I built a SaaS. 38 engineers signed up before I could charge them.

I launched on February 1st with a half-broken product.

No payment system. No fancy landing page. Just a text box that said “paste your logs” and a button that said “generate report.”

By February 14th, I had 38 signups. All organic. Zero dollars spent on ads.

And I still couldn’t take their money.

The problem I didn’t know I was solving

Here’s what I thought I was building: an AI tool that writes incident reports.

Here’s what I actually built: a way for engineers to stop spending 2 hours at 3 AM doing something a computer should do in 30 seconds.

The difference matters.

I found out the hard way when my Redis cluster failed at midnight. Twelve minutes to fix it. Two hours to write the report. My manager wanted it by 9 AM, and I had four different Slack threads, a CloudWatch dump, and some Kubernetes events I’d screenshotted but couldn’t find anymore.

I remember thinking: “I just fixed a production incident. Why am I now fighting with Google Docs?”

The launch that wasn’t supposed to work

I didn’t have a launch strategy. I had a URL and a problem I knew 10,000 other engineers had because I’d lived it.

So I shipped it incomplete. No Stripe integration. No email automation. Just the core thing: paste chaotic logs, get a structured report.

Free trial. Three reports. That’s it.

First signup came 6 hours after launch. A Tech Lead at a 200-person company.

By day 7: 18 people.

By day 14: 38 people.

Most of them asking the same question: “When can I pay you?”

The math that made me panic

Let me show you the numbers that kept me up at night:

Who signed up:

13 decision makers (Founders, CTOs, Directors, Tech Leads)
11 people at companies with 200+ employees
9 people who explicitly said “yes, company budget”

What they wanted:

83% said they’d pay (either personally or with company money)
15 people requested a demo
One Director called his engineers “cowboys” and said my tool finally helped them understand structured troubleshooting

What I didn’t have:

A way to charge them

Stripe integration was sitting in my backlog. “I’ll build payments after I validate the idea,” I’d told myself.

Turns out, 38 people asking to give you money IS validation.

The one message that changed everything

Day 9. An email from someone with a 200+ person team:

“Your demo is ready — 200+ team solution”

Wait. I didn’t write that subject line. That was MY email to HIM.

He replied:

“Perfect timing. We handle weekly incidents and our reports are a mess. When can we start?”

That’s when I realized: I wasn’t selling a tool. I was selling two hours of their life back.

Another engineer wrote about his biggest struggle: “Finding proper error messages, coordinating team, writing clear reports, tracking action items.”

All four problems. One incident. Every week.

What “product-market fit” actually feels like

Here’s what they don’t tell you about PMF:

It doesn’t feel like success. It feels like panic.

Because you suddenly have people who want to pay you, and you’re not ready. Your product is 85% built. Stripe is “coming soon.” You’re manually responding to demo requests.

But here’s what PMF actually looks like in the data:

2.4 signups per day (organic, no ads, no launch on Product Hunt)
35% decision makers (not tire-kickers, not students, actual budget holders)
“When can I pay?” instead of “Can I get a discount?”
People signing up, then emailing to ask for a call
Companies with 200+ people saying “this solves a real problem”

I didn’t manufacture this. I just put a solution in front of people who were actively suffering.

The mistake I won’t make again

By day 15, I had Stripe live.

The first payment came in at $9.50/month. Early bird pricing. 50% off forever.

That customer had been waiting 6 days to pay me.

Six. Days.

If you’re building something and people are asking “when can I pay?” — stop everything and build payments. Not next week. Today.

I lost revenue because I thought “validation” came before “monetization.”

Wrong.

Monetization IS validation.

What I learned about engineers

The data told me something surprising.

I surveyed everyone: “What’s your biggest struggle during incidents?”

I expected: “Writing reports.”

I got: “Finding proper error messages.”

8 people said that. Only 5 said “writing reports.”

Here’s what that means: Engineers don’t hate documentation. They hate the 90 minutes of archaeological dig through Slack, CloudWatch, and Grafana trying to remember what the hell happened.

The actual writing? That’s 20 minutes.

The finding, copying, organizing, verifying? That’s the nightmare.

So that’s what ProdRescue actually solves. You paste the chaos — Slack threads, JSON dumps, whatever you have — and it finds the panic, the timeout, the decision, the person who made it.

Then it writes the report with evidence links so your manager can verify every claim.

The writing is just the last mile.

The numbers that matter

Here’s what 38 organic signups in two weeks actually means:

If this continues:

72 signups by end of month 1
20% convert to paid = 14 customers
Average $15/month = $210 MRR month 1

Conservative year 1:

$48K-72K ARR
Break-even: month 3
Profitable: month 4

Optimistic year 1:

15% MoM growth (realistic for product-market fit)
$120K ARR
First enterprise customer (200+ team): month 6

Not unicorn numbers. But profitable SaaS numbers.

And it started with 38 people I couldn’t charge.

What I’d do differently

Ship with payments on day 1. Even if it’s just a Gumroad link.

I thought launching incomplete meant “no payments yet.” Wrong. Incomplete means “missing features.” Payments aren’t a feature. They’re proof people value what you built.

Everything else? I’d do the same.

Launch before you’re ready. Talk to every single person who signs up. Ask what they struggle with. Then fix that, not what you think they need.

Where it’s going

Right now, ProdRescue does one thing: turns messy incident logs into executive-ready reports in 30 seconds.

Next month: Slack integration, so you don’t even need to copy-paste.

Next quarter: Team workspaces, so your whole SRE team can search past incidents.

Next year: The tool I wish I’d had at 3 AM when Redis failed and I knew I’d spend the next two hours fighting with Google Docs instead of sleeping.

The real lesson

You don’t need 10,000 users to validate a SaaS.

You need 10 people who would pay tomorrow if you let them.

I have 38.

And now, finally, I can let them.

Devrim builds tools for engineers who are tired of 3 AM documentation. Try ProdRescue free (3 reports, no credit card). Or don’t. But if you’ve ever spent two hours writing an incident report, you know why it exists.

Related: If you’re curious what a 30-second AI-generated incident report actually looks like (with evidence mapping so you can verify every claim), try pasting your worst incident log. See if it beats your 2-hour manual version. It probably will.