until it is working. Effort split across several channels produces a gentle slope that ranking and recommendation systems do not react to, whereas concentration produces the acceleration they are built to detect.
Should a startup hire a marketing agency or build in-house?
An agency suits a specific motion with a date attached, such as a launch or a funding announcement, because that is what agencies are bought and measured on. In-house suits a motion you have already run once and intend to keep running. Before product-market fit the founder does it, since the conversations that produce positioning cannot be delegated without losing the output.
How much should an early startup spend on marketing?
Before positioning is proven, close to nothing, because spend multiplies a message you have not validated yet. Once it is proven, the useful question is what an amplified moment costs rather than what a monthly budget should be, since reach concentrated on a launch behaves differently from reach spread thinly across a quarter.