Fighting for Open Source Sustainability: Introducing Code Sponsor | Hacker News Reader

Fighting for Open Source Sustainability: Introducing Code Sponsor

medium.com

69 pointscavneb31 comments

The authors are mistaken about the impact of employer-paid open source. It is orders of magnitude bigger than anything else, and it has a decades-long history of success in projects like Linux.

Companies aren’t doing it out of charity and don’t see it as such. They see (correctly) that it’s easier to get what they want by paying some people to work on an open source project than to start from scratch and pay for 100% of the cost.

I’m all in favor of more models, but until you’re channeling many tens of millions of dollars per year, your impact is much smaller than employer sponsored open source.

This is great feedback. Thank you.

Full disclosure: I'm a long time sponsor (https://rollbar.com/) and supporter of Code Sponsor. It's a fantastic resource for the Open Source community, notably the smaller, independent projects. Supporting OS sustainability while generating awareness for Rollbar was a total no-brainer decision. The transparency in the system between sponsor and project owner is exceptional. They've built an interesting, novel solution to help ease any burden felt, maintaining and growing a project. Bravo!

Maybe I'm missing something, but I couldn't find a source for their claim that "Two-thirds of the top OSS projects are maintained by one or two people".

How many projects? How are they rated "top"? When you make such a bold claim, you should be able to back it up with verifiable sources IMHO.

Also, being a Red Hatter that has welcomed thousands of people in our New Hire Orientations, I dare say that we actually do employ quite some developers that continue their good work in a lot of communities. Describing that as "the 1% approach" feels a bit wrong.

This data comes from a talk given by Nadia Engbhal entitled “Consider the Maintainer”. This data originally comes from here: https://github.com/aserg-ufmg/Truck-Factor/blob/master/READM...

Thanks. Found the slides. But they also don’t give any link to rating criteria or a list of projects.

https://www.slideshare.net/NadiaEghbal/consider-the-maintain...

Nadia left her job in venture capital last May and started work on a thesis. What is not venture backable in tech right now, that tech absolutely cannot do without?

Projects that had present, not future, value. Were actively used by Facebook, Instagram, Pinterest, Netflix, even governments. Directly responsible for tech’s rapid rise, as Mark Suster once explained but they hadn’t captured the financial value they deserved.

Capturing of financial value for these projects is extremely important yet isn't realised and can have devastating effects. Remember heartbleed?

I encourage you to speak with a maintainer heart to heart of any moderately successful opensource project to tell it like it is. Here's a 4min clip from the PythonAU keynote delivered by Russell Keith Magee that does exactly that.

https://youtu.be/0t85TyH-h04

It's a never ending stream of notifications that contain abuse and demands for unpaid work with occasional moments of the highest of highs (ie someone sending you a thankyou postcard via snail mail or asking how they can help you out)

Open-source keeps burning out brilliant software engineers. For the longest time opensource has been talked about user freedoms but has been mums the word to "at what cost"? Production is still expensive and the cost is too high.

0) https://youtu.be/VS6IpvTWwkQ

1) https://www.fordfoundation.org/library/reports-and-studies/r...

2) https://medium.com/@nayafia/how-i-stumbled-upon-the-internet...

3) http://nadiaeghbal.com/oss

> "... while part goes to maintaining Code Sponsor itself."

If their intentions are "pure" and their primary purpose really is to help open-source developers, they should either be a non-profit and/or completely open and honest and publish the sizes of each "part". Preferably, that would look (at minimum) something like:

  October 2017

  Income from sponsors:  $ x
  Paid to developers:    $ y
                         -----
  Kept by CodeSponsor:   $ z

Where z == x - y, of course.

  October 2017

  Income from sponsors:  $ 7,437.20
  Paid to developers:    $ 3,413.20
                         -----
  Kept by CodeSponsor:   $ 4,024.00

This does not take into account the $6,000 we spent to become a sponsor for Github Universe.

With those numbers, you kept about 54% of October's income.

So how is that split determined? That is, is it a fairly flat overhead so if the income had been $100,000, you still would have kept ~$4000? Or are you taking a percentage and it would have been ~$54,000?

For example, the sponsor pays $1,000 for 1 month of sponsorship. The developers gets 40% of that ($400). The remainder ($600) goes into building Code Sponsor.

Understand that this has been a side project since day one. I have not taken any distributions at all.

The split used to be 50/50, however I could not sustain myself with that split and still keep the CPA low for our sponsors. We are now setting a minimum CPC for all new sponsors to $1.50+. This means that the developer will get $0.60/click.

Finally, part of the 60% goes into the server costs and fraud prevention. We run every click through vigorous fraud detection prior to marking it as “payable”. By doing so, we provide much higher quality clicks to the sponsors. This makes them scale their funds. Remember, we are working with marketing funds, not sponsorship funds.

Thank you for answering. I think it's perfectly reasonable at the current income level, since if that's a typical month's income it would mean that you're still only taking in about $50k/year, which is difficult to do much with.

However, if you're intending to keep that split as revenue increases, I think it conflicts with the way you're framing the company in places like the linked post and the Code Sponsor homepage. Stating a mission like "fighting for open source sustainability" and remarks like "Why do we do it? Basically, we want to help" don't mesh with the fact that your company alone keeps more of the money than you pass on to all of the projects combined.

I recognize that Code Sponsor isn't a charity/nonprofit, but it feels like you're trying to give the impression that it's a similar sort of values/mission-based organization. So it's not a fair comparison, but as a point of reference, Charity Navigator would give a financial score of zero to an organization that only used 40% of their funds for their "purpose", and they'd be solidly in the bottom 10% (https://www.charitynavigator.org/index.cfm?bay=content.view&...).

Overall, your take doesn't seem to be out of the ordinary for advertising. The World Advertising Federation says that 60% of advertising cost is taken by various companies that an ad passes through before it's actually displayed (https://3.bp.blogspot.com/-AkjTy8a6FJ8/WGqb2nGR0pI/AAAAAAAAF...), but those companies certainly don't try to describe themselves as philanthropic ones.

I hope I don't come off as too negative. I think it's great that you're working on improving the current situation of open-source sustainability, which really is a major issue. But I hope that if your income reaches the point where Code Sponsor is sustainable (which certainly includes paying yourself), you consider adjusting the split so that more funds are given to the projects you want to help, instead of just keeping it as profit.

Hi Diemorz,

I’m very glad for your comment. This is something that I can honestly say I have not considered. My biggest fears were 1) making sure that the whole idea isn’t rejected by the developer community and 2) making sure I have enough money in place to give to developers.

I started off initially by doing a 50/50 split with developers. I also started by charging $1/click to sponsors. I quickly realized that it would take a very long time before I could get enough sponsors and developers to support myself full time. Because of this, I dropped the distributions to 60/40. It still is not making much money at all. Code Sponsor may take 60%, but I take 0%.

Now in the future, if it does continue to grow at a healthy rate and if Github will allow us to sustain OSS through README’s, then I would definitely consider a more developer-sided split.

All I want is to help sustain OSS and do it as a living.

Thanks for your openness about the finances, and thanks for building something to address this issue, whatever the difficulties and drawbacks are.

Thank you for the transparency! Apologies if this is already posted somewhere and I just didn't see it.

Oh.. it wasn't. I posted it in response to your query :-D Like I said, I'm an open book. I just want to help sustain OSS.

I've been using Code Sponsor for a month or two now.

The best way to describe it is something between a generic Google Ads network and a invite-only Carbon Ads network. And it's really meant to go in your README.md file.

So this results in a few things:

1. Broker. Sponsors can select which repos/languages/tech/etc they want to advertise to and developers can select which sponsors are relevant to them.

2. Transparent. Developers can see which ads a sponsor will display and how much they pay per click (usually $0.40). The ads are usually just text with a couple emoji that fit well into the README.md design and are loaded as an svg.

3. Limited. Sponsors pay a fixed amount so if a repo is really popular and displaying ads, it may stop displaying after the money is paid out to the developers. This means you may just get a generic ad that recommends using Code Sponsor near the end of the month, like this repo: https://github.com/styfle/geoslack

Caveats: it seems the Code Sponsor is not open source :(

After reading the article, I _think_ this is an advertising network for oss project owners. Not a bad idea except why aren’t existing ad networks meeting this need? And...if most of the pages displaying these ads are served by GitHub, isn’t this stealing money GitHub has deliberately left on the table, and won’t they be upset by that and hence take action? Perhaps the plan is to succeed so quickly that GitHub buys them?

> why aren’t existing ad networks meeting this need

Because most ad networks are data-gathering scum, and developers know this, thus avoid them at all costs.

This company is at least _promising_ not to track the ads, and only provide basic and relevant conversion data to advertisers. This is a big difference that would mean at least I would be willing to give them a try.

Owner here. I agree 100%. From day one I have been promised to never do any tracking or remarketing. I don’t sell data. I have followed the Ethically advertising guidelines established by Read the Docs. If you are a developer on the platform, you know to the second what your payout is. If you are a sponsor, you know to the second how much you owe.

Owner here. Github is responsible for about 60% of the overall traffic. We support websites as well such as https://jsbin.com and https://codesandbox.io.

I don’t believe Github would ever want to buy Code Sponsor. The goal from the beginning was to provide scalable sources of funding. This has not existed prior.

We do not consider ourselves an advertising network. Here's why:

  * We do not display sponsors based on a bid system.
  * We only present sponsors that the developers choose.
  * Sponsors can select which projects they want to support and which they do not.

I'm a big fan of Code Sponsor! Eric who created Code Sponsor is passionate about his project and surely at your disposal for any query.

If you have a project with more than 100 stars, try it, you'll really appreciate the transparency and the personalization of the ads you can have. Long life to Code Sponsor!

Does anybody know what the click / view ratio is like on small projects? I know it depends on a lot of things but I really have no idea at all... Would it be below one percent or more around five, or something different?

I have one project[1] with 0% CTR.

I have another project[2] with 3% CTR.

[1]: https://github.com/styfle/copee

[2]: https://github.com/styfle/magnemite

Both have less than 200 impressions total.

Depending on the project, we have seen between 0.4 - 1.2% click rate on average.

Thank you both for your replies!

Another upvote for Code Sponsor! Been using it for almost a month or so, and it's pretty awesome.

Styfle already describes it very well, so I'm just gonna add that although I'm seeing a very shaky impression/click rate, and my clicks are below what I expected to receive, the payout is quite big for an ad/sponsor network like this. Also, it's much better than ad networks - no detailed tracking and more (listed below by cavneb)!

Unfortunately, it's limited to projects with 100+ stars. Cool idea though.

It didn't used to be limited, but I believe it's because of frauds and scams. Maybe if you contact cavneb he can work something out?

I totally love code sponsor ️

Thank you, Kent! Your OSS efforts deserve to be funded!

Reply on news.ycombinator.com