What kind of content should an early-seed startup create first?

The first asset is the one a live call is missing

The first content an early-seed startup should create, once a live call exists, is the screen a founder can send this week, plus a short note from those conversations. A stranger must be able to say who it is for, the job, how a start works, what differs from the current workaround, and the one next step.

If you are not yet in live sales conversations, do not publish. Keep a shared log of the line, the role, and the workaround until you can name the buyer and send the screen on a call you already have.

Every piece gets one of three jobs. Make the offer legible. Test whether a problem sentence comes back in the buyer’s words. Remove one objection that already showed up on a call. Those three cover a live sales call. If a user can start without one, the job is the first run inside the product. Camille Ricketts starts with what the audience wants to accomplish, and with material that helps them do it. Sonal Chokshi says most of the work is killing what should not ship, and that you measure only what you will change. Sam Altman says growth is not step one. Focusing on growth before people love the product creates a leaky bucket: users come in and do not stay. He says the same for enterprise and hard tech. Content will not patch that. The homepage message is the screen.

Match the asset to the motion you already run

There is no single first asset for every early-seed startup. Ship the piece that removes the block in the motion you already run.

Startupik’s early-stage SEO guide starts most teams with a homepage, product pages, use cases, and comparisons, and tells them to skip a large blog. The same guide says SEO is a weak primary channel when buyers do not search yet, messaging is unstable, the motion is enterprise outbound with custom deals, the product changes weekly, or nobody can maintain the pages and the site. In those cases it points to founder-led sales, partnerships, communities, paid search, or product-led distribution.

Situation	Ship first	Then, only if	Not yet
No live sales conversation yet	A shared log: the line, the role, and the workaround	You can name the buyer and send a screen on a call you have	Any public post, including the screen
No buyers, and you are selling	Screen: who, job, how a start works, the difference, one next step, plus a note with a broken belief	A second company names the same pain	A case study, a public price, a post cluster
Design partners, no result you can publish	Their workflow as a use-case page, labeled as early	They will confirm the situation and the change	An anonymous result you did not measure
Buyers name a tool, a hire, or a spreadsheet	One comparison: when you win, when you lose, who should stay	That alternative showed up in a call or in search	A ten-tool roundup, or a rival you never meet
No shared word for the job	Old job, the break, your approach, who stays with the old way	Calls keep showing the same confusion	A coined name with no contrast
Buyers search a known category	One high-intent use-case or comparison page	Those pages draw the right people, not just visits	Trend posts and an industry glossary
A user can start without a call	The empty state and one guide inside the product	The same stall shows up in support or session notes	An essay about the market

Pier ranks a case study first because it answers whether you have done this for someone like the buyer. Hold that sequence until an account can give a situation, what you did, and an outcome. Pier already names the failure: content aimed at other founders, investors, or the founder’s own network, while the buyer never appears. Define the ICP and apply that test before you pick the row.

A use-case page has six lines or it is a slogan

A use-case page is ready when a stranger can name the buyer, the current workflow, the break, the change, the tools it sits beside, and who it is not for. Write it in the buyer’s verb and end on one next step.

Ali Hajimohamadi cuts any piece a rival could write with no access to your customers, and treats clarity as the early tactic and volume as the later one. Developers get docs and implementation notes. Operators get comparisons and playbooks. Executives get a one-page memo, not a playbook. The founder supplies the lines from the calls. Someone who has not heard those calls does not originate the page.

Founder-led B2B content is the next call, written down

Founder-led B2B content is the sentence the founder will reuse on the next call: the problem in the buyer’s words, and later a proof or comparison a champion can forward.

Gallium publishes native first, reads whether the right roles reply, then turns the winning theme into a page. Add help when that signal repeats, not when the founder is tired of posting. Contentify says stay where that audience already gathers, in a format the founder can sustain, and refuse a podcast, a video show, a short-video app, and a blog at once. Before product-market fit, a piece you already send is a listening device. The surface can be paper.

One source, two surfaces: a page, and the note the founder pastes. If recent buyers did not find you on a network, do not open one. Count named-role replies, lines that land on the page, and meetings that cite the note, not likes or pageviews. Founder-led GTM before a full team is the loop. Content that supports outbound is the note.

A pain becomes a page only after a second buyer names it

A customer pain becomes GTM content when a second buyer, at a different company, describes the same workaround in their own verb. One vivid quote is a hypothesis, not a page.

1. Log the line, the role, and the workaround. Do not translate it into your product noun.
2. A second company must name that workaround without your wording. One design partner can justify a labeled page about their workflow, not a market claim.
3. Write the break, the fix people try first, and how to tell the cases apart. Skip a definition article.
4. Put the line on the homepage or the use-case page the same week. A post that never changes the page is not go-to-market content.
5. Send it on the calls you already have. If the next buyer says that is not them, kill it or rewrite the buyer.

Do not invent a time cost, a failure rate, or a calculator you did not measure. Founder interviews capture the line. Testing positioning angles separates a repeated pain from a phrase you prefer.

Teach the category as a contrast, not a new noun

To help prospects understand a new category, publish one contrast: the job they already recognize, where the current workaround breaks, your approach in one sentence, and who should stay with the old way.

April Dunford treats positioning as context, not a vision or a tagline. Start with what the customer would do if you did not exist, including nothing, a spreadsheet, or a hire, and do not position against rivals you never meet in a deal. Choose the category last, as the frame that makes the value obvious. A frame you invent must be taught before it can orient anyone.

Ship one contrast page and one worked workflow, with no figure you did not measure. Map three of your terms to words they already use, and put the “what do you mean” answers on that page. If an existing category already makes the value obvious, skip the new noun.

Run one month, then retire what nobody sent

The first month finds which single asset enters a real conversation. Keep that asset and retire the rest.

1. Week one. The screen a stranger can use: who it is for, the job, how a start works, the difference from the workaround, and one next step, plus one founder note that includes a belief the interviews broke. No second channel.
2. Week two. One problem or use-case page in the logged words. Send it on every live call. Record the sentence people repeat and the one they reject.
3. Week three. A comparison, a proof note, or a short FAQ only if that question already appeared, including how a start works. Publish a price only when you will keep it.
4. Week four. A second channel waits until the same buyer asks the next question.

Do not scale until a few pages draw the right people.

These formats wait until a buyer asks

Leave a broad blog, a podcast, a brand film, company news, a gated paper, and a comparison with rivals you have never met until a buyer asks the question that asset answers.

A build-in-public post has to prove a pattern, a refusal, or a lesson. A launch line does not.

Stop when the product changes every week, or when nobody will answer a reply. Do not publish a new page to fill that gap.

The first set I would ship

For an early-seed company I would not open a content program. I would ship the screen from calls you already have, and one page a buyer can be sent, with a loop that turns the next call into the next edit.

I am Piet Baudoin, one person under the name Poldermarketing. I work fully remote, in Dutch and English, and you can hire me freelance or for a few days a week. I am an AI-native marketer who builds and executes. I am strong in AI, content, automation, and building workflows. Google Ads and Meta Ads are newer for me. I can set them up and review them. I do not run them at scale.

One freelance marketer can cover the marketing, the AI, and the automation for a freshly funded startup anywhere, from the first message to the first customers, without a separate specialist for each. If the site cannot say who it is for, the free growth scan shows how it reads. Positioning and messaging is where the sentence starts. How I work is the shape of an engagement.