I'm a solo founder who hates being on camera. I made 40 product videos anyway. Here's what actually moved signups.
For the first 18 months of my SaaS, I shipped zero videos. Not one. Not even a screen recording.
It wasn't a strategy decision. It was vanity. I hate my voice, I cringe at my own face, and I was convinced a janky solo-founder video would make the product look smaller, not bigger. So I hid behind text — landing page copy, docs, changelog posts. All words, no motion.
Meanwhile the numbers told the same story every week. People landed on my page, read for about twenty seconds, and bounced. Trial signup rate: 2.1%. I blamed positioning. I blamed pricing. I blamed the market. I never blamed the missing demo — because I didn't want to make one.
Then a customer asked the question I couldn't dodge: "How does the thing actually work?" I had a 900-word doc for that. They wanted to see it.
So I ran an experiment. Six weeks. No new features, no new ads. Just video.
I made 40 clips — product demos, onboarding walkthroughs, and a handful of talking-head takes I was too embarrassed to post at first. Equipment budget: a $40 mic I used exactly once. Then I tracked everything.
Here's what happened:
- Landing page with a 90-second embedded demo: signup rate went from 2.1% to 4.2% in four weeks. Same page, same copy. The only variable was the video.
- The demo that converted wasn't the polished one. It was a screen recording with my voice over it, walking through a real task, fumbling included. The animated explainer I spent three days on performed worst of all.
- Onboarding beat marketing. Users who watched the 3-minute setup walkthrough activated at roughly double the rate of users who skipped it.
- Not one person commented on my voice or my face. The cringe was entirely in my head.
The uncomfortable part: I wasn't protecting my product's image by avoiding video. I was protecting my ego. Users don't need a filmmaker. They need to see the thing work in under two minutes.
I still hate being on camera. But I stopped letting that be a product decision. Most of my videos these days are generated from scripts — it's the exact problem I built vidmachine.ai to solve: turning a script into a product video without a camera, a studio, or a personality transplant.
If you're a solo founder avoiding video out of self-consciousness — what's one question you keep answering in text that you could answer in a 60-second clip instead?
Your result makes the best next video obvious: one 60–90 second ‘first successful outcome’ clip for every high-friction task, then place it at the question’s point of origin—landing page, onboarding email, or help article. I’d measure not only signups but whether support questions drop and time-to-activation shortens; that tells you whether the video is setting the right expectation as well as improving conversion.
The buried lead is the onboarding walkthrough doubling activation, not the landing page going 2.1 to 4.2. Signup rate is a number you can move with a headline; activation decides whether anyone is still there in week four, and almost nobody makes a video for it. On the stuck clip, I would not reshoot: trim to the seconds before the bad execution and end on the confirm step, because the fumbling is what makes it credible and the bug is the only part that contradicts your pitch.
Starting with a customer's question gives the video a manageable job: show one task on screen. I appreciate you sharing that you're still uncomfortable on camera. Someone reading this can try a small demo without waiting to feel confident.
The "fumbling beat polished" result matches something I didn't expect either. My only real screen recording that's gotten traction is a raw, unedited one that still has a visible bug in it — an app that says "Couldn't find Mom in contacts" immediately followed by "DONE — Executed," which is a genuinely bad look if you think about it for two seconds. It still outperformed a polished mockup I made with stock footage. I think the fumbling isn't just more trustworthy in the abstract — it's proof the thing is real, running, unscripted, in a way narration can't fake.
The part I haven't solved: that same clean-except-for-the-bug recording is now stuck. I can't post it publicly without contradicting my own "confirm before it executes" pitch, since the bug shows execution happening on bad data. So I'm sitting on my best-performing format, blocked by the one clip that would prove the case, until I either fix the bug or shoot a version that avoids that specific contact.
The interesting lesson is that video is a distribution experiment, not a personality test. Comparing signups is useful, but I would also track whether viewers arrive with a better understanding of the problem and need less explanation during onboarding. Did the videos that converted best share a common moment in the user journey, such as showing the first successful outcome or handling an objection?
I also use AI to make my demo videos.
The problem with using a SaaS for this is how does a remote service know the context of what your app does? To make a demo video it needs to know features by looking at code, drive your app to make the recordings, and then put everything together locally with a voiceover and avatar. That's why I didn't make a product out of my plugin, seemed like something that works best running with a local agent in your terminal.
The "fumbling included" screen recording beating the polished 3-day animated explainer is the real lesson here — feels counterintuitive but makes sense, since a real task with real friction is more trustworthy than something obviously staged. The question I keep answering in text that should probably be a clip: "what does the output actually look like before I commit." Curious if you found demo length mattered — did shorter clips outperform, or was 90 seconds close to a sweet spot?
The signup and activation lifts are strong, but the more important question seems to be what happens after activation. Did users who came through the video actually retain or pay at a higher rate, or does the lift mostly stop at signup?