"4 weeks. 1 SaaS. 3 upvotes. An honest Product Hunt postmortem."

Ykiki — Fri, 01 May 2026 11:05:56 +0000

Introduction

Product Hunt taught me something this week, but it was not what I expected to learn.
24 hours after launch: 3 upvotes (one was mine), 0 comments. By every metric the launch failed.
Here is why I am not pivoting, not quitting, and not building an AI wrapper.

Who I am, briefly

I am Ykiki. AI Lead at a startup by day, indie hacker on the side. I run two small products called Respo and PaperTracker. The third one I just launched is MakerPulse — a metrics dashboard for indie makers that pulls Stripe, App Store Connect, and Google Analytics 4 data into one screen.

I built it because every Monday morning I was opening five tabs and spending 30 minutes copy-pasting numbers into a Google Sheet whose formulas I had already broken twice. The closest existing tool, Baremetrics, starts at $129/month, which is fine if you are doing $50k MRR but absurd if you are doing $50.

So I built the version I wanted. Free plan, Pro at $19/month, four weeks of evenings and weekends. Shipped on time. Launched. Flopped.

The launch numbers

Let me put the result on the table first, because too many launch posts dance around it.

Day 1 on Product Hunt
- Upvotes:   3 (one self)
- Comments:  0
- Signups:   single digits
- Revenue:   $0

I had been quietly optimistic. I imagined I would be replying to a healthy thread of comments, watching the upvote counter tick up while I pretended to do my day job. Instead I refreshed the page every 20 minutes and watched nothing happen.
It stings. Not going to pretend otherwise.

Why it did not land

I keep coming back to one observation. Scroll Product Hunt today and the front page is almost entirely AI tools. Dev tooling and indie SaaS, unless they have a striking hook, just do not get the algorithmic lift they used to.

The tempting conclusion is "I should have built an AI tool." I do not buy it. Here is why.

The AI wrapper space is saturated. Every day brings another GPT-flavored variant of an existing workflow. Differentiation gets harder, not easier.
The model labs eat their own wrappers. Anthropic and OpenAI ship at a pace that no solo dev can outrun. If your moat is "convenient interface to someone else's model," your moat is on a timer. The labs will use their own AI agents to ship the thing you wrapped, and they will ship it better.
The boring problem I am solving does not go away. Indie makers will still spend 30 minutes a week reconciling numbers across five tabs in 2027. AI does not solve that for them. A unified dashboard does.

So MakerPulse staying as MakerPulse is not stubbornness. It is a bet that non-AI problems still pay rent, and that the AI gold rush will, for a lot of indie devs, turn into the AI consolidation aftermath.

What I would actually change

Honest list, not a humble brag list.

The launch had no audience pre-built. I shipped quietly for 4 weeks and then waved at the launch day crowd. Of course nobody waved back. The makers who win Product Hunt today have spent months building an audience before launch day. I had not.

The hook was too rational. "Stop checking 5 tabs, see your metrics in one place" is true and useful and completely forgettable. There was no surprise, no contrarian take, no number that made you stop scrolling. Compare it to "I shipped a SaaS in 4 weeks and got 3 upvotes" — which, I notice, is the headline of this post. That hook works. The product page hook did not.

I aimed at the wrong day. I picked launch timing for myself, not for my audience. Tuesdays in Pacific time are crowded. A weekend launch into a smaller field probably would have given me more visibility, even if absolute traffic is lower.

I did not ship the screenshot. Until the night before launch I kept polishing the dashboard screenshot for the listing, and that polish was zero-value. The screenshot is not the product. The product is the product. Hours I spent on the screenshot were hours I did not spend telling anyone the launch was happening.

I am not sure all four would have changed the result. But I am sure they each cost me upvotes.

What I am keeping

The interesting part: there is also a list of things I would not change.

The 4-week scope. Constraint produced clarity. Every "wouldn't it be cool if…" got cut. The product that shipped is the product I would have built in 12 weeks too, except 8 weeks earlier.
The technical decisions. Supabase Auth + RLS over NextAuth, Stripe Connect OAuth over API key entry, ASC API over an iOS SDK integration. These were all chosen by one rule: what reduces friction for the user? That rule will outlast Product Hunt rankings.

The price. $19/month is correct for indie makers doing under $5k MRR. I am not raising it for revenue and I am not lowering it to chase signups. Wrong customers at a low price is worse than no customers.

Dogfooding. I am the first user of MakerPulse and the harshest reviewer of it. If I lose interest in checking my own dashboard, the product is broken. That feedback loop is faster than any user interview.

The plan from here

Product Hunt is a one-day event. SaaS is a 365-day marathon. I keep repeating that to myself.
Here is what I am committing to publicly, on this post, so I cannot quietly walk away from it:

Every Friday for a year: a build-in-public post on X, in both Japanese and English.
Twice a month: a technical or strategic write-up here on Dev.to and on Zenn (Japanese).

Within 3 months: 3-5 SEO long-form articles on the topics indie makers actually search for — "how to calculate Stripe MRR correctly," "App Store Connect API gotchas," that sort of thing.

Within 6 months: public MRR pages — let MakerPulse users opt in to publishing their own MRR badge, which doubles as distribution.

None of this is glamorous. None of it bets on a single launch day. That is the point.

If you are reading this and you are about to launch, or you just launched and it underperformed: the launch is the start of the work, not the end of it. The makers I respect most have all had the underwhelming launch. The difference is what they did on day 31, day 90, day 365.

Talk to me

Honest reactions are the most useful thing you can give me right now.

"This won't work because…"
"I would actually pay for this if it did X"
"Have you seen [competitor]? You are reinventing it"

All welcome. X DM, Dev.to comment, email — whatever is least friction for you.
Onward.

🔗 MakerPulse (free plan available): makerpulse.app
🐦 Personal X: @ykiki_dev
🐦 MakerPulse X: @MakerpulseApp

DEV Community: Ykiki