How to get your first 25 customers for a vibe-coded app
No audience, no ads budget, no sales experience — the manual loop that actually lands your first paying customers, and the honest metric to count while you do it.
7 min read ·
You shipped. That was supposed to be the hard part, and nobody warned you it was the easy half. The app works, the landing page is up, and the internet is perfectly, completely silent. This guide is the loop that breaks that silence — by hand, this week, without an audience or an ads budget.
Define who pays, from the product you already built
Not a persona deck — one sentence. Who has the problem your app solves badly enough to pay a stranger for it? Look at what you actually built: every feature is a bet on someone's pain. Write the sentence down. If you can't, your outreach will read like it — vague notes get deleted on the first line.
Find them on the live web, not in a lead database
At 25 customers, lead databases are the wrong tool — they're built for teams buying ten thousand rows, and your buyers are too niche to be a column in one. Search the way a customer would: the directories they list themselves in, the communities they ask questions in, the "best X in Y" pages they fight to appear on. You're hunting a few dozen real companies, not a spreadsheet.
Sort likely buyers from everything else
Half of what you find won't be buyers — it'll be competitors, press, communities, and noise wearing buyer costumes. Be ruthless: a likely buyer is a company that fits your sentence from step one and shows a live sign of the problem. Twenty of those beat two hundred maybes, because every step after this one costs you time per company.
Write like one founder to another
One email per company, and the first line must prove you actually looked: name the thing on their site a template couldn't know. Two or three sentences total — what you noticed, what you built, one soft question. No "quick call?", no calendar link, no feature list. You're a founder who made something useful, not a rep with a quota.
Send small, from your own inbox
Your regular Gmail is the right tool at this scale — replies land where you already live. Keep it to five to ten a day, verify each address before you hit send, and put a real unsubscribe line and postal address in the footer. That's not bureaucracy; it's what keeps your domain trusted enough for email #200 to still land.
The deliverability rules have their own guide: cold email from your own Gmail without burning it.
Follow up — the replies live there
Most founders stop after one send. Most replies come from touch two and three. A follow-up is one line — new detail, same soft question — three or four days later. Two follow-ups, then stop; past that you're not persistent, you're spam.
Count replies, not clicks
The honest metric at this stage is the reply. Opens are noise, clicks are curiosity — a reply is a human deciding you're worth a sentence. Ten percent reply rate on well-chosen companies is normal; every reply is a conversation, and conversations are where the first 25 customers actually come from. The close happens on your product, not in the email.
Run the loop until the number, then automate it
Do a week of this by hand — it teaches you your buyer faster than any analytics dashboard. Then be honest about whether you'll keep doing it, because the loop only works run daily, and you have a product to build. This exact loop — model the buyer, find likely ones, verify, write personal, send small, follow up, count replies — is what ProfitLayer runs autonomously. Point it at your site; your first 15 finds every month are free.