I finally plugged GPT-3 into @pastelapp to generate color palettes from text prompts, and it just does exactly what I hoped it might do.
-
The next version of Pastel tackles a bunch of things, including revamped undo/redo support. I recently added pinch-zoom to the image importer, too. Paying attention to lots of little corners of the app that haven't seen love in a while — time for a polish pass now that Liquid Glass is no longer taking up all of my development time
Venturing out into the Great Unknown, I have finally made the switch from single-purchase to subscriptions in @pastelapp, for new customers, after talking about it for years.
I should have done it years ago, but the risk of the floor falling out from under me scared me off. Better late than never!
I'm much better prepared to handle the initial financial shock today, and as I've migrated everything to StoreKit 2 I've been really impressed by how easy Apple has made it to implement subscriptions
-
Venturing out into the Great Unknown, I have finally made the switch from single-purchase to subscriptions in @pastelapp, for new customers, after talking about it for years.
I should have done it years ago, but the risk of the floor falling out from under me scared me off. Better late than never!
I'm much better prepared to handle the initial financial shock today, and as I've migrated everything to StoreKit 2 I've been really impressed by how easy Apple has made it to implement subscriptions
Why subscriptions? A one-time purchase revenue graph, even for a moderately successful app, looks like this. Yet work required only increases in complexity — Pastel has had 40 multiplatform updates since it launched in 2020, including two system UI redesigns, and a whole new OS (visionOS).
One-time purchases just aren't sustainable, unless the product itself is also one-and-done — and that's just not possible on a moving target like Apple's OSes. Revenue can't be inversely proportional to time
-
Why subscriptions? A one-time purchase revenue graph, even for a moderately successful app, looks like this. Yet work required only increases in complexity — Pastel has had 40 multiplatform updates since it launched in 2020, including two system UI redesigns, and a whole new OS (visionOS).
One-time purchases just aren't sustainable, unless the product itself is also one-and-done — and that's just not possible on a moving target like Apple's OSes. Revenue can't be inversely proportional to time
I've tried to make Pastel's subscription pricing as unobtrusive as possible. As it stands:
• Existing paid users are grandfathered in and don't have to subscribe
• New users can subscribe for $2/mo or $10/year, with two-week free trials
• If you are on an OS version that I know won't be supported going forward (though I haven't decided to drop support yet), you won't see subscription options at all, only the legacy one-time purchase (now $20)TL;DR, if you already own the app, you won't notice
-
I've tried to make Pastel's subscription pricing as unobtrusive as possible. As it stands:
• Existing paid users are grandfathered in and don't have to subscribe
• New users can subscribe for $2/mo or $10/year, with two-week free trials
• If you are on an OS version that I know won't be supported going forward (though I haven't decided to drop support yet), you won't see subscription options at all, only the legacy one-time purchase (now $20)TL;DR, if you already own the app, you won't notice
All of this is subject to change in the future, but I have no plans to do so. And this model, by and large, is what I will be doing with @broadcastsapp with Broadcasts 4.
There is so much I'd like to be able to do in both apps that requires investment, and the only way I can justify it is if the revenue line is going in the right direction, otherwise that time needs to go into new products. I would like to do both things! So Pastel gets to be my test case, and we'll see how it goes
-
All of this is subject to change in the future, but I have no plans to do so. And this model, by and large, is what I will be doing with @broadcastsapp with Broadcasts 4.
There is so much I'd like to be able to do in both apps that requires investment, and the only way I can justify it is if the revenue line is going in the right direction, otherwise that time needs to go into new products. I would like to do both things! So Pastel gets to be my test case, and we'll see how it goes
Taking another crack at the 'Gallery' tab in Pastel, while I have the codebase open. The idea is to provide a vehicle to offer new/refreshed curated collections of palettes without stomping on the built-in presets. Also a way to recover deleted presets. I'm reasonably happy with the idea so far (though content and UI is still temporary). Ideally they would have illustrated key art of some kind, but in lieu of that I'll have to find some other way to visualize collections
-
Taking another crack at the 'Gallery' tab in Pastel, while I have the codebase open. The idea is to provide a vehicle to offer new/refreshed curated collections of palettes without stomping on the built-in presets. Also a way to recover deleted presets. I'm reasonably happy with the idea so far (though content and UI is still temporary). Ideally they would have illustrated key art of some kind, but in lieu of that I'll have to find some other way to visualize collections
53,000 devices have updated to the latest version of Pastel over the past day, and it has done 350K (unique) units over its lifetime (which is several multiples of Broadcasts). I've always been generous with the free tiers in my apps, but I think I might have massively overdone it with Pastel. If even 1% of Pastel's users paid for a subscription, that would be a salary unto itself and allow me do so much more with the app. Hard not to feel sheepish about how much I've left on the table
-
53,000 devices have updated to the latest version of Pastel over the past day, and it has done 350K (unique) units over its lifetime (which is several multiples of Broadcasts). I've always been generous with the free tiers in my apps, but I think I might have massively overdone it with Pastel. If even 1% of Pastel's users paid for a subscription, that would be a salary unto itself and allow me do so much more with the app. Hard not to feel sheepish about how much I've left on the table
(Broadcasts, in fact, is my third most-popular app by unit count. Fourth, if I include an app I sunset long ago. But it does make the most money, by far)
-
(Broadcasts, in fact, is my third most-popular app by unit count. Fourth, if I include an app I sunset long ago. But it does make the most money, by far)
Working on a bunch of quality of life updates and bug fixes for Pastel this week, including a little bit of new UI, and some improved behavior
-
Working on a bunch of quality of life updates and bug fixes for Pastel this week, including a little bit of new UI, and some improved behavior
It's far too early to do any real data analysis on Pastel's switch to a subscription model, but I can say that for as much as people complain about subscriptions online, the actual numbers don't seem to back it up. Actual unit sales of Pastel are up 3X from when it was a non-subscription payment, and most of them are the yearly sub not the monthly option. It's fascinating to watch. Makes it so much easier to budget development time to go towards improving Pastel, too, so I've been updating it
-
It's far too early to do any real data analysis on Pastel's switch to a subscription model, but I can say that for as much as people complain about subscriptions online, the actual numbers don't seem to back it up. Actual unit sales of Pastel are up 3X from when it was a non-subscription payment, and most of them are the yearly sub not the monthly option. It's fascinating to watch. Makes it so much easier to budget development time to go towards improving Pastel, too, so I've been updating it
@stroughtonsmith I have one minor wrinkle. When I copy the hex code of a color from Pastel to an Adobe program I have to delete the octothorpe or Indesign/Illustrator/Photoshop doesn’t accept the color correctly. I think it always changes go black because it is assuming the code is without the #. This is an Adobe problem I realize, but could we have a pref to toggle so I can copy the next code without the leading octothorpe?
-
R relay@relay.infosec.exchange shared this topic
