Your second cohort: find one repeatable acquisition channel (UK SaaS, Claude Code 2026)

You raised your price, you saved a few of your first churners, and your retention has finally stopped looking like a slide. The product holds water. That changes which problem deserves your attention this week, because the constraint on the business has quietly moved.

It is no longer "is this good enough to keep". It is "where does the next customer come from, again and again".

Most solo founders get their first ten or twenty customers from a flurry of one-off effort: a launch post, a few DMs, a friendly community, the founder's own audience. That works once. Then it dries up, and you are staring at a flat signups graph wondering why growth stopped. The honest answer is that you never had growth. You had a launch. This is the runbook for the next stage as a UK founder building with Claude Code: turning sporadic, lucky acquisition into one channel you can run on repeat.

Why your first cohort is not a channel

Your first customers came from a burst. You posted somewhere, told some people, rode a wave of goodwill, and a handful signed up. It felt like momentum. It was a one-time withdrawal from a small, finite account: your existing network, a single launch audience, a moment of novelty.

The reason it does not continue is mechanical. A launch reaches a fixed pool of people once. When the pool is spent, the signups stop, because nothing about that effort renews itself. You cannot post the same launch to the same subreddit next Tuesday and expect the same result. The audience has already seen it, and the algorithm has moved on.

A channel is different. A channel is a loop: a repeatable set of actions that reaches new people for roughly the same effort each time, with roughly the same payoff. The whole game at this stage is converting your one-off launch instinct into one loop you can pull every single week. Not three customers this month from luck, but two or three customers a week from a process you control.

One channel, not six

Here is the mistake that eats months. The founder, sensing that acquisition is now the job, decides to "do marketing" and spreads across everything at once: posting on X, answering on Reddit, starting a newsletter, trying a bit of SEO, sending some cold emails, joining three communities. Each gets a fraction of attention. None gets enough to work. Twelve weeks later there are six half-built channels and still no reliable source of customers.

Do the opposite. Pick one channel and commit to it for four to six weeks before you even consider a second. This feels slow and it is the fastest path, because channels have a threshold. Below a certain amount of consistent effort, almost any channel produces nothing, and you conclude it "does not work". Push past the threshold on one channel and it starts to compound: you learn the audience, you find the angle that lands, your name starts to recur, and the loop begins to turn on its own.

One channel that reliably brings five customers a week beats six channels that each bring zero. You can always add the second once the first is genuinely repeatable. You almost never recover the months lost to doing all of them badly.

Pick the channel where your buyers already are

The right channel is not the easiest one or the one a generic marketing list ranks highest. It is the one where your specific buyers already gather, doing the thing that makes them want your product. Your job is to be present there, usefully, on repeat.

You already have the data to find it, sitting in your first cohort. Ask your first ten customers two questions: where were you when you found us, and what were you trying to do at the time. The answers cluster. Maybe they were searching a very specific problem on Google. Maybe they were lurking in one niche community venting about exactly the pain you solve. Maybe a single person kept recommending you. That cluster is your channel, evidenced by people who actually paid, not guessed from a blog post about "the top ten growth tactics".

The shape changes with who you sell to. If your buyers are UK developers, they live in technical communities, on GitHub, and in search results for specific errors and tools. If they are small UK trades, freelancers, or local service businesses, they are in niche Facebook groups, trade forums, and word-of-mouth networks, not on X. If they are hobbyists, the channel is wherever that hobby congregates online. Match the channel to where the buyers breathe, and acquisition stops feeling like shouting into a void.

Instrument the path before you scale it

You cannot improve a channel you cannot see. Before you pour weeks into one, wire up the ability to tell, customer by customer, where they came from and how far they got. Without this you are flying blind, and you will scale the wrong thing or kill the right one.

You need three numbers per channel, in order:

• Signups — how many people created an account from this source.
• Activated users — how many hit the value moment you already identified during your retention work.
• Paying customers — how many converted to a subscription.

The last number is the only one that counts as success. A channel that drives a hundred signups and zero paying customers is worse than one that drives ten signups and three customers, because the second one is a business and the first is a vanity spike. Traffic is not the goal. Paying, sticking customers are.

The plumbing is small. Tag every inbound link with its source, capture that tag on signup, and store it against the user. Then you can group your Supabase signups by source and read the three numbers off directly. This is exactly the kind of narrow, well-specified job to hand to Claude Code: "Add UTM source capture to our signup flow. When a new user signs up, read the source from the URL or a cookie set on landing, and store it in a source column on the users row in Supabase. Then write me a query that, grouped by source, shows total signups, activated users, and paying customers." One focused session and you can finally see which channel is earning its keep, instead of arguing with a feeling.

Run the loop, judge it on customers

With one channel chosen and the path instrumented, the work becomes weekly and boring in the best way. Define the loop concretely enough that you could hand the steps to yourself next Monday and expect the same result. That repeatability test is the whole point. If the steps only worked because of a one-time audience or lucky timing, it was a launch wearing a channel costume.

Run the loop for four to six weeks without flinching at week-one silence, because the threshold is real and early numbers always look discouraging. Watch the three metrics accumulate, and steer on the last one. If a channel produces signups but no paying customers, the problem is usually a mismatch between who the channel attracts and who the product is for, or an activation leak you can route back to product work. If it produces paying customers for an effort you can sustain every week, you have found it. That is the asset you were missing: not more customers, but a known, repeatable way to get them.

When not to chase a channel, and when to stop tuning

A few honest cautions, because the order matters more than the effort.

Do not work on acquisition if retention is still broken. A channel fills the bucket, and a leaky bucket just empties faster with more water in it. If customers churn within weeks, your highest-leverage work is still the product, not the funnel. Pouring acquisition on top of poor retention burns money and goodwill and teaches you nothing, because the customers leave before the channel can prove anything.

And once a channel is working, resist the urge to optimise it to death or to immediately bolt on five more. You do not need a full marketing stack, an attribution dashboard, or a content calendar across six platforms at fifteen customers. You need one loop that turns, instrumented well enough to trust, run consistently enough to compound. Get that, keep it turning, and only then earn the right to add the second channel. Repeatable beats clever, and one working loop beats a wall of dashboards every time.

You shipped, you launched, you kept and saved a cohort. Now give the business the one thing it has been missing: a reliable way to find the next customer. Pick the channel, wire up the path, and run the loop until it runs itself.

---

Want a validated idea to build next? Every week IdeaStack publishes one deeply researched UK business opportunity, with real keyword volumes, competitor pricing, SERP analysis, and copy-paste Claude Code builder prompts. Read this week's free report and start building.