7 plugins, 328 downloads, $34.98 in the last 30 days — and the gap I only just noticed
I've been building small Obsidian plugins solo for a few months now. Seven shipped so far, all the same shape: free tier capped at 3 lifetime uses, paid Pro unlocks unlimited via a one-time Gumroad license. No server, offline license check, BYOK for the AI part.
Last 30 days on Gumroad: 295 visits, 2 sales, $34.98. Across the directory, 328 total downloads spread across the 7 plugins.
The thing I spent the last week fixing wasn't a new feature. It was something dumber than that.
When someone hit the free-tier limit, the plugin showed a toast notification. No link. No "here's how to upgrade." Just "free limit reached" and then nothing. I'd written the whole licensing system — Ed25519 keys, offline verification, the works — and then put zero effort into telling anyone it existed once they actually hit the wall.
I noticed it on one plugin first, while checking something unrelated. Then I went and checked the other six. Same thing, every single one.
So this week was: add a proper upgrade prompt (a modal, not a notice that disappears in five seconds) plus a persistent "upgrade to Pro" section in settings, across all seven repos, one at a time. While I was in there I also found a couple of free-tier counters that weren't as bulletproof as I thought — one was quietly resetting monthly instead of staying lifetime, and at least one had a real race condition where clicking fast enough could squeeze out a 4th free run. Fixed those too.
None of this is live long enough to show up in numbers yet — I only finished the last plugin today. So I can't tell you it worked. I can tell you the gap was real and embarrassingly obvious in hindsight: you can't convert someone to a paid plan if they never see that the paid plan exists.
Will report back once there's actually something to report.
The "free limit reached and then nothing" gap is the kind of thing that's invisible until you see it, then can't unsee it. Good catch. But the deeper issue your numbers point at: 295 visits to 2 sales isn't only an upgrade-prompt problem, it's a "hit the wall hard enough to care" problem. 3 lifetime uses might be too few to prove value, so people bounce before they're hooked, not because they didn't see the upgrade.
Worth testing: does a slightly more generous free tier (enough to get someone dependent on the plugin) convert better than 3-then-paywall, even with the new modal? The upgrade prompt matters, but only if they've felt the value first.
Which plugin gets the most downloads? That one's your real product, the other six might be dilution.
Fair point on the free tier — but the data doesn't quite support it in my case. The plugin with the most downloads (109) is AI Journal Coach, which has zero sales. The two that actually sold have fewer downloads each. The pattern I keep seeing is that tools working on accumulated external backlogs (articles, highlights, videos) sell; tools that process the user's own writing don't, regardless of how polished the upgrade prompt is. So I think the 3-use limit isn't the bottleneck here — category fit is.
Which is a different (and harder) problem.
That's a sharper diagnosis than mine, and the data backs you. "Tools that clear an external backlog sell, tools that process your own writing don't" is a genuinely good insight, and it's really a willingness-to-pay pattern.
Underneath it: backlog tools attack a pain that's already piled up and visible (147 unread articles, a folder of highlights you'll never revisit). The pain pre-exists your plugin, so the value is instant. Tools that process the user's own writing require them to first generate the input, then feel a pain they may not have yet. You're asking them to create the need before you can solve it. Backlog = existing pain you relieve. Journal coach = latent pain you have to convince them they have. The first sells itself, the second needs education, which is expensive and rarely converts at a one-time price.
So AI Journal Coach at 109 downloads and zero sales isn't a pricing or prompt failure. It's that "coach my own writing" is a vitamin and "clear my backlog" is a painkiller. Same build effort, completely different willingness to pay.
I ran your situation through Hivemind (the AI strategy tool we're building) to pressure-test this, and the read was blunt: stop spreading effort across seven plugins, double down on the backlog category where buying intent already lives, and treat the other six as evidence that polish doesn't fix a vitamin. If you want to point it at your own positioning: https://hivemind.myosin.xyz
The question that decides your next month: within backlog tools, which specific backlog do people already pay to clear elsewhere? That's the niche to own.
That's a clean way to put it — painkiller vs. vitamin explains the pattern better than "category" did. Matches what I'm seeing exactly: the two plugins that sold both clear an external backlog (reading list, meeting notes) that already existed before the plugin. The one that's asking people to generate new writing first and then value my feedback on it — that's the vitamin, no argument there.
On the "which backlog" question — already mid-repositioning one plugin around exactly this. Read-it-later services (Pocket, Omnivore) shut down over the past year and a lot of people migrated their reading into Obsidian without solving the actual backlog problem, just relocating it. That's the angle I'm leaning into rather than picking a brand new niche.