Get sustainable funding

[ErikBjare]

In pursuit of securing the financial sustainability of the project and accelerate development we'd really benefit from getting funding one way or another.

Avenues:

• Research grants
  • The most likely source that I can imagine right now in the projects current shape and focus. We should ask the SSB people who seem to have extensive experience with this.
• Open source grants:
  • FUUG.fi, see Submit FUUG.fi grant application #343
• Donations from users
  • The only source right now, unsatisfactory as it only barely covers current costs (not including development costs).
• Consulting
  • Similar to the model used by Syncthing. Helping companies use ActivityWatch.
• Building an enterprise-focused version
  • Similar to the model used by GitLab. Might follow consulting efforts, we've developed a basic business plan in a course I took but not sure how good it is.
• Raise money from investors
  • We'd need to develop a better business plan for this.

Related to #236

[xylix]

The FUUG application has been sent and waiting for the outcome should probably happen here instead of in that issue. I'll close #343 , and update here with their response when it happens (could take a few months).

[ErikBjare]

Update on the fuug.fi grant: We got 3000€ from them!

Huge thanks to @xylix for pulling this across the finish line, and to @teklak for helping me write the first drafts of the application.

Now we need to:

• Write a short blurb for their website (after which they'll send us the money)
• (Once the money is spent) Write an update on what the money was used for

[xylix]

Blurb now written, waiting for their approval and comments.

[ErikBjare]

I just applied to Mozilla Fix the Internet Spring MVP Lab (thanks for the tip @Powersource!).

The internet's amazing, weird, sometimes a little icky. Let’s face it: it could use some fixing. Are you a student, coder, designer, creator who wants to help us build new services and platforms that foster collaboration during this crisis, or put privacy first, or combat filter bubbles and polarization, or AI platforms that work for the end user, the citizen? Mozilla’s Fix-The-Internet MVP Lab is an 8 week-long incubator-style program this Spring to mobilize & fund around products and technologies that enable everyone to connect and build a better society.

See the attached application: fix_the_internet_spring_mvp_lab_response_3B4HN8suUmty.pdf

In the best-case scenario that would give us 2500€ per person (me, @johan-bjareholt, and @xylix) as well as being part of their incubator program (which would definitely be interesting!). There's additionally prizes for the best projects:

First: $25,000, Second: $10,000, Third: $5,000 (per team)

Update: We did not receive the grant, looks like it wasn't a good fit for us.

[CrazyPython]

I have a suggestion: Sell ActivityWatch.

I don't mean to make it closed-source. I mean to sell the bundle of implied support and quality guarantees inherent when you sell something.

I bought another month of Timing instead of using ActivityWatch because of #380. Today I read that and found the workaround.

Most users do not have the technical savvy or the patience to do like I did.

When I pay for software, I expect support and the guarantee that it will work. I would expect someone to show me step-by-step instructions on how to fix my issue, linking me directly to the workaround. So I suggest selling (guarantee of support and assurance of continued development of_) ActivityWatch.

[pcuci]

Yeah, it's tricky @CrazyPython - I think https://signal.org/ has an embedded privacy-preserving org structure that's got a good defense against market forces.

Here's @dhh's take on the whole lot of reciprocal vs. permissive licenses wrt. to the idea of debt I think you're proposing. RailsConf 2019 - Opening Keynote by David Heinemeier Hansson

Part of the reason I use AW is precisely because it is free self-hostable software, anything less would be a trojan horse of sorts, orthogonal to the project's underlying privacy-preserving principles.

The choice of MPL I find interesting. Personally, I would've went for an MIT-esque https://ethicalsource.dev/ license to prevent, say, corporations spying on their employees with AW... but that's a different can of worms, not unrelated to the debt/business-model conversation :-)

The other part of why I am comfortable using ActivityWatch to spy on my own behavior and gain personal insight into my own psyche is because @ErikBjare and @johan-bjareholt and the contributors are essentially gifting it to everyone, with no expectation of this being a contract with us users.

Can an ethical business be created on top of this tech? Sure. I might even trust it more as it's going to be deployed at larger scale (though it would arguably be more secure as a fringe project; YMMV depending on your threat model). Moreover, folks from https://tidelift.com/ and https://opencollective.com/ are constantly working at making f/loss more (financially) sustainable.

@CrazyPython we can't assume the developers want the headache of user support; quite the opposite might be true. I also wouldn't be surprised if it is precisely because of this freedom-from-user-support that the project ends up being more creative and keeps growing; which reminds me of https://comma.ai/ 's business model and approach—there might be some lessons there to port over here.

Think of why more people donate more blood in countries where they don't get paid for it. Strange? Perhaps. Also true.

[CrazyPython]

The other part of why I am comfortable using ActivityWatch to spy on my own behavior and gain personal insight into my own psyche is because @ErikBjare and @johan-bjareholt and the contributors are essentially gifting it to everyone, with no expectation of this being a contract with us users.

I didn't say to make it closed source. I said, keep it open-source (OSI, FSF approved), just don't provide a download link to pre-built binaries. Nothing in the free software (F in FOSS) definition says you can't charge for a download. Other people are still allowed to host the binaries and have full freedom to do so. People who impulse buy, are too lazy to search for a free download, want the faster download speed, want to support you, etc. are much more likely to give you money if you ask for it up-front.

GCC, a FSF project, was once sold on CD-ROMs for hundreds of dollars. It was still FOSS.

In other words, don't restrict it to paying users, but don't provide a link to download it for free either! They can still legally get a copy from a friend.

[ErikBjare]

@CrazyPython I think that's a reasonable strategy, but I don't think we'll ever stop publishing binaries for free.

However, I do agree that there is probably a large part of the user base who would gladly pay, if only we were more upfront about asking for it.

I'm kinda thinking of the WinRAR case here, where you didn't really need a license (for non-enterprise use), but the license nag made a lot of people buy a license anyway! (although I hope we can improve upon the notoriously annoying "license nag" UX)

I should add that I really appreciate you both engaging on this very long-term strategically important issue for the project. Very helpful to have this discussion!

[pcuci]

@CrazyPython why not fork the project and do exactly what you're proposing? That's what I'll do should the binaries become unavailable here. (CentOS vs. RedHat)

You are free to build a profitable business with all the ActivityWatch code. My point was that the maintainers might not want that responsibility associated with running a business, and be subject to negative user reviews, regulation, etc. - I can respect that choice.

@ErikBjare there are a few data-driven business models to perhaps consider:

To @CrazyPython 's point, there might be an equivalent of a default donation to developers https://github.com/muhamadridwanid/Masari-Mining-Pool (notice the 0.2% configuration default).

We can think of users working on their computers as "work"/"mining"/"creating data" and that data has value. So they're not users, they are creators(!), let's call them "players" for lack of a better term. Perhaps the players want to sell some of their data? I wouldn't mind selling some aggregate of my behavior to trustworthy parties that guarantee to not resell my data further (it is my data, my body clicking on the mouse button created it), re: https://www.goodreads.com/book/show/51781479-privacy-is-power - the infrastructure for this already exists: https://www.humanapi.co/ (simply plug in the AW data stream, and charge 10%)

Also, the ideas expressed in the MDL https://www.montrealdatalicense.com/en might be informative towards understanding the scope of data licensing

I would be comfortable allowing my de-anonymized data to be used for general Research purposes, and publishing, even if the researchers make a profit.

I wouldn't want to contribute to either output or model commercialization, however, without express privacy guarantees and fair compensation.

Turn AW into a data mining operation. I suppose that's the business idea I'm trying to articulate here. :-)

Of course, do it in an ethically responsible way to build a privacy-focused brand.

[CrazyPython]

why not fork the project and do exactly what you're proposing? That's what I'll do should the binaries become unavailable here. (CentOS vs. RedHat)

Keep dev builds available as FOSS on GitHub releases. In fact, making binaries non-redistributable makes software nonfree, so I would advise against a RedHat strategy.

That's what I'll do should the binaries become unavailable here. (CentOS vs. RedHat)

It takes time and effort and marketing energy. How will people visit your site instead of AW's? How will you pay the hosting fees?

I am not sure I agree with @pcuci's idea, but perhaps you could reuse code from Brave's heavily scrambled Privacy-Preserving Product Analyics. brave/brave-core#3242

[CrazyPython]

I am on macOS Mojave. I would have paid for a download of ActivityWatch with this workaround #380 (comment) via an AppleScript in Login Items. Even as a developer, I just do not like the mental overhead of finding workarounds on GitHub and applying them.

That AppleScript and corresponding Login Item would not have taken the maintainers or volunteers long to write.

This is an example of why someone would pay directly.

[CrazyPython]

I will pay the developers of ActivityWatch for:

• fixing easy bugs that hurt my UX
• implementing small features I need
• improving UX in small ways
• discussing my needs as a user

In return, the project receives:

• bug reports from a real user
• feature requests from a real user
• enthusiastic user
• funding

I alone can only provide so much funding. Please join me.

My main requests are these, but would like to discuss it with devs:

• UX improvements for adding categories
• Coloring activities by category instead of application color
• Make it possible to make categories from the timeline view (tooltips icons)
• Display icons and use icon colors

(My requests don't have to go in the main app, they can be enabled via plugin or setting.)

[Potherca]

regarding "Research grants", I don't have any concrete suggestions (yet) but since this tools is aimed at being open and de-centralized, you might want to look into grant regarding breaking out of silos, privacy and user-owned data, data portability, and grants regarding decentralized web.

Dapsi comes to mind, but I am sure there are others!

[zplizzi]

I would suggest trying the "issue bounty" approach as an additional source of funding. I see there may have been some attempt at this, but the integration (and promotion) of this approach seems limited. Eg I placed a BountySource bounty on #751, but nothing really happened. There should probably be some automatic integration that posts a message on the issue like "@zplizzi placed a $50 bounty on this issue. If you would like this issue solved faster, you can contribute to the bounty, or if you are able to solve the issue, you can collect the bounty.". You could also add something to the issue template to inform people that they can place a bounty on the issues they open, and on the readmes/donation pages (which would help raise visibility). Plus, official recognition of the program would make a potential bounty-poster more confident that the bounties are a meaningful way to incentivize work.

I also don't love BountySource - they make you pay upfront to place a bounty, which means you have to lock up money on issues that may not be solved for a long time/ever - this definitely makes me less likely to place bounties. There's a bunch of other options (and I don't know which is best), eg boss.dev seems to charge once the issue is completed.

In general, I would be willing to pay ~$50 to fix simple issues, or put up a similar amount as part of a shared contribution to larger features. And I'd guess others would be too - if it felt like the bounty actually incentivized work on the issue, was promoted so that others would pitch in collectively, and charged when the work was done.

[pktiuk]

I think that idea of @zplizzi about issue bounties is very good, but it would be very difficult to select a good platform for posting bounties. Unluckily these platforms are not very popular ( Gitpay, Gitcoin )
boss.dev you mentioned looks abandoned, their issue site is not very active, it is a similar case to IssueHunt.

[zplizzi]

Good point, BountySource seems to have gone downhill (lots of reports of them not paying out bounties, etc). Gitcoin might be the best of the bunch, they are actively paying out millions/year (https://gitcoin.co/results) in bounties, although the UI/UX isn't great - feels like it was designed to support contract work/contests in addition to github issues, and thus is a little less streamlined (plus needing to use crypto). Not sure what level of integrations they provide, if any.

[snehal-shekatkar]

@ErikBjare When it comes to user donations, which channel is preferred? For example, OpenCollective or GitHub sponsors?

[Jeremiah-England]

I don't see @polarsource mentioned in this issue yet.

[pktiuk]

Polar looks great, but this is still just an alpha.

[hmijail]

I've been looking for some way to offer a bounty / sponsor issues, but looks like there is nothing current nor maybe much interest? Particularly, even the gitcoin link in the Donate page seems broken.
So I'll just donate.

[kleenkanteen]

I currently pay $120 yearly for rize. I switched from rescue time because it was when seeing aggregate data (top apps used this month) and the UI sucked.

I'm looking to change because 2 reasons. 1. It seems overpriced for just a time tracker. 2. A simple feature I want, which is my detailed activity log going top down as beginning of day to end of day. Currently it's end of day to beginning of day so I have to scroll all the way down then go up to read it in the right order. Seems like a 1 line code change to change the sort order, but on the canny board my feature request has only 3 upvotes so I understand it's low priority. If I had a open source tracker, I could make the change myself. If this project manages to do that I'm willing to switch to it and pay $60 a year for it to get priority feature request and bug support.

[BelKed]

I currently pay $120 yearly for rize. I switched from rescue time because it was when seeing aggregate data (top apps used this month) and the UI sucked.

I'm looking to change because 2 reasons. 1. It seems overpriced for just a time tracker. 2. A simple feature I want, which is my detailed activity log going top down as beginning of day to end of day. Currently it's end of day to beginning of day so I have to scroll all the way down then go up to read it in the right order. Seems like a 1 line code change to change the sort order, but on the canny board my feature request has only 3 upvotes so I understand it's low priority. If I had a open source tracker, I could make the change myself. If this project manages to do that I'm willing to switch to it and pay $60 a year for it to get priority feature request and bug support.

I suggest you open an issue describing exactly what you want your feature to look like :) And maybe take a look at the ActivityWatch to see if something similar has already been implemented ;)

[kleenkanteen]

I'm trying it out rn. So far looks a bit limited and worse in UI. Will use it for a couple days.

…

On Mon., Jun. 3, 2024, 3:22 a.m. BelKed, ***@***.***> wrote: I currently pay $120 yearly for rize. I switched from rescue time because it was when seeing aggregate data (top apps used this month) and the UI sucked. I'm looking to change because 2 reasons. 1. It seems overpriced for just a time tracker. 2. A simple feature I want, which is my detailed activity log going top down as beginning of day to end of day. Currently it's end of day to beginning of day so I have to scroll all the way down then go up to read it in the right order. Seems like a 1 line code change to change the sort order, but on the canny board my feature request has only 3 upvotes so I understand it's low priority. If I had a open source tracker, I could make the change myself. If this project manages to do that I'm willing to switch to it and pay $60 a year for it to get priority feature request and bug support. I suggest you open an issue describing exactly what you want your feature to look like :) And maybe take a look at the ActivityWatch to see if something similar has already been implemented ;) — Reply to this email directly, view it on GitHub <#259 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQF3ZKXKZGWJETTVK3YM6MLZFQKTBAVCNFSM4GNGDPSKU5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TEMJUGQ2DKOBQG4YA> . You are receiving this because you commented.Message ID: ***@***.***>

[brayo-pip]

An alternative could be charging for cloud storage of the db + syncing features. Obsidian does the same, I think enough users would be willing to pay for such a feature. This is of course way later in the future after users have an "AW account", encryption is implemented or some other way to preserve the privacy of the data stored. The db size is small, I bet under 50MB for most users, cloud storage would be cents per user for the entire year.

[pcuci]

charging for cloud storage of the db + syncing features [when] users have an "AW account"

so that I now must watch for a https://en.wikipedia.org/wiki/Warrant_canary ? No thanks. 😄

this business model seems antithetical with self-hosting, which is where the value of AW is, IMO

if you really want this setup, it is fairly trivial to put together with Pulumi/Terraform on your public cloud of choice, today. YMMV

[brayo-pip]

Depends on how it is set up, with the right protocols in place, the information is useless without the decryption key, assuming we decide encryption is the way to go. Just as secure as self hosting in that the authorities could just storm into your home and take the hard drives themselves. If encrypted they won't prove much use to them. Also note that this will be an option to reach more users that aren't as techy savvy. Not everyone can manage to even set up aw-sync with just rclone or syncthing. Anyone with privacy concerns or any other reason will still be able to self-host.

[melroy89]

Try to look how Bruno is doing.. Bruno is going great! Tons of people using it, while they provide a golden edition (which is a small fee for some premium features). Golden edition last forever. I think money wise Bruno is big business now. While also still providing an open source solution mostly under MIT. https://github.com/usebruno/bruno

EDIT: Also read: usebruno/bruno#269