97 Upvotes, 45 Downloads, 1 Paid User: 3 weeks after launching a native Mac app

2% is brutal but the real leak is download friction not features. a web version would probably 3x your funnel. skip the privacy mask, ship windows - those 10 requests are your actual signal

the 97 upvotes to 45 downloads to 1 paid line is the most useful thing in this post, and you read it right: the upvotes are approval, the download is interest, but the paid user is the only one that's pain. and 1 paid from 45 downloads isn't a bad conversion for week three on a local-first tool with a permissions wall, it's a signal that the core value lands hard for the right person, you just don't have many of the right people in the funnel yet. that reframes both your questions.

on features vs porting: at a 2% download-to-paid rate with this little volume, you don't yet know whether 2% is your ceiling or just your current sample. porting to Windows widens a funnel whose conversion you haven't validated, you'd be multiplying an unknown. so i'd go deep before wide: get enough Mac users through that you can trust the 2%, and learn what the 1 paid user did that the other 44 didn't. the cross-platform requests are real demand, but "i'd use it on Windows" is approval again, cheap to say, and porting is the most expensive bet you can make on it. make them wait and watch how many keep asking.

on privacy masking: i'd let your own positioning answer it. you built this local-first for privacy, that's the thesis the 1 paid user bought. so masking isn't a power-user editor feature, it's coherence with the promise that converted them, if a user has to leave a privacy tool to hide PII, the tool broke its own pitch. but the frictionless instinct is also right, so the resolution is scope: ship the dumb version, manual blur/black-box, not an editor, the minimum that keeps the privacy promise intact without becoming an image app. the test for whether a feature belongs in a minimal tool isn't "is it powerful," it's "does its absence break the core promise." masking does. a full editor doesn't.

On your second question I have a clear opinion, because I did the wrong choice here myself one time: with 1 paid from 45, porting to Windows gives you only the same leaky funnel, just bigger. Then you have two platforms where people download and nobody pays, and you still don't know the reason. The 44 who didn't pay are in the moment the most valuable thing you have. I would talk with maybe 3 or 4 of them before you write one single line of Windows code. To make the top wider has only sense when the middle is already converting.
To the masking feature and your frictionless philosophy: I run myself a tool where "no setup, no friction" is the whole point, so I feel this one. What worked for me was not to half-build the big feature, but to be honest about the limit. A small note like "we don't edit images, use [your OS tool] to blur before upload" makes the user often more happy than a clumsy editor inside the app that you must then maintain forever. A tool is allowed to say where it stops. Day-one necessity I would say no, honest signpost yes.
Curious in which direction you lean now, after the PH dust is settled.

The upvote-to-download ratio actually tells a useful story: people liked the idea but the install friction held them back. For native apps that need system permissions, that gap is almost always steeper than for web tools — it's the cost of going local-first, not a sign the product is weak.

On the privacy masking question: I'd keep it off the day-one roadmap. A blurring editor adds real surface area and edge cases that could slow down the core experience for most users. Better to ship a lightweight version — even just a right-click "hide before sharing" shortcut — once you've seen what the first cohort actually does every day.

On platform vs. features at 2% paid conversion: I'd talk to the one paid user before deciding anything else. Was it price? A specific use case? Did they fit your ICP exactly? That answer probably tells you more about whether to go deep on Mac or wide to Windows than the 10 Windows requests do.

Honest answer to your roadmap question: privacy masking is a day-one thing for any tool that handles screenshots. People don't think about PII until they accidentally share a bank statement or a Slack DM with a name in the corner. Once they hit that wall once, the tool is "the app that leaked my info" and they uninstall.

That said, you're right that a full editor is feature creep. The middle path I'd take: a one-click "blur detected text" button that runs OCR locally and pre-blurs anything that looks like an email/phone/card. 4-5 hours of Rust work, no UI complexity, kills 80% of the privacy anxiety.

On the platform question: 10 requests in 3 weeks is loud signal, but Windows users are a different audience. Your 1 paid user picked you BECAUSE local-first Mac, not despite it. Going Windows is a coin flip on whether you keep the persona or trade it for volume. I'd wait until request count hits 30 before porting.

This is a very clear breakdown, especially because you shared the real funnel: 97 upvotes → 45 downloads → 1 paid user. From a first-paying-customer perspective, the most valuable next step may not be adding more features, but understanding the exact conversion threshold of that first paid user.

I would treat that person as your first potentially repeatable buyer. They paid because, at some point, the product created enough trust and value. I’d work backward and validate three things:

1. What was the first key step they completed after downloading?
2. What message or experience made the product feel worth putting into a real workflow?
3. What moment moved them from “curious download” to “willing to pay”?

If those three points repeat across more users, they become a model for finding the next paid user. Until then, expanding platforms may just amplify the download-to-paid gap instead of fixing the funnel.

If I were pushing this forward, I’d frame the 2% conversion as two small experiments:

• Can a new user complete one meaningful outcome quickly after first use?
• Is the pre-payment trust message — local privacy, data safety, control over the result — actually felt by the user before they decide to pay?

Your breakdown here is really useful — the 97 → 45 → 1 funnel is a good diagnostic framework because each drop tells you something different. The upvote-to-download gap is partly a trust/friction problem (Mac permissions feel invasive to new users), but the download-to-paid ratio of ~2% actually isn't terrible for a local-first tool at this stage. On the cross-platform question: I'd be cautious about porting too early. The one paid user and early retention is your signal to learn what's driving that conversion before you dilute focus by expanding platforms.

That 97 → 45 → 1 is actually a clean diagnostic if you read each drop on its own. 97 upvotes to 45 downloads is fine — upvotes are the cheapest possible signal (no commitment, costs nothing), so I'd weight them near zero. The real signal is 45 downloads to 1 paid, and that gap is almost always one of two things: the value lands but price/timing is off, or people grabbed it out of curiosity and never hit the moment where they'd pay. So the question that matters isn't downloads, it's: how many of those 45 opened it twice? Activation, not download count, is the thing to obsess over at this stage. Do you have any second-session data — and what's the price point?

I’d treat privacy masking as part of activation, not just a feature. If local-first privacy is the promise, the key metric is probably not downloads to paid yet, it is how many people reach the first “I can safely use this in my real workflow” moment before you widen to Windows.

This is a really clear-eyed breakdown, the friction gap point especially. 97 to 45 is actually a reasonable conversion for something that requires a download and Mac permissions, people don't realize how much that step alone filters out anyone who wasn't already fairly serious.

On the privacy masking question, I'd lean toward day one necessity over power user feature, but only because of how you framed it yourself. If someone has to leave your app to protect their own data, that's not really a usability gap, it's a trust gap. For a tool that's positioning itself on local first privacy, shipping without that feels like it undercuts your own pitch. Power user features are things that make the tool better for people who already trust it. This sounds more like a baseline expectation for the category you're in.

On Windows versus retention, I'd actually go the other way from what most people will probably tell you. Porting to Windows multiplies your top of funnel, but you still only have 1 paid user out of 45 downloads on the platform you've already built for. Until you understand why that number isn't higher (pricing, onboarding, the gap between "downloaded" and "saw the value"), more top of funnel just means more people falling into the same hole.

I'd want at least a few user interviews with people who downloaded but didn't pay before deciding platform expansion is the lever to pull.
Following along, curious what you land on.

What stood out to me is that the post seems to contain several different lessons that could all end up competing for influence.

The paid user.

The download friction.

The cross-platform requests.

The privacy feature debate.

Each one feels capable of justifying a different next move.

That's usually where things get interesting.

The 10 cross-platform requests are the most dangerous number on this list, because requests from people who haven't paid aren't demand, they're a wishlist. I wouldn't build Windows or Linux off 10 non-buyers. The number that actually matters is buried in your funnel: 97 upvotes to 45 downloads. Half the people who liked it won't even install, and for a local-first Mac tool that install plus permissions step is your real leak. Going cross-platform doesn't fix that friction, it triples it across three platforms instead of solving it on one. I'd go deeper on Mac, obsess over the download-to-first-value moment, and get that 1-in-45 paid rate climbing before porting anything. And put the Windows crowd to a test: a waitlist with a small deposit. Requests are free, a $5 commitment tells you who's real.

Really useful breakdown. The signal I'd try to separate is "download friction" vs "workflow incompleteness."

Before choosing Windows or a full privacy editor, I'd instrument one tiny event: after someone captures/imports assets, do they come back to retrieve or share the same asset later? If yes, retention is real and privacy masking is probably a completion feature for the existing workflow. If no, a Windows port just sends more people into a leaky first-run experience.

For masking, I wouldn't build a complex editor first. I'd ship the smallest version that proves the need: one-click blur/blackout rectangles, maybe no fancy annotations. Then compare:

• first successful retrieval rate
• second session within 7 days
• % of users who use masking before sharing/exporting

That should tell you whether privacy is part of activation or just a power-user layer. My bias: keep the Mac product narrow until you can describe the paid user's repeat workflow in one sentence.