I built six Obsidian plugins. Two sold. I couldn't have guessed which.
Quick context: I'm a textile engineer with no coding background, building and selling small digital products with AI. Here are my real numbers from last month, because the split surprised me and might be useful to someone.
Across all six plugins: 204 views total, 2 sales, around 35 dollars. Not a figure to brag about. But the total isn't what kept me up. It was where the sales landed.
All six plugins run on the same idea: take the scattered material in your vault and turn it into one clean summary with AI. Technically they're siblings. On the sales side they're nothing alike.
I split them in two. One group works on stuff that piles up from outside you: unread articles, web clippings, book highlights. The other works on notes you produced yourself: journal entries, meeting notes, weekly periodic notes.
Here's how it came out. The outside-the-vault group (Reading Inbox, Literature Review, Highlight Inbox) pulled 52 views and made 2 sales. The your-own-archive group (Periodic Notes, AI Journal Coach, Meeting Notes) pulled 152 views and made zero.
Three times as many people looked at my own-archive plugins, and not one bought. My most-viewed product was Meeting Notes: 68 views, a flat zero.
My theory, and I'm not sure of it: people pay to clear a pile that builds up from outside, because that pile produces guilt. "I was going to read those 40 articles. I didn't." That pressure is real, and a wallet opens for something that lifts it. Synthesizing the journal or meeting notes you wrote yourself doesn't create the same pressure. Most people don't even keep those notes regularly, let alone buy a tool to process them.
Honesty check: you can't do statistics with two sales, n is tiny. Maybe the problem is my positioning, not the category. But 152 views and zero sales stops feeling like a coincidence.
And the products I was proudest of weren't even these. I had SaaS boilerplates I'd put together over a few days. They didn't sell either, but I'll be fair: they only pulled 27 views total, because they have no community sending traffic the way the Obsidian plugins do. So their zero isn't the same evidence. Still, the same shape keeps repeating: small, specific tools with a community around them get seen and sometimes bought. Big do-everything stuff has neither.
Now I'm hunting for the next inbound thing: unwatched videos, the podcast queue, the unopened newsletter folder. Same guilt, and I already have the engine that lifts it.
Full breakdown with the per-plugin numbers here: https://ibrahimharbutlu.com/writing/six-plugins-two-sold
Anyone else seen this split between "proud of" and "actually sells"? Curious if it's just me.
Hi, the guilt-pile framing is one of the cleaner product insights I've seen articulated on here. People pay to make something stop accumulating, not to do more with what they already have.
It maps to something I've noticed on the automation side too. The workflows that actually get built are the ones solving for a pile that grows without you — unread leads, unanswered follow-ups, orders nobody tracked. The ones that would make existing work more meaningful? They stay on the wishlist.
Your meeting notes zero doesn't surprise me for another reason: most people who would buy a meeting notes synthesizer don't keep consistent meeting notes to begin with. The tool presupposes a habit that the buyer hasn't formed. The reading inbox doesn't — the pile is already there whether you're organized or not.
The podcast queue and newsletter folder angle you're hunting next feels right. Both have the guilt property and neither has a clean tool yet.
Curious whether you've looked at the n8n or Make side for any of this — some of those same guilt-piles (newsletter backlog, YouTube queue) are interesting to hit at the workflow layer rather than inside Obsidian specifically. If you want to think through it, happy to connect: https://teams.live.com/l/invite/FBAk3iOSJkDyS11JQ?v=g1
dead on about the meeting notes. i built it for a version of the user i wish existed, not the actual lazy user. the n8n/make angle for newsletters is fascinating. the trade-off there is distribution. building an obsidian plugin means you get immediate eyeballs from their internal community store. building a standalone n8n workflow means you have to drive your own traffic from zero. but for the newsletter problem specifically, fixing it at the routing layer rather than the storage layer is probably the only way that actually works. appreciate the link, i'll reach out when i get to the newsletter phase.
The guilt-driven purchase pattern makes a lot of sense. People pay to clear piles they feel bad about ignoring, not for tools that improve things they already manage.
I'd bet the video queue and podcast backlog will follow the same pattern. The guilt is already there, you just need to show up with the fix.
Have you started building for those yet?
That's the exact hypothesis. i haven't built them yet, but the youtube 'watch later' playlist is currently at the top of the backlog. processing text was cheap and easy in obsidian, but shifting that same synthesis engine to handle audio/video inputs is the next technical wall i need to climb with the ai tools i'm using. definitely the next logical step though.
The YouTube watch later list feels especially painful because people keep adding to it faster than they consume it.
What I'd be curious about is whether users actually want the full content processed, or if they just want confidence they're not missing something important. Those are very different products.
I've definitely caught myself saving videos with every intention of watching them later, then never touching them again.
this is the exact pivot i needed to hear, honestly. building a tool that tells you "you didn't miss anything, it's mostly filler, safe to delete" is probably way more valuable (and cheaper on api calls) than generating dense summaries. shifting the focus from 'learning' to 'clearing the backlog' changes the whole product. appreciate the insight!