Shipped my first paid product 36 hours ago. 0 sales, a 4-hour patch, and the most useful feedback from someone who didn't buy.
I shipped a $99-349 productized kit on Gumroad on Monday. It's for AI agencies building dental voice agents. Solo founder in Estonia.
36 hours later, here's what actually happened.
The launch itself was unremarkable.
I posted on r/SideProject. Got 131 views in the first 90 minutes. 4 substantive comments. No sales.
Statistically expected for a niche kit with no audience. Day 1 isn't about sales. Day 1 is about finding the people who'll tell you what's wrong before you build the next thing.
That's what happened next.
The Reddit comment that forced a 4-hour patch.
In a Reddit reply, I made claims about three features: barge-in handling, NATO phonetic spell-back, and KB fallback.
The KB fallback was real. The other two were not. I'd written them into marketing copy but never built them into the actual prompts.
Caught it at 10 PM. Patched all three language files, tested, re-uploaded to Gumroad by 2 AM.
Lesson: audit your marketing claims against your actual product before launch. If you wouldn't bet $1000 the feature works, don't say it does.
The expert feedback I didn't see coming.
I cold-DM'd a few builders on LinkedIn the day before launch. Not to sell. Just to introduce the kit and ask for honest feedback.
One of them told me the landing page is missing a demo. People want to hear the agent before they trust the kit. He's right. Adding it this week.
The most valuable feedback came from a builder I cold-DM'd, not a buyer.
Three things I'd tell myself 36 hours ago:
- Audit your claims against your code before launch, not after.
- Cold-DM builders before launch, not just to find buyers, but to find people who'll tell you the truth.
- 0 sales on Day 1 isn't a problem. The problem is launching without knowing what you'd change after Day 1. I now know.
The kit is at dental-voice-bot-kit.vercel.app if you're curious. But the post isn't to sell. It's the boring real version of Day 1, not the highlight reel.
What's the most useful feedback you've gotten from someone who didn't buy?
The useful signal here is not “0 sales.”
It’s that the first real objection was trust, not price.
Nobody said the offer was too expensive.
They said “I need to hear it first.”
That usually means the product is not being judged like a template.
It is being judged like infrastructure.
And infrastructure without proof does not convert.
That is the real lesson in the first 36 hours:
the missing asset was not another feature.
It was proof.
The second that becomes true, the current name starts capping trust too.
“dental-voice-bot-kit” explains what it is, but it also traps it at template-tier.
Useful for discovery, weak for trust.
The moment this becomes something buyers evaluate like operational infrastructure instead of a downloadable kit, the product will want a name that can hold more weight.
Exirra.com would carry that much better once he moves beyond the “kit” framing.
The trust-vs-price read is sharp. The demo coming this week is exactly the proof asset you're describing. Appreciate the diagnosis.
Exactly.
The demo solves the first trust gap:
“does this actually work?”
But after that, the second trust gap becomes:
“is this a serious operational system or just a downloadable kit?”
That’s where the name starts mattering more.
“dental-voice-bot-kit” is fine while you’re validating demand.
But once buyers are watching a demo and imagining this inside their front desk workflow, the kit framing starts working against the infrastructure story.
The demo proves capability.
The name has to carry seriousness after that.