0 real signups in 29 days — killing FacelessFlow (the post-mortem)
I pre-committed to kill thresholds on Jul 21 so I couldn't move the goalposts: 50+ emails by Aug 9 → build the MVP, 15–49 → iterate, under 15 → kill it and publish the post-mortem.
Final count: 0 real signups. (1 email total, and it was my own test entry.)
So: kill it. Here's the post-mortem.
The idea: FacelessFlow — niche/format discovery + flat-fee rendering + official-API publishing for faceless-video creators. Landing page first, product second.
What I did:
- An X post every 2 days (automated), each one a single insight from a research drop
- Value-first posts in YouTube-growth subreddits (new account, so karma-building came first)
- A real lead magnet: 10 video formats from small channels that beat their sub count 5:1+ and their own median views 3x+, screened against live YouTube data
What I got wrong:
- Launching the page is not distribution. It sat untouched for the first week — the actual push didn't start until Jul 18, which left less than three weeks before the deadline. Not enough runway for any channel to compound.
- I never got real signal on whether the pitch itself lands, because traffic was too thin to tell "bad pitch" from "no one saw it."
What's next: no MVP, no iteration on this pitch. If I come back to faceless-video tooling, distribution goes first next time, not last.
Page (up for now, may come down): https://facelessflow-self.vercel.app
Thanks to everyone who gave feedback on the original page-copy post — genuinely useful even though the numbers didn't move.
Huge respect for actually sticking to your kill rule. Most founders (myself included) would just move the goalposts and spend another month tweaking the landing page or building features nobody asked for. Knowing when to pull the plug based on pre-set data takes serious discipline. On to the next one!
This is a hard post to write but an important one — thanks for sharing honestly.
The demand validation step is where most of us skip too fast. I made a similar mistake before and learned to check HN, GitHub, and Reddit for real complaints before writing any code.
What signals did you have that people actually wanted this before you built it?
Disclosure: three products of mine, and the relevant one is a lead-list tool - so I've already run the experiment you're scheduling for next time.
"Distribution goes first next time" is right, but a warning about the obvious way to do it. I bought 483 clicks across Bing and Google in July specifically to get past the not-enough-traffic problem. Real browsers, plausible referrers, pages rendered. Number who ran a single search in the product: zero. About a third of that site's lifetime traffic turned out to be parked-domain syndication I'd paid for.
So I got the volume and still couldn't separate a bad pitch from a bad audience. Buying traffic converts "nobody saw it" into "the wrong people saw it", and those two read identically at the bottom of the funnel.
The more hopeful read on your 29 days: the lead magnet is the asset here, and it survives the product. Ten formats from small channels beating their sub count 5:1, screened against live data - that's real research, and it's the kind of thing that earns an audience before you have anything to sell them. You clearly can find these people and you're clearly reading the niche properly. What you didn't have was three weeks. Lead with that next time and you start from a warm list instead of a cold page, and then a zero would actually mean something.
The pre-committed threshold is the disciplined part - but the bigger insight is hidden in what you're measuring. You set a bar on "0 real signups" but that's measuring the output of distribution, pitch quality, and timing all compounded together. When it hits zero, you know something failed. You don't know what.
The other comments are right that you should split reach from conversion. But there's a layer beneath that: the measurement system you chose tells you what you were actually trying to test.
By leading with the landing page, you implicitly measured "can I get people to see this and want it?" That's a compound question. Once you ask it that way, zero signups could mean:
You discovered one of those things: distribution came last. But you discovered it by process of elimination after 29 days, not by measurement.
The winning move for next time isn't just splitting the threshold. It's instrumenting each layer separately before you declare the test failed. "Did we get visitors?" is answerable in three days. "Did they click through?" is answerable in a week. "Did they read the full page?" is answerable in ten days. Only after you can measure each of those should you ask "did they convert?"
Your discipline on the kill threshold meant you didn't waste another month. That's good founder behavior. But the leverage is in building a measurement system that tells you which month to cut short, not just when."
Your threshold design is honest, and the conclusion is the right one. I made the same mistake two months ago with a different shape.
I built an agent that generates and publishes SEO articles daily — fully automated, 18 posts in two weeks. The writing was fine. What I never verified until later was how many of those URLs Google had actually indexed. The answer was far fewer than 18. An article that isn't in the index has exactly zero value no matter how good it is.
But that wasn't really the failure. The failure was that I delegated production and called it distribution. The agent could write and publish around the clock, and I treated that velocity as progress. It wasn't. Nobody was at the other end. Zero traffic, zero signal, zero revenue.
Your split between reach and conversion is the thing I wish I'd instrumented from week one. In my case, the split would have shown me on day two that no one was arriving — same as yours — and I could have killed it fourteen days earlier or redirected the effort toward an actual channel.
The one thing I'd add to your framework for next time: set the traffic floor before you set the conversion threshold. If you can't hit 300 visitors in N days using a channel you can name and measure, the experiment hasn't started yet — it's still in the distribution-discovery phase, and the signup threshold isn't the right scoreboard.
The pre-committed kill thresholds are the best part of this post — most people who set a bar like that quietly move it when the number comes in low, you didn't. One thing on the diagnosis though: you don't actually need much traffic to tell "bad pitch" from "no one saw it" — those two failure modes leave different fingerprints even at low volume. Near-zero sessions in the first place is purely a distribution problem, no signal needed. A few sessions with instant bounces points at positioning/headline. A few sessions with full scroll-through and no signup points at the pitch itself. Sounds like this was closer to the first case given the push didn't start until Jul 18 — did you have enough session data from that window to tell, or was it too thin even for that?
The thing I'd change next time is the shape of the threshold rather than the number. "50 emails by Aug 9" bundles reach and conversion into one figure, and those two fail for completely different reasons, so a zero can't tell you which one broke.
A floor underneath it separates them. Something like: 300 unique visitors on the page, and 5% of them converting. Miss the traffic floor and the experiment simply didn't run, so the honest verdict is "distribution failed" rather than "the offer failed". Hit the floor with 1% converting and you've genuinely tested the pitch. Same discipline, but now both outcomes are decidable.
The other thing worth instrumenting is the step before the page, because that's almost certainly where this one died. Impressions and outbound clicks counted per channel would have told you in week one that nothing was arriving, while there was still time to change something. Signups is a lagging measure of about five stacked things, so it's the one number you can't debug from.
One detail in your write-up is quietly good news: the single email was your own test, which means the form and the delivery path both work. That rules out the most annoying class of silent failure and narrows this to reach.
The part about not knowing whether the pitch was bad or simply unseen is probably the most useful takeaway here. Zero signups feels like a clear result, but with almost no traffic it really isn't enough evidence to diagnose the offer itself. The kill threshold still makes sense though — sometimes the disciplined decision is to stop before spending another month trying to manufacture signal.
The pre-committed kill threshold is the most disciplined thing I've seen a founder do. Most people move the goalposts. The lesson about distribution going first is one most founders only learn after killing something — respect for publishing it honestly.