In short6 min read
- The founder reported $469 in about six weeks on a €50 ad budget.
- Two of the first 14 trials bought without ever running a live game.
- Corporate email scanners can burn one-time sign-in links before a human clicks.
- Typed answers, rounds and Excel import all came from host requests, not the roadmap.
Plenty of people now build small tools with AI coding assistants. Far fewer turn one into something strangers pay for. This case is worth a close look because the founder published the boring, useful part: the numbers from the first weeks and the three assumptions that turned out wrong. By the end you will know which of those mistakes you can test for before your own launch.
The founder runs live quizzes at corporate events and trainings and paid for Crowdpurr and Kahoot for years. In February they built their own quiz tool with Codex and Claude, only for their own sessions. This autumn they opened it to other hosts under the name EventQuiz.
In this article
- Why an internal tool makes a strong first product
- The first six weeks in numbers
- Mistake 1: Tracking the wrong activation moment
- Mistake 2: Sign-in links that expired before anyone clicked
- Mistake 3: Assuming everyone answers every question
- Let real events write the roadmap
- Pre-launch risk check for AI-built micro SaaS
- Finding the next customers without bigger ad budgets
- Steps to turn your tool into a product
Why an internal tool makes a strong first product
The tool started as a fix for the founder’s own work, not as a startup idea. That matters. By the time it went public it had already been used in real sessions, and the founder knew the job it had to do because they did that job every week.
It is also a clean example of the AI-assisted build pattern this site covers in its AI agents for solo businesses hub. The coding assistants lowered the cost of building. They did not remove the work of finding buyers, which is where most of the founder’s lessons sit.

The first six weeks in numbers
Here is everything the founder shared about early traction.
| Metric | What the founder reported |
|---|---|
| First sale | September 17, a $39 event pass |
| Revenue, first ~6 weeks | $469 |
| Ad spend | €50 |
| Trial-to-paid | About 25% |
| Early trial sample | 14 trials, 2 purchases |
That is a modest start, and the founder presents it that way. The useful part is that the product earned money from people outside their own network while still very young.
Mistake 1: Tracking the wrong activation moment
Standard SaaS advice says a trial user must reach an aha moment before paying. For a quiz tool, the founder assumed that moment was running a game with real people.
Then they looked at the first 14 trials. Two bought, and neither had played a single live game first. The buyers paid before the event, which makes sense for a host who needs the tool ready on a date.
The lesson is not that activation metrics are useless. It is that your activation event should come from watching what buyers actually did, not from a template. For event products in particular, the purchase may happen during planning, so onboarding should help someone prepare an event, not only run one.
Mistake 2: Sign-in links that expired before anyone clicked
A customer at a large company reported that the sign-in link dropped them on the homepage. The database showed they had requested two links that morning and both were already used. Their corporate email scanner had opened the links before the person did.
If your buyers are corporate trainers or HR teams, this is not an edge case. It is your core segment.
The fix the founder shipped
In the thread, the founder explained the change: the link still exists, but opening it now only shows a Sign in button. Nothing is used up until a person clicks, because scanners open pages but do not press buttons.
One commenter described the same pattern in more technical terms. The page load should be harmless, and only the button’s form submission should consume the token. They also suggested putting a 6-digit code in the same email as a fallback and logging which email domains end up using it, so you learn which companies run aggressive filters.
Mistake 3: Assuming everyone answers every question
At a 32-player event with 20 questions, only 42% of the possible answers came in. The founder admitted they had expected that having the question on each player’s own phone would pull more people in.
A commenter replied that this sounded normal, since part of any room just watches. Either way, the design takeaway holds: scoring, leaderboards and host screens should look good when many players skip questions, not only when everyone plays.
Let real events write the roadmap
Three features shipped because hosts asked for them during real events: typed answers, rounds, and importing questions from Excel. None came from the founder’s own roadmap.
That is the quiet advantage of selling early. Each paid event produced a specific request from someone who had already paid, which is a far stronger signal than a feature idea from a survey.
Pre-launch risk check for AI-built micro SaaS
The three mistakes share a pattern: each was a hidden assumption about how buyers behave. Before you open your own internal tool to the public, run through this check.
| Hidden assumption | How it fails | Cheap way to test it |
|---|---|---|
| Users must try the core feature before paying | You optimize onboarding for the wrong step | Look at what your first paying users did before checkout |
| Login emails reach a human first | Corporate buyers cannot sign in | Make link pages require a click, add a code fallback |
| Users engage the way you do | Reports and screens look broken in real use | Review results from your first real sessions, not demos |
| Your roadmap reflects demand | You build features nobody pays for | Ship what paying customers ask for during real use |
| A small ad test proves a channel | Spend grows faster than revenue | Test one direct channel alongside ads |
The last row is the open question the founder raised. They said plainly that €50 in ads will not scale by itself.
Finding the next customers without bigger ad budgets
The founder asked other sellers which channels worked for reaching event organizers and trainers. The replies gave practical options.
- Direct outreach. One commenter noted that corporate trainers and event hosts are active on LinkedIn and suggested messaging 20 of them personally before spending the next €50 on ads.
- Ask existing buyers. Another suggested asking hosts who already paid what triggered their search and where they found the tool, then testing one channel aimed at that exact organizer type.
- Keep the offer fixed while testing. The same commenter warned that changing the demo and offer during a channel test makes it impossible to tell whether the channel or the message did the work.
If you plan a direct outreach push, the setup lessons in 300 cold emails, 12 replies and a broken deal are a useful companion. For a wider look at how another solo founder judged each channel, see this solo SaaS channel audit.
Keep expectations grounded. This founder had years of experience as a host and a real problem to solve. Your results will depend on how well you know your buyers and how consistently you reach them.
Steps to turn your tool into a product
- Pick a tool you already use for paid work. Build it for your own sessions first with an AI coding assistant, and use it on real jobs before selling it.
- Price around a real unit of value. The founder sold a $39 event pass; choose a unit that matches how your buyers budget.
- Harden sign-in for corporate inboxes. Make email links open a page with a button, and add a short code fallback.
- Study your first paying users. Record what they did before buying and rebuild onboarding around that step.
- Design for partial engagement. Check that reports and screens still make sense when many users skip steps.
- Log every feature request from paying customers. Ship the ones that come up during real use first.
- Run one direct channel test. Message 20 likely buyers personally, keep your offer unchanged, and compare the result with your ad spend.
List the tools you patched together for clients
Today, list the tools you built or patched together for your own client work. Pick the one you use most, and write down who else does the same job and how they currently pay for a solution.
Get new AI automation playbooks by email
One email when new guides and case studies go up. No spam, unsubscribe anytime.

Leave a Reply