Launched on Product Hunt today (solo, no network) — what actually moved your first 100 visitors?
Kotoba Dojo is a game: describe a target image in your own words, AI redraws it from that text only, you get scored on the gap.
Solo in Tokyo. No hunter, no email list, no audience. Today I seeded replies in "what are you building" threads on X — the first organic PH comment came from that.
For people who launched with no audience: where did the first 100 real visitors actually come from? X replies, Reddit, directories, something else?
I'll post day-1 numbers in the thread as they come in.
i’d probably care more about where the first 20-30 relevant people come from than getting 100 random visitors
if 10 come from reddit, 10 from x replies and 10 from direct outreach you can actually compare which group sticks around and uses the product
that seems more useful early on than optimizing for a bigger traffic number
For a game like this, I would not optimize for 100 random visitors yet. I would aim for 30 people from three distinct sources and compare behavior.
Try 10 from language-learning communities, 10 from AI art communities, and 10 from direct outreach to people who already share word or image games. Give each group a separate link or code.
The useful signal is not just clicks. Track who starts a round, finishes it, tries a second image without being prompted, and shares a result. If one source produces voluntary second rounds, that is probably where the next 70 visitors should come from.
Your X reply already gave you one small piece of evidence. I would repeat that format with a specific challenge image rather than a general launch link. The game itself is the content.
This is the most actionable reply I've gotten anywhere today — thank you.
Two things I'm taking from it immediately: separate tracked links per source (I can see rounds started/finished/retried server-side, so measuring voluntary second rounds is very doable), and the challenge-image format on X — you're right that the game itself is the content, and I've been leading with the launch link instead.
I'll run the three-source test this week and report what the second-round data says. Genuinely appreciate this.
Really glad it was useful. You already have the most important part in place: the ability to measure behavior after the click, not just traffic.
One thing I would define before starting is the threshold that makes one source worth pursuing. For example, not only which source produces more second rounds, but how large the difference needs to be before you concentrate the next batch there.
Looking forward to seeing the results. The comparison between sources should be much more useful than the total visitor count.
Good push — defining it now, before the data can tempt me to move the goalposts:
Each source gets a minimum of 10 visitors before I judge it. A source "wins" the next batch if its voluntary second-round rate is at least 2x the runner-up, with an absolute floor of 3 unprompted second rounds. If no source clears that bar, the test was inconclusive and I change the offer, not the channel.
Writing this in the spec file today. Results in this thread either way.
Locking this in before any numbers come in — thank you for forcing the pre-commitment.
Criteria, written into my spec file as of today:
One practical note on my setup: signups per source are recorded in my database, and spontaneous return visits will come from analytics. If return-visit data is too thin at this traffic level to be meaningful, I'll note that in the report rather than quietly swapping in a friendlier metric.
I'll post the numbers in this thread either way — including if the result is "inconclusive."
One distinction worth preserving before you start: voluntary second rounds and spontaneous return visits test different things. A second round in the same session measures immediate pull; a return visit measures recall and habit, and will be much noisier with 10 visitors per source.
Since your server already records rounds, I would keep the pre-registered primary metric as unprompted second rounds per source. Treat spontaneous return visits as a secondary signal. That way a thin analytics sample cannot force you to swap the success criterion after the test begins.
The rest of the threshold is clear. Once each source reaches 10 visitors, post starts, completions, second rounds, and returns separately. Then the decision will be auditable even if it is inconclusive.
Agreed, and thank you for catching that — I was conflating two different things.
Locking the primary metric as unprompted second rounds per source (server-recorded). Spontaneous return visits stay as a secondary signal only; they will not change the pass/fail call.
Honest status: the launch-day traffic did not convert into any tracked signups, so the three-source test hasn't produced usable data yet. Source tracking is now in the database, and I'm starting the outreach for the three groups this week. I'll post starts / completions / second rounds / returns per source here once each group reaches 10 visitors — inconclusive or not.