Marketing for SaaS founders

SaaS marketing advice is overwhelmingly written for companies that already work. That have a funnel to optimise, a cohort to retain, a CAC to bring down. If you have forty users and no paying customers, none of it applies, and following it is how founders lose their first six months.

Early-stage SaaS has its own sequence. It looks like this.

Stage one: get ten customers by hand

Not through a channel. By hand. Finding individual people with the problem and talking to them, in threads they started, in replies to their complaints, in emails that aren’t templated.

This feels wrong to engineers because it doesn’t scale, and we’re trained to distrust things that don’t scale. But you don’t need scale yet. You need ten.

If you can’t get ten customers by hand, a channel won’t get you a hundred. Channels multiply what already works.

There’s a full version of this in how to get your first SaaS customers.

Stage two: steal your positioning from those ten

The single highest-leverage marketing asset you will ever have is a sentence your customer said out loud, used verbatim on your homepage.

Ask each of the ten one question: “What were you doing about this before?” Their answer tells you what you actually replaced. For most early SaaS it isn’t a competitor. It’s a spreadsheet, a group chat, a person doing it manually on a Sunday night, or nobody bothering.

That answer is worth more than a competitive analysis, because it tells you what the customer was willing to change a habit for.

The trigger matters more than the persona

“Backend engineers at Series A companies” tells you nothing about when to show up. A trigger does:

• “We broke prod with another migration.”
• “The person who used to do this has left.”
• “Our spreadsheet stopped coping at eighty rows.”
• “I’ve just taken over organising this.”

Each of those is a moment when somebody goes looking. Write down three for your product. If you can’t, that’s your finding for the week, and it means more conversations rather than more promotion.

Stage three: one channel, understood properly

The instinct at this point is to test five channels for a month each. Don’t. Ten channels tested badly are worse than one channel understood properly. You end up with five inconclusive results and no idea which failure was the channel and which was the execution.

Pick the one where your trigger is most visible. Do it fifteen times. Then look at what it produced. That’s a real result; a month of “trying LinkedIn” isn’t.

How to choose it, without a framework

Three questions, in this order. The first channel that survives all three is your channel.

• Can you see the trigger happening there? Not your audience, the moment. Somewhere people post the thing that means they have just started looking.
• Can you do it fifteen times in a fortnight, around the rest of your job, without buying anything?
• Will you be able to tell whether it worked? If the answer arrives in nine months, it is a real channel and it is not this quarter’s channel.

For most early SaaS this eliminates paid, search and launches immediately, and leaves being useful in the specific rooms where the problem gets discussed. That is not a fashionable answer, which is part of why it is still available.

Fifteen attempts, then read the result

One attempt tells you nothing, and everybody quits somewhere around three. Fifteen is roughly the point at which the difference between “this channel does not work” and “I was bad at this channel” becomes visible.

Read it against one number decided in advance. Not impressions, not upvotes: replies from people with the problem, or signups you can trace to the attempt. If fifteen good attempts produce none of either, the channel is answered and you have earned the right to try the next one.

Pricing is marketing, and most early SaaS gets it backwards

Hidden pricing costs you more early-stage customers than a high price does. A founder evaluating your product at 11pm will not book a call to find out what it costs; they’ll close the tab.

The same goes for plan count. Four tiers before anyone understands the product is a decision you’ve outsourced to someone with no basis for making it.

Which numbers mean anything yet

Early SaaS gets handed the metrics of a company with a sales team, and most of them are noise at your size.

• Worth watching. Paying customers. Whether the people who signed up four weeks ago are still using it. How many conversations you had with somebody who has the problem. Time from landing page to a working account.
• Not yet. CAC, computed from a spend of zero. LTV, computed from a churn rate you cannot observe with eleven customers. MRR growth rate, which at these numbers turns one customer into a triple-digit percentage and tells you nothing.
• Actively misleading. Signups, visits and anything with the word engagement in it. All three go up when you post somewhere busy and none of them predict revenue at this stage.

What to ignore until you have paying customers

• Content calendars and programmatic SEO
• A launch on any platform
• Paid acquisition
• Referral programmes
• Anything described as a growth loop
• Rebuilding the landing page for the fourth time

All of those multiply something. Get the something first.

Then measure, or none of this compounds

The reason most founders never get good at marketing isn’t effort. It’s that nothing tells them whether last week worked, so they never develop instincts. Write down what you did and what happened, even when what happened is nothing. Especially then.

Six entries in, you have something no marketing article can give you: evidence about your own business rather than advice about businesses in general.

What a week actually looks like

Concretely, at stage one, with a job and a product to maintain:

• Monday, twenty minutes. Find five people who described the problem in public in the last month. Write the five names down.
• Tuesday to Thursday, twenty minutes a day. Be useful to two of them. No link in the first message.
• Friday, ten minutes. Write down what you did and what happened, including nothing.

That is under two hours, it is repeatable, and thirty of those weeks is a business. Compare it against what a marketing plan asks of you in week one.

Where to go from here

The stage-one mechanics are in how to get your first SaaS customers. If the line is flat with users already on it, reading a flat line starts with retention instead. If you are selling to developers, the developer tool version inverts several of the rules above. And if the temptation this week is to buy an agent platform to do it all for you, that category is worth understanding first.