Should a seed-stage startup outsource marketing or hire internally?

Staff seed marketing as a portfolio

A seed-stage startup should not outsource marketing as a department, and should not hire a department either. Before users actively recommend the product, founders keep the buyer, the story, and what good looks like. Outside help is one bounded build, a page or a tracking setup, not a channel. Once a few customers are buying and one motion repeats, one generalist owns the brief and one specialist keeps the craft.

Emily Kramer, writing in MKT1, puts generalists on the team and uses agencies as specialists. Contractors on the wrong jobs, and outsiders managing outsiders, are how the work goes wrong.

Work	Who holds it at seed	Hand it over only when
Buyer, story, positioning, motion	Founders. A facilitator may edit.	Never as a whole. A later marketer joins the judgment.
One asset or one channel test	A contractor or a small niche shop	The job repeats weekly and a founder is only coordinating
Paid, specialist SEO, press, brand, burst content	An outside specialist, after the foundation	Rarely. Tactics move fast or the work comes in bursts. A full week is not the trigger.
Weekly plan, brief, edit, kill rule	A founder, then one generalist	That generalist is the bottleneck on a channel that already works

Sort foundation, fuel, and engine before you pick a contract

Keep the foundation inside. Buy fuel only when something can distribute it. Buy distribution only when the fuel and a page exist.

Kramer’s foundation is the audience and competitors, the story’s tenets, basic positioning, and the business model, lifecycle, and go-to-market motion. Handed to an agency, that research often comes back unusable, because the inputs live with the founders. A contractor may facilitate or edit, not own the conclusion. Brand design is the exception, and it lands better after the rest exists.

Fuel is the content. The engine is distribution and conversion. Without an in-house editor and a plan for email, social, paid, or a community, content only checks a box. Search content also needs a keyword list and a page that converts. A paid shop fails without a page, content, and ads you wrote first.

Sam Altman, in a 2016 Y Combinator essay, blocks the engine until people love the product. Otherwise you fill a leaky bucket: users arrive and do not come back, and ads, marketing, and press carry what recommendations should have carried. Raising leaves the leak in place. His signal is that users actively recommend the product to friends. A net promoter score is one measure he names. Early use can still teach you what to build.

Buy a named craft, with a person and an end date

Outsource execution you can describe, date, and grade. Do not outsource the choice of buyer.

Arielle Jackson treats a one-time activity as freelance or agency work: a naming pass or a short paid test. Freelancers who never talk to each other turn the work into a time sink, and it looks uncoordinated. If the need cannot wait, the outsider is only a bridge: a first full-time search takes about three to six months unless the person is a friend or former colleague you already trust.

Kramer caps the bridges. One or two contractors work if you brief, edit, and read the numbers. A crowd does not, and a full-time marketer usually comes before agencies take the website, content, paid, and SEO.

Prefer a contractor or a small niche agency, the first two of Kramer’s five buys (then a small generalist, a large niche shop, and a large generalist). Get names, same-stage clients on your motion (self-serve and top-down differ), references, and a scope, because pitch seniors often do not do the work. Stay inside the craft: brand design is not web design, web design is not web development, ebook and case-study writers are often weak at SEO content and short web copy, and paid managers are often weak at design and copy.

• Paid search and paid social, sometimes two suppliers. Judge the full funnel, not clicks. Do not add design or landing pages. A switch repeats the ramp, so do not switch to escape a bad month.
• SEO is keyword strategy, content, on-site or technical work, or links. Name which. A dedicated SEO hire is rarely among the first 10 to 15 marketing hires, except programmatic SEO. Inbound or the site manages the agency.
• Brand at seed can be a contractor and a templatized site. Around Series A, Kramer places a brand studio, then a separate web studio.
• Burst content turns interviews and webinars you already ran into assets. Give a written brief, edit the first pieces before you trust the voice, and do not let the contractor take the customer conversation.
• Press is a partner once you have a story, not the owner of strategy. A first funding note does not always need a retainer when the story is more than the round: an opinionated origin as a blog post, not a press release, a summary sent as an exclusive, and the post and the exact details held until a reporter agrees.

See a growth marketer versus an agency and when the first growth marketer earns a seat.

Put one generalist inside to direct the specialists

Hire internally when founders are the bottleneck and you can name the problem. The seat directs specialists, not a vice president with nobody to lead.

Jackson times the hire by motion. Marketing-led work, including a consumer brand, hires sooner, near a team of about ten. Enterprise software waits for a repeatable sales motion, then adds marketing when you look for a sales leader. Hardware that needs distribution, packaging, and pricing before launch starts earlier. Content or community can justify the seat while paid spend is small. A few ad tests do not. Seek five to eight years as Head of Marketing or Director, not vice president, and ask when they last did the tactic. Equal strength in product, performance, creative, and communications is rare, so rank one area and rent the rest.

Kramer and Kathleen Estreich call the miss too many firsts: new in at most one or two of hands-on work, strategy, an early company, your model, managing people, and the sub-function you need. A pi-shaped marketer is expert in product marketing, content, or growth, and proficient in a second. The usual first pair is product marketing plus growth. Content plus growth is rarely first.

Maya Spivak wants a T: one deep skill, still doing the work for a few quarters, because the motion can change within about six months. Product-led work needs familiarity with ads, email, social, and search. Sales-led work needs case studies and credible pages. A growth hire needs a site and a signup, and is out if they skipped yours. Founders who already design should hire the analytical gap.

Hold the target, the kill rule, and the readout once a channel repeats. See fractional cover, freelance versus a full-time hire, and expertise without a full-time seat. Where ending employment is slow, including the Netherlands, prove the weekly job on a freelance project before a permanent contract. The timing above assumes you can end the seat quickly.

Run seven checks, then stop

If users do not actively recommend the product, do not staff an engine. If you cannot write the buyer and the story, keep that work and stop at the first miss.

1. No active recommendations from users: a page or a tracking setup can go out. A channel cannot.
2. Cannot write the buyer, story, positioning, and motion: that stays inside. See how a founder stops doing every marketing task.
3. A project ends in an artifact. A job owns next week.
4. One or two projects, and a founder who briefs and edits: a specialist, on a fixed scope, a maximum number of hours, and an end date. Accounts stay in your name.
5. No distribution plan, no content shop. No page, no content, and no ads you wrote, no paid shop.
6. A bottleneck, a named problem, and a non-revenue signal (repeat use, an active recommendation, or a qualified conversation): open the search and cover only the urgent artifact.
7. Rank one area and rent the rest. Hire again only when this marketer blocks a proven channel.

Whoever manages outsiders must run the projects, edit quality, and read the analytics. If nobody holds that trio, add no supplier.

Refuse the contract that hides the work

Walk away if the brief is to do the marketing, if press or a growth owner comes before the story and the signup, or if the accounts live only in the vendor’s login.

• The brief names no buyer and no question.
• Press with no headline or reporter list is Jackson’s cart before the horse.
• Spivak says the growth hire disseminates the site and the signup, and a candidate who never tried your signup is out.
• The proposal names nobody who logs in, or the references are a later stage or another motion.
• The ad account, the site, the CRM, or the prompts sit only in the vendor’s login, or the report stops at clicks.
• The candidate describes a future team, cannot show work they did lately, or expects a large team and a large program you do not have.
• An investor said to hire. Jackson hears that often. It is not a job.

Shut both options in four situations

Skip both when the product is not loved, the buyer has no name, nobody reviews the work each week, or a founder only dislikes the task. Dislike is not a brief: until you have done the work, you cannot tell a weak specialist from a weak offer. A permanent seat is the wrong test of a channel.

The part I will hold

I am Piet Baudoin, one person under the name Poldermarketing, an AI-native growth marketer. The offer is an all-in go-to-market solution for freshly funded startups anywhere: marketing, AI, and automation, fully remote, from the first message to the first customers.

I work in Dutch and English. Hire me for a project or for set days each week. I build and execute. I am strong in AI, content, automation, and workflows. Google Ads and Meta Ads are newer for me: I set them up and review them. A large paid program is the wrong brief.

The free growth scan shows how the site reads now. Positioning and messaging is the job when the sentence is the constraint. How I work is a project or a few days a week.