Marketing for founders

Every founder marketing guide opens the same way: pick your channels, build a content calendar, define your ICP, set up analytics. Then you have a plan, and about four days later you have a plan you aren’t following.

The problem isn’t discipline. It’s that a plan assumes you already know what will work. You don’t, and neither does anybody writing those guides, because the right answer depends entirely on which part of your business is currently broken.

Marketing is a multiplier, not an engine

Whatever your product and page already do, marketing does more of it. That’s the whole mechanism. If your homepage converts one visitor in two hundred, doubling traffic gets you one more customer and twice the confusion about why it isn’t working.

Multiplying something broken gets you more broken, faster.

So before choosing a channel, work out what you’re multiplying. There are five candidates and they have a strict order.

The five bottlenecks, in order

1. Nobody can tell what it is

Your homepage names a category, “the platform for modern teams”, rather than a problem. Test it: show the page to someone for five seconds, hide it, ask what it does and who it’s for. If they can’t say, nothing downstream matters.

This is the most common bottleneck and the cheapest to fix, and almost nobody starts here because rewriting a headline doesn’t feel like marketing.

2. There’s no way to become a customer

No visible call to action, a signup that breaks on mobile, pricing hidden behind a demo request, four plans to choose between before anyone understands the product. Sign up on your own phone, on cellular, in a private window. Time it. Most founders find something broken inside ninety seconds and most have never tried.

3. Genuinely nobody is arriving

Clear positioning, working signup, under about a hundred visits a week. This is the only case where “I have a marketing problem” is the right diagnosis, and even here the answer is rarely a channel strategy. It’s finding the specific places your problem gets discussed and turning up in them.

4. People arrive and leave

Traffic is fine, signups are near zero. Under about 1% of visitors doing anything is a conversion problem. More traffic here is the most expensive mistake available to you: every additional visitor is another data point confirming what you already know.

5. People sign up and disappear

This is the one that looks most like a marketing problem and least is. Acquiring customers who vanish on day two is paying money to learn nothing. The fix is five conversations, not a channel.

Why that order, and not another one

The order is not a preference, it is a dependency chain. Each stage gates the one below it, which means work spent below your actual bottleneck cannot show a result even when the work is good.

Traffic to a page nobody understands produces nothing. A clear page with no signup path produces nothing. A working funnel with no visitors produces nothing. Retained customers cannot exist before any of it. Every stage multiplies the one above, so the first zero in the chain sets the product of the whole thing to zero.

Find the first zero. Everything else is arithmetic on top of it.

Running the diagnosis in thirty minutes

You need four numbers and one stranger. None of it requires a tool you do not already have.

• Visits last seven days. Any analytics. If you have none installed, that is your first job, and it also means every decision you have made so far was made blind.
• Signups last seven days. A database count is fine.
• People who signed up four weeks ago and used it in the last seven days. This is the retention number and it is the one nobody has.
• Time from landing page to a working first action, measured by you, on your phone, on cellular, in a private window, with a stopwatch.
• One stranger’s answer to “what does this do and who is it for”, after five seconds on your homepage, with no explanation from you first.

Those five readings resolve the list above almost every time. The stranger’s answer settles bottleneck one, the stopwatch settles two, visits settle three, the ratio of signups to visits settles four, and the retention number settles five.

The one move for each bottleneck

A diagnosis is only useful if it terminates in a single action with a way to tell whether it worked.

• Cannot tell what it is. Rewrite your first two lines using the words your stranger used when they got it wrong. Success: three of the next three strangers describe it correctly.
• No way to become a customer. Fix the single worst step in your own stopwatch run. Success: the run completes under two minutes on a phone.
• Nobody arriving. Twenty individual messages to people who have publicly described the problem in the last month. No link in the first message. Success: five replies.
• Arrive and leave. Watch three people use it without helping them, then remove whatever made all three hesitate. Success: the signup rate moves at all.
• Sign up and disappear. Five conversations with people who left, asking only what they hoped it would do. Success: you can state the reason in one sentence that is not a guess.

Notice that none of those five is a channel, and that all of them fit in a week.

Two failure modes that survive the diagnosis

Doing all five at once. A diagnosis that produces five workstreams is a plan again, and you will be off it by Thursday. The output has to be one thing.

Re-diagnosing instead of acting. The list is comfortable because analysis feels like progress and carries no risk of being wrong in public. Run it, act, and do not run it again until the action has a result.

What this rules out, this month

• SEO. Six-month payback on a this-month problem. It’s a good investment once you know what converts. You don’t yet.
• Paid ads. You’d be paying to test a page free traffic hasn’t validated.
• A launch. One non-repeatable traffic spike, aimed at a page with nothing on it suggesting another human has used this.
• Another feature. Building is procrastination with a Git commit attached.

What to do instead

Pick the first failing item. Fix that one thing. Then check again, because fixing it usually moves the bottleneck somewhere else, and the next right action changes with it.

That’s the actual loop, and it’s why plans don’t survive: a plan written in week one assumes a bottleneck that has already moved by week three.

Two hours a week is a different problem

A founder with two hours a week and a founder working full-time on distribution should not receive the same advice, and almost all marketing advice is written as though everyone is the second one.

With two hours, sequencing is everything. One well-chosen action completed beats four started. Which is, again, an argument for diagnosing rather than planning: a plan gives you forty things and no way to rank them.

Where to go from here

Pick the page that matches the reading you just took, not the one that sounds most interesting.

• Launched and the line is flat: launched, and nobody came splits the two cases that look identical from your chair.
• Some users, no movement: how to read a flat line starts with retention rather than acquisition.
• Building SaaS specifically: marketing for SaaS founders has the pre-fit sequence.
• Building for developers: marketing a developer tool inverts most of the advice above, for good reasons.
• Just shipped, nothing happening: how to market an app starts with the page rather than the channel.
• Wondering why this is the hard part now: startup distribution is the long version.

And if you would rather have the diagnosis run against your actual site than run it yourself, the Product Check is free and needs no account.