Blog · By  · · 7 min read

My launch flopped. What do I post now? A 7-day plan

Your launch got almost nothing. Here's what to post in the week after, day by day, with fill-in templates and numbers from 2,000 top build-in-public posts.

Short answer

Post the flop itself, with your real numbers, then spend the week showing the work instead of announcing it again. In 2,000 of the most-viewed build-in-public posts on X, posts owning a failure had a median of about 41,000 views, while launch posts had about 17,000, the lowest of 18 formats. So a plain results post, replies where people already talk about the problem, a screenshot, one specific question and a ship log will usually do more than a second launch.

The post's title, My launch flopped. What do I post now? A 7-day plan, over Claude Monet's Stormy Weather at Étretat (1883)
Painting: Claude Monet, Stormy Weather at Étretat, 1883. Public domain.

If your launch got a handful of upvotes and no signups, don't launch again next week. Post about what happened, with the numbers, and then show the work for a week. In the posts I studied, owning a failure was read more than announcing a launch.

You're not the only one this happened to. Last week a maker posted on X that their Product Hunt launch flopped, while a two-line reply they had written about turning GIFs into WhatsApp stickers got 27,078 impressions. Every post about the thing they actually built: under 105. Today someone else asked, simply, "Which launch site sent your agent-built app zero signups?" On Hacker News, a founder used to B2B opened an Ask HN with "Product Hunt is dead" and said their Show HN threads brought mostly sales emails from other startups.

And on Indie Hackers, a founder with four upvotes, all from people they knew, put it plainly: "Nobody talks about the week after launch." So here's a week.

Why did my launch get nothing?

Mostly because there are far more launches than places to be seen. Product Hunt Pulse, a third-party tracker, counts about 196,000 Product Hunt launches in 2026 so far, against 43,027 in all of 2024. About 7,100 of this year's were featured on the front page, roughly 4 in 100. In 2023 it was 37 in 100 (numbers read on 16 November 2026).

So a quiet launch day usually tells you that few of the right people saw it. It doesn't yet tell you whether the product is any good. One Indie Hackers thread on quiet launch days splits the silence into three questions: did the right people see it, did they understand it in one sentence, and did they have a reason to act today. That's a useful way to read your week.

Is it worth posting about a launch that failed?

Yes. I sorted 2,000 of the most-viewed build-in-public posts on X, by 117 authors, written between January 2023 and November 2026, into 18 formats. The full table is in what to post when you build in public. These rows matter for the week after a flop:

FormatWhat it isPostsMedian views
FailureSomething that didn't work, owned plainly5540,510
ScreenshotThe screenshot is the post, with a few words7933,721
Lesson listNumbered lessons from a stretch of work6630,543
Ship logSeveral small things shipped, as a short list8822,278
Ask for feedbackOne specific question about work in progress1518,128
LaunchThe problem, what you built, where to get it21716,846

Launch posts were the most common of these and had the lowest median of all 18 formats. Failure posts had more than twice the median. That's the case for posting your flop instead of hiding it, and for not just reposting the launch.

Two cautions, the same ones I give with the full table. These are winners, mostly from people who already had an audience, so the table shows what popular posts look like, not what will make yours popular. And the sorting is rough, so small gaps between rows are noise. The failure row is also only 55 posts. Treat it as a direction, not a promise.

What should I post each day of the week after?

Here's a plan for the seven days after launch day. Each day has one job. Skip a day if you have nothing real for it; an empty post is worse than none.

DayPostIts job
1The results, numbers and allOwn it before anyone else frames it
2No post. Ten repliesShow up where the problem is already being discussed
3A screenshotShow the product working, with one line
4One specific questionGet feedback from the people who did look
5The one fixChange one thing and say why
6Rest, or a reply dayNothing new to say is fine
7Ship log or lesson listClose the week with what changed

Day 1: the results post

Say what you launched, where, and what came of it, in numbers you could screenshot. Then say what you think went wrong and what you'll try. This is the failure format, the best-performing row in the table above that you can write the day after a flop.

Day 2: replies, not posts

That maker's reply about WhatsApp stickers got over 27,000 impressions while their posts about the app stayed under 105. That's one maker's week, not a rule, but it fits what the Indie Hackers threads keep saying: find people who already have the problem and help them, without pitching. Search X, Reddit or Hacker News for the problem in your users' words and write ten useful replies. Don't link your app unless someone asks.

Day 3: a screenshot

One picture of the product doing the thing it's for, and one line about what it does. In the study, screenshot posts had a median of about 34,000 views, double the launches. A screenshot also answers the "did they understand it in one sentence" question better than a feature list does.

Day 4: one question

Ask the people who did look one thing you need to know. Not "thoughts?", which gives nobody anything to answer. A real question, about one screen or one sentence.

Day 5: the one fix

Change the thing the week pointed at: the headline, the first screen, the price page. Post what you changed and why. One change at a time, so you can tell what worked.

Day 7: close the week

A ship log of what you changed since launch day, or a short lesson list. Either one tells people the project is alive, which a launch post can't.

Fill-in templates for the week

These are templates, not real posts. Replace every bracket with something that actually happened, and delete any line you can't fill honestly.

Day 1, results:

I launched [product] on [where] on [day].
Result: [visitors], [signups], [upvotes or replies].
What I think went wrong: [your best guess, in one line].
What I'm trying this week: [one change].

Day 3, screenshot:

[What it does, in one line.]
[image of it doing that]

Day 4, question:

[Product] is for [who], to [do what].
People who looked at it left at [screen or step].
Question: [one specific thing you'd change, A or B?]

Day 5, the fix:

Changed [thing] after launch week.
Before: [old version].
Now: [new version].
Why: [what you saw that made you change it].

Day 7, ship log:

A week after a quiet launch, shipped on [product]:
- [change]
- [change]
- [change]
Next: [the one thing you're working on].

If a draft of any of these sounds generic, how to write build-in-public posts that don't sound like AI has a before and after you can copy from.

Should I relaunch, or launch somewhere else?

Not this week. The usual advice, from Product Hunt's own help center to the 30-day plans that rank for this, is to keep replying, follow up with anyone who signed up, fix the biggest drop-off, and launch again only after a real change. That holds up. Another launch site with the same page and the same pitch tends to get the same result.

What a flop can't fix is having nobody watching beforehand. When I built Mat, my iPhone photo widget, in public, there were 500K+ impressions and 1,700+ test users before it went live, and 1,386 users four days after. Those people had watched it get built for weeks. A launch was the last post of many, not the first.

Where Moneta fits

The hard part of this week is having something real to post every day when it feels like nothing happened. I made Moneta for that. It's a Mac menu-bar app that takes a screenshot every minute and keeps your workday in a folder on your Mac. The agent you already use, such as Claude Code, Codex or the Claude app, reads that folder and drafts posts from what you actually did, using 31 templates built from the same 2,000 posts. You read the drafts and press Post.

A step in Moneta's setup that reads: Moneta is in Claude Code. In any chat, start with Hey Moneta. Below it, the line Hey Moneta, what did I do today? with a copy button
The last step of Moneta's setup, from my design in Paper.