Poldermarketing
How can a founder stop doing all the marketing themselves?
A founder stops doing all the marketing by handing one repeating, teachable job to a single owner, while staying in customer conversations and on the message. Write down who buys and what good looks like. Move production, routing, and reporting first. Do not hire in order to disappear, and do not stay the approver of every line.
Stop the production. Keep the learning.
Hand one repeating job to one owner, and stay in the conversations and on the message.
YC’s essential advice: do work that does not scale, talk to users, and grow the team only after people want the product. Use one or two metrics. Ansa, 16 April 2026: founders most likely to blow up that hire did it to escape the work. See when to hire the first growth marketer.
Hand off a motion only after it has repeated
Hand over when the same action has won several buyers, a colleague can explain the yes, follow-up waits on you, and the week shows qualified conversations, a trusted conversion, or revenue tied to the work. Hold for one post, a shifting offer, vanity counts, or a dated surge.
Breakthrough3x: stress is not the rule, and there is no universal revenue or headcount line. Have source materials, access, context, and feedback, so the hire is not left to repair every growth problem. Jackson: about 10 people for a consumer brand or a marketing-led path; for enterprise SaaS, after a repeatable sales model, around the sales-leader search. This is not a law for a freelancer. Ansa wants some closed business for most B2B companies.
Stop being the bottleneck
You stop being the bottleneck when routine marketing no longer waits on your approval, your memory, or your hands.
Breakthrough3x calls it routine decisions, approvals, or knowledge that must pass through you before work can move. For one week, sort each item: you alone decide, someone else decides inside limits, it belongs in a procedure, or it stops. Count routine decisions still reaching you, approval wait, repeated questions, and milestones that slip while you are away. Muthukumar calls opens, visits, and post engagement vanity. Prefer revenue tied to the work, or product usage.
Sort every task before you name a seat
Keep what needs your network, your judgment, or your name. Give away repeating work that still runs if you are offline for a day.
Muthukumar: closing most conversations means a demand hire; mixed comprehension means product marketing first. Spivak: a growth hire does not build the site. Have a decent page and an easy signup already. If those people do not convert, the growth hire runs experiments to raise conversion. Ansa: demand generation comes after a stable product, buyer, and claim. Hire for demand first only when you already own the story, the message is stable, one channel already works, and pipeline is visible. Extra traffic on a confused offer brings unqualified conversations. Jackson: press, before the headline and who would care, is a match error.
Whatever the seat, if the break is the page, onboarding, or follow-up, fix that before you buy more traffic.
One-page brief, with decision rights
Put decision rights on one page, taken from one recurring task you did in the last two weeks. Name who decides, what they flag, what you still approve, and how fast you reply.
- Who you want, and the problem you solve.
- Phrases from their calls, not a competitor homepage.
- How they find you now.
- Two or three recordings or pieces to repeat.
- Claims you will not use.
- Proof, no customer name: a buyer sentence and a product fact.
- When your name, face, or a quotation may be used.
- Where accounts, files, and prompts live, with a second person who can open them.
Lead lists and call recordings are personal data. An outside party needs a written processor arrangement: documented instructions, confidentiality, and return or deletion when the work ends. Do not use a personal account. Have someone qualified check that, and the employment status if the seat looks like a job. This page is not that check.
Delegate without a full-time team
A founder delegates marketing without a full-time team by giving one person one outcome, and buying one skill you already understand. See more marketing without a full team.
Jackson: a defined project goes to a freelancer or an agency. A full-time search takes about three to six months unless they are a friend or former colleague you already worked with closely. If it cannot wait for that search, use a freelancer or an agency as a stop-gap. About five freelancers who do not talk to each other put you back in the middle. About five to eight years, still doing the work, as head of marketing, not an early VP. Spivak wants hands-on work for at least a few quarters. Muthukumar wants a T-shaped person, deep in product marketing or demand generation. The diagnosis can change within about three months. Ansa: a senior marketer covers more writing, research, and reporting, so a manager of agencies is a weaker fit. A fractional lead before a head of marketing sets the order. Content or community can justify a full-time person before much paid media. A small paid test, by itself, does not.
| Seat | Fits when | Wrong when |
|---|---|---|
| Freelancer | One skill you already understand | Finding the buyer, or you still coordinate them |
| Hands-on generalist | A repeated motion, you are out of hours, and they own shipping plus one number | No recent hands-on work |
| Channel specialist | You have seen that channel work twice | The channel is a guess |
| Part-time lead plus executor | The gap is the order, and someone else ships | A title, and nobody logs in |
| Marketing leader | A clear motion, a team to direct, and real tradeoffs | No team and no clear motion |
Judge a specialist on judgment and the business tie, a generalist on writing and ordering the work, and a leader on tradeoffs and developing people. Skip the 30/60/90-day plan. Whiteboard the funnel and see if you learned something. End the interview if a growth candidate has not tried the signup. For the writing, ask who it was for, what changed, and why.
Build a process that does not depend on you
A marketing process does not depend on the founder when every repeating motion has one owner and a written path, and a normal week ships while you are away.
Breakthrough3x says skip the large manual. Name the trigger, the owner, the inputs, the steps, the quality bar, and the escalation. The person doing the work keeps it current. See a repeatable marketing process for a small team. A tool cannot name the owner.
Automate the tasks you still do by hand
Automate a founder task only after you have run it by hand and the step is the same every time. One workflow first. See how to automate repetitive marketing work.
| Move | First candidates |
|---|---|
| Automate | Form to the company list, a confirmation, a follow-up that stops on a reply |
| Stop | A channel you have not seen work twice |
Give it a trigger, actions, an output, and a stop. It does not choose the claim, stop a test, or take the call. Breakthrough3x: name capacity, authority, or a missing procedure before you add software.
Let AI remove work, not add a queue
AI reduces a founder’s marketing work when a draft ships under rules you already wrote. A draft that waits on you lengthens the queue. Use it without losing the founder’s voice.
McKinsey, June 2026: 90 percent of CMOs are experimenting with AI, and less than 10 percent have scaled it or captured value across the workflow. The miss is a bolt-on approach: point solutions. Only 28 percent of the surveyed organizations are rewiring teams and workflows. On your team, saved time counts only when the piece goes live under the one-pager.
A model may draft from a call. You still choose the claim, the offer, and the call. A model does not publish a result, a customer, or a legal claim.
Move the work in three passes
Breakthrough hands work over in three passes: recurring execution, then choices inside your limits, then you as a planned input.
| Pass | They own | You still touch | Done when |
|---|---|---|---|
| Execution | Calendar, scheduling, routing, numbers | One-pager and exceptions | A normal week ships without a rewrite |
| Choices | Tests, a decision log, one new channel with an end date | Forbidden claims, and when a test stops | You read the log, not every thread |
| Input | Cadence and the number | Calls, your words, launches, partnerships, the hire | The work survives a week away |
Each week: customer calls, one anchor you record or write and they cut into posts, pages, and emails, one sync, and one pass on the number. Meet on what shipped, what customers did, what you learned, and what is next. The first handoff is one named outcome for a set period, such as the next month. Do not open the seat if you cannot keep the person and the tools through a quiet period.
Teach pass one by watching, then beside you, then alone. At about 90 days, Ansa wants a clear ICP, a positioning or messaging document, and one channel experiment with an early read. Missing those, check access to you and to customers before you replace the person. Muthukumar: month one is a funnel note, real calls, and a launch.
When to take the job back
If the number moves the wrong way, name the cause first: strategy, execution, resources, measurement, or normal market variation. Offer and follow-up are company problems, not that diagnosis. Score shipping, message quality, the agreed number, their decisions, and what stops when you are away. Agree before the bad week what triggers coaching, a process change, or a temporary return. Take the whole job back only when the motion was never real, or the person still cannot ship inside the rules after that chance. Narrow the seat when one person is asked to fix the offer, follow-up, and product, or when you still rewrite every line.
What I take off your plate
I run one recurring task from the last two weeks against that brief, and review exceptions weekly. Accounts and prompts stay in your name.
I am Piet Baudoin, one person, Poldermarketing, fully remote, in Dutch and English. For a freshly funded startup anywhere, I am one freelance marketer for marketing, AI, and automation, from the first message to the first customers, for a bounded job or a few days a week, without a separate specialist for every part. I am an AI-native growth marketer, strong in AI, content, automation, and workflows, and I build and run the work. Google Ads and Meta Ads are newer: I set them up and review them. A large paid program is the wrong job.
The free growth scan is the first look at the site. How I work is the engagement. Positioning and messaging is where I start when the offer is hard to say.
Questions people ask
Should I leave marketing once someone else owns it?
No. Routine work should keep moving when you are away. You still take customer conversations, you still decide which claims the company will stand behind, and you still appear where your name is the reason someone listens. If prospects start using words you would refuse, you left the story too early. If every asset still waits for your evening, you handed over the title and kept the job.
What should I hand over first?
Start with one recurring task from the last two weeks that already leads to a conversation, a signup, or a sale. Write how you do it, what good looks like, and which exceptions come back to you. Editing, page builds, scheduling, reporting, and lead routing are the usual first moves. Keep who you sell to, a major launch, and any channel you have not seen work twice.
Can an agency replace me as the marketer?
An agency fits when you already know the buyer and you need several defined channels run at once. It is a poor fit when the job is to discover the market or to sit in your sales calls. You still own the message and the weekly review. A single skill belongs with a freelancer. Work that spans the message, the tests, and the coordination belongs with one hands-on person.
How do I know the handoff is working?
Judge it by whether a normal week ships without you rewriting it, and by whether the number you agreed has moved: qualified conversations, a conversion you trust, or revenue you can tie to the work. Likes, opens, and raw visits are weak proof. If the week still waits on your rewrite, the handoff is not done.
When is it too early to stop doing the marketing myself?
It is too early when you cannot say who buys and why they say yes, when a single lucky post is the only proof, or when the offer still changes after every call. It is also too early when you want marketing to hide a product people leave, or a sales follow-up that nobody owns. Talk to users and do the manual version until one motion has worked more than once. Then hand off the part you can teach.