Marketing Before Product-Market Fit: What to Skip

Almost all the marketing advice a founder reads was written for a company that already knows what it sells and to whom. Build the content engine, set up the attribution, hire the growth person, run the experiments. It is good advice for that company. Applied eighteen months too early, it is an expensive way to look busy while the actual question goes unanswered.

The actual question before product-market fit is narrow: does a specific kind of person have a problem painful enough that they will change what they do to solve it. Marketing before fit exists to answer that faster, not to grow anything. Everything that does not serve the answer can wait.

What actually changes at the moment of fit

Before fit, you are searching. Every conversation is a source of information and the product is still moving. After fit, you are scaling something that already works, and the job becomes finding more of the same person at a lower cost.

Those two modes want opposite things from marketing. Searching wants a small number of high-information conversations, positioning you can change next week, and no infrastructure that would be painful to tear down. Scaling wants repeatability, measurement and systems.

The tell that you have crossed over is usually not a metric. It is that people start describing your product to each other accurately, in words you did not give them, and that they come back without being reminded.

The four things to skip until you are sure

Skip the content engine

Two posts a week aimed at a category you might exit is a large, slow bet on a hypothesis you have not tested. Search takes months to reward you, which means you will be publishing for a positioning you no longer hold by the time it pays.

Write, by all means. Write the three pieces that make your specific argument, the ones you would send to a prospect who is not yet convinced. That is a different activity from a content calendar, and it is useful immediately.

Skip the paid acquisition experiments

Paid channels are a way of buying more of a result you already get. If the conversion from visit to signup to use is not working, paid traffic buys you a more expensive version of the same failure, and it hides the diagnosis behind noise.

There is one exception worth allowing: a very small spend used purely as a message test, where the thing you are buying is not customers but a comparison between two ways of describing what you do.

Skip the full brand build

You need a name, a mark, a colour and a site that does not look abandoned. You do not need a brand system, a tone-of-voice document, or a logo iteration that takes six weeks. Founders reach for the brand project because it is the most enjoyable form of avoiding customer conversations.

Skip the analytics stack

At ten signups a week, no attribution model will tell you anything a phone call would not tell you better, and every event you instrument now describes a product that is about to change. Install one simple analytics tool, watch signups and activation, and get the rest from talking to people.

What to do instead, all of it founder-scale

Every item here is deliberately something that does not scale, because at this stage information density beats reach.

1. Talk to thirty people who have the problem, not thirty people who like your idea. The difference between those two groups is the difference between a company and a hobby. Ask what they do today and what it costs them, not whether they would use your product.
2. Write the one page that explains what you do in the customer's words. If a stranger cannot tell in ten seconds who it is for and what changes, no amount of traffic will help. This page is your positioning made testable.
3. Go where the problem is already being discussed. A community, a forum, a group, an industry association. Be useful there for weeks without pitching, and the pitch stops being necessary.
4. Do the manual version for your first ten customers. Onboard them personally. Do the work by hand behind the scenes if you have to. You are buying the highest-quality product feedback available anywhere.
5. Write down what they say, in their words. The exact sentences customers use become your homepage, your ads and your emails later. Founders who keep this file skip a whole quarter of positioning work.

The one page, and how to tell if it is working

The page needs four things and nothing else: who it is for, what changes for them, some evidence that it is real, and one action. Evidence at this stage is a screenshot, a short demo, or two named customers, not a wall of logos you do not have.

Then use it as an instrument. Send it to ten people who fit the target and ask them to say back what it does. When their description matches yours, the positioning is holding. When it does not, you have learned something more useful than a conversion rate.

Rewrite the page every time your understanding changes, which early on may be monthly. A page that has not changed in six months at this stage usually means nobody has been asked to read it.

Before fit, marketing is a research method that happens to produce a website.

The signals that say you can start building the machine

There is no single threshold, but a few things tend to arrive together, and none of them is traffic.

• People come back without being reminded. Retention is the only number at this stage that cannot be faked by effort.
• They describe the product to others accurately. Word of mouth means the value is legible without you in the room.
• The same objection stops appearing. Objections converging is what fit sounds like from the sales side.
• You can predict who will buy. When you can look at a prospect and call it correctly most of the time, you know the segment well enough to spend money finding more of them.

When those show up, the standard advice becomes correct and you should follow it. Build the content programme, instrument the funnel, hire someone whose job it is. The skipped work was never wrong, it was early.

The trap worth naming

Building marketing infrastructure feels like progress and produces artefacts: a site, a calendar, a dashboard, a plan. Talking to thirty strangers about a problem produces a text file of notes and a lot of uncomfortable moments, and it is the only one of the two that answers the question you are actually stuck on.

So the rule is simple enough to check weekly. If a marketing task would not change what you build or who you sell to, it can wait. If it would, do it today, at the smallest scale that gets you the answer.

Once you do have fit, the mechanics become ordinary, and ordinary is the point: a fixed slot, a small routine, published consistently. That version is in The 30-Minute Marketing Week, and if AI assistants are already part of how your buyers research, AI Search Is Already Reading Your Website is worth reading before you write that one page.