Hostea: forge hosting supported by a collective

Running Gna! Business development
#1

Bonjour,

Providing hosting services for a fee (dedicated or shared) is a loss leader (i.e. not sustainable on its own) that helps sell services (which makes it a profitable combination) and that is what GitHub & GitLab are doing. Hostea was initially set to only provide hosting (dedicated & shared) and no services but it would have to host a very large number of customers to be sustainable (thousands of not tenths of thousands), it could take years and create centralization problems. Fundraising such a bootstrap by organizations that expect to make profit out of their monetary investment is unlikely (and has undesirable side effects) because the profitability is low, years away and unlikely to be attractive.

About a month ago discussions led to the idea that Hostea could provide services in addition to hosting, thus mimicking the GitHub & GitLab business model (without any proprietary software or user lock in whatsoever, it goes without saying but it is better said). The fact that both do that based on proprietary software does not significantly change the equation. In all cases (proprietary software or not) the hosting part of the operation is a loss leader and the service part is profitable. Free Software based service companies have been profitable and it is a well established business model since the early 2000. There is however one condition for selling services: potential customers have to know the software already, it has to be a recognized brand. Which Gitea is not. But GitLab is which led to the proposal that Hostea supports both GitLab and Gitea.

Today it occurred to me that Hostea could also be a loss leader for the benefit of the Free Software service companies supporting it. Here is the value proposition:

  • Hostea provides dedicated & shared forge hosting based on Gitea and GItLab
  • Free Software service companies providing services based on Gitea and GitLab support Hostea financially or in kind
  • Hostea notoriety helps Free Software service companies with their sales as they can claim to run Hostea

Here is an example:

Realistically Free Software service companies are unlikely to be convinced to contribute financially, at least not in the first few years, because Hostea is not an established brand and it won’t help much with selling services. But they may be persuaded to contribute in kind (resources or work) if they get funding to do so via governmental research grants in the short term.

Free Software service companies cannot obtain R&D grants for providing services because it does not qualify. But the highly experimental nature of federation and Hostea being in its bootstrapping phase makes it possible. As Hostea matures and becomes more widely known this will no longer be possible but that will also mean that Hostea notoriety can effectively help with selling services. The incentive shifts and Free Software service companies can keep supporting Hostea.

The Hostea timeline could look like this:

  • Year 0 to year N: Hostea is not known and Free Software companies contribute “in kind” because they can get R&D grants
  • Year N+: Hostea is known and Free Software companies contribute financially and “in kind” because it helps them sell services

In other words that business model bridges the gap years and allows Hostea to grow to a threshold where it becomes sustainable and able to hire staff even though it is a loss leader. In the first years most of the day to day work to run Hostea is contributed in kind by Free Software service companies (which is currently the case) and in the end it will become a mix of:

  • Hostea paid staff funded by
    • Hostea paying customers and,
    • financial donations from Free Software service companies
  • Resources and staff provided “in kind” to Hostea

Interestingly such a setup is well aligned with the current revenue sharing model. It also does not require centralization: any number of organizations duplicating the Hostea business model can exist at the same time. Very much in a post growth spirit, the growth phase is before the organization gains a notoriety. Once this threshold is reached, it should stop growing and thrive, aiming for excellence rather than size.

What do you think?

2 Likes
YC Co-Founder Matching Platform | Y Combinator
#2

You asked my feedback, and I find it appealing. Before I explain more, I should remark that for the enterpreneurial types you are dating as co-founder this may be a tougher sell. It is deviating from well-understood business models into more experimental territory.

My affinity to this model stems from the fact that it aligns fully with innercircles philosophy that I’ve been thinking a lot about. One core principle of innercircles is (holistic) Sustainabilty and an intrinsic value is Humanity. At the heart of innercircles is an activation mechanism based on a virtuous cycle of Mutualism and Reciprocity. One tenet or objective is to “De-emphasize the role of money”, and be always in search of the triple-win i.e. creating win-win situations and continually maximize the synergy in them in a methodical manner.

That is what is happening in this model. A new business has a ton of needs that can all be and usually are acquired via upfront monetary investment. By exchanging “in kind” you can take money out of the picture on certain aspects of the business.

In innercircles the Triple Win is a kind of fictitious balance sheet between partners, established on a “trust-first” basis, where monetary entries are just one of the many ‘currencies’ on the ledger. They are the most mundane. More valuable are non-monetary deeds and exchanges expressed in ‘currencies’ of Humanity. Also, importantly, this ledger need not be balanced and it is encouraged that parties pro-actively and magnanimously do more for their counter parties, e.g. by providing Gifts. By doing so deep bonds of Trust can be forged, and counter parties are stimulated to actively look how they can redeem the unbalance, and further the triple win’s synergy in the process. They’ll be reliable partners and in times of need you can count on their help.

You have sketched some triple win exchanges in the examples above. If you make thinking about these triple wins part of daily business, along the way, you’ll find many more ways to help Easter-Eggs other than displaying them in your About page. And you just DO them if there’s time and opportunity, there’s often no need to inform them or expect a direct exchange for your favor. It is part of the natural triple win bond between you.

This may all sound weird, but I feel a very strong new post-growth model underlies the innercircles philosophy. Here I’ll also have to mention that I have a weird approach to fostering and evolving the innercircles philosophy. For reasons it can’t have the same transparency requirements you adhere to in all your work. But innercircles has the same levels of transparency on its long-term outlook.

1 Like
#3

What you are proposing is risky, but I like the idea!

Many companies selling proprietary, paid software have an “open source version” that they give away for free to support the moment. The gratis offering could be our version of it.

There’s no telling how long “N” will be. While I don’t mind the uncertainty and certainly love the altruistic nature of this proposal, I believe we should continue trying to come up with ways to sell Hostea.

Gitea hosting, IMHO, seems too general. Gitea is fairly easy to install and doesn’t break very often, so our current offer is not attractive enough. There should be ways to make impress regular folks enough to buy Hostea services. There should be something that we can do, with our current product, that nobody else is doing. There should be a pain point which is uniquely solved by us.

I’m currently taking a break from Hostea and working on other projects, but I’m trying to think of ways to pivot.

2 Likes
#4

I’m intrigued because I see that as risk free. What do you have in mind?

#5

Agreed. I don’t think finding someone external to the Free Software world is possible to implement this. It is unique and requires a kind of creativity that involves intimate knowledge of the ecosystem. Very different from selling GitLab services. It is about convincing Free Software service companies to participate in the adventure. Even if they are funded to do it and like the idea in theory, that’s a challenge.

Agreed. I would like to stress that while the service companies contribute in kind they are funded to do so via R&D grants. This is essential to make it work for as long as necessary: it would only go so far if done without funding.

While I certainly share that philosophy, I also acknowledge that it is so rare at present that it would be hazardous to base a business model on that. it will hopefully change in the future but that’s not going to be tomorrow.

On the positive I think the proposed business model is compatible and encourages that kind of behavior. Which is better than other business model that simply break if people help each other out of the kindness of their heart :wink:

Right. The core of the Hostea activity would be to provide:

  • hosting to users (gratis & paying)
  • value to the members (via about or any other way) so that they, in turn, have a clear and measurable incentive to contribute to Hostea (in kind or financially)

and the value it brings to the members needs to be adjusted and reevaluated on a regular basis to make sure it is real. Otherwise members will no longer contribute. In the same way users will go elsewhere if the service does not provide the value they expect.

1 Like
#6

I find it risky because it changes the perception of what Hostea is. If customers come to believe that Hostea can be gotten for gratis, why would they pay? The free offer might be good enough to do forge-things.

And if the gratis offering becomes too expensive for the org and any revenue(cash + kind) that we generate is insufficient to keep it alive, then we might have to resort to making change to our offerings, right? Any changes that will negatively impact the gratis offering will not be well received by our customers.

#7

That’s not what is proposed here though.

By including a shared hosting offer, it would make sense to have a gratis offering similar to framagit, codeberg and others, constrained by resource usage. If a user wants more resources they could pay for them, either on the shared instance or by renting a dedicated instance.

1 Like
#8

Note that there’s “basing on” and merely being an “aspect of”, and the latter doesn’t carry the same hazard.

Even if it were
 there’s the Sourcehut approach that makes it very clear that for things ultimately a proper compensation is required. A form of managing of expectations. Here’s some recent HN comment by Drew Devault on that: > I think Sourcehut does this in an ideal way. On the site it very clearly says ... | Hacker News

1 Like
#9

Pitch for potential Hostea members (Autonomic, Wearchitects, IndieHosters)

English

  • Hostea is a software forge hosting service exclusively dedicated to Free Software.
  • Individuals and organizations can create an account for free within limits.
  • If they need more storage or continuous integration, they can buy additional resources.
  • This is how GitLab or GitHub started but 
 it is not sustainable, reason why they sell services.
  • The hosting helps sell services.
  • Part of the income derived from services is used to keep the hosting part running.
  • Organizations providing Free Software service based on GitLab or Gitea could also provide a public hosting service but they don’t because it is too expensive to do that individually
  • This is why Hostea is set as a collective of Free Software providers
  • Together they have the resources to run a software forge hosting service where users pay their rent monthly with their credit card
  • As the reputation of Hostea grows its reflects positively on its members and helps them sell services
  • In return the members contribute to Hostea
  • But there is a catch: building such a reputation takes years
  • Fortunately there is a solution to bridge this gap: organizations can rely on governmental research and development incentives and grants
  • That is what Easter-eggs, my employer, does: that’s what allows me to work for Hostea and all other Hostea members can do the same
  • Eventually, when Hostea’s brand is well known and actually helps sell services, Easter-eggs will also be able to share these profits with Hostea
  • When Hostea reaches this level of maturity, it will stop growing and stay small, it is not designed to scale up
  • It is expected instead that other collectives will be created using the same model
  • Together these organizations can scale globally and make centralized software forges obsolete

French

  • Hostea est un service d’hĂ©bergement de forges exclusivement dĂ©diĂ© au Logiciel Libre
  • Des personnes et des organisations peuvent s’y crĂ©er un compte gratuitement, dans certaines limites
  • Si elles ont besoin de plus d’espace ou d’intĂ©gration continue, elles peuvent acheter des ressources supplĂ©mentaires
  • C’est comme cela que GitLab ou GitHub ont dĂ©marrĂ© mais 
 ce n’est pas un modĂšle durable, raison pour laquelle ils vendent du service
  • L’hĂ©bergement aide Ă  vendre les services
  • Une partie du revenu de ces services est utilisĂ© pour financer l’hĂ©bergement
  • Les personnes ou sociĂ©tĂ©s fournissant des services Logiciel Libre Ă  base de GitLab ou Gitea pourraient aussi fournir un service d’hĂ©bergement accessible au public mais elles ne le font pas parce que c’est trop couteux de le faire seul
  • C’est pourquoi Hostea est un collectif de fournisseurs de services Logiciel Libre
  • Ensemble ils ont les ressources nĂ©cessaires pour opĂ©rer un service d’hĂ©bergement ou utilisatrices et utilisateurs payent au mois avec leur carte de crĂ©dit
  • Lorsque la rĂ©putation de Hostea grandit les membres en bĂ©nĂ©ficient et cela les aide Ă  vendre du service
  • En retour les membres contribuent Ă  Hostea
  • Mais il y a un soucis: construire une telle rĂ©putation prend des annĂ©es
  • Heureusement il y a une solution dans l’interim: les organisations peuvent s’appuyer sur le crĂ©dit impot recherche et des subventions
  • C’est ce qu’Easter-eggs, mon employeur, fait: c’est ce qui me permet de travailler pour Hostea et tous les autres membres d’Hostea peuvent faire la mĂȘme chose
  • Finalement, lorsque la marque Hostea est connue et aide vraiment Ă  vendre des services, Easter-eggs pourra aussi partager ses profits avec Hostea
  • Lorsque Hostea atteint ce niveau de maturitĂ©, elle arrĂȘtera de grandir et restera de taille modeste, elle n’est pas conçue pour grandir indĂ©finiment
  • Au lieu de ça l’espoir est de voir apparaĂźtre d’autres collectifs construits sur le mĂȘme modĂšle
  • L’ensemble de ces collectifs peut grandir sans limite et rendre les forges logicielles centralisĂ©es obsolĂštes
1 Like
#10

I understand that a shared hosting service would have to be flexible enough to allow participation from non-customers. Maybe I’m being stingy here, but any such mechanism that we deploy shouldn’t be too expensive.

I was going to mention the same but Sourcehut has it easy: their workflow is based on email. So anyone with an email address, Sourcehut customer or not, can send in patches and create tickets.

If we are doing shared hosting, then we have to make it possible for non-customers to participate. But IMO, their capabilities should be limited to participating in our customers’ projects(send patches, create tickets, etc).

That said, Sourcehut is also flexible. They have a process to offer their services gratis for that can’t pay for forges:

If you require financial aid to use sourcehut, please send an email explaining your circumstances and we’ll do our best to accommodate your needs.
src

Overall, I think Hostea can benefit from shared hosting, but we should be careful about our gratis offerings. It should be flexible to accommodate less-fortunate folks and enable customer projects to benefit from non-customer contributions, but it should also let people know that running a quality, sustainable customer-centred service costs money.

1 Like
#11

I understand your concern better now, thanks for articulating it. Would you say that imposing the same limitations as Codeberg on gratis hosting is sensible? Would you see that as an acceptable risk?

1 Like
#12

Codeberg has gratis limitations?

#13

Yes, there are limits, here are some from the top of my head:

  • no CI by default and you have to make your case to use it so that it is not overloaded. And if it is I assume you would be kindly asked to reduce the usage
  • limit of 100MB (or so) for private repositories
  • repository mirroring is disabled (for everyone I think because it can’t be enabled for some repositories and not others) because people used a massive amount of disk mirroring large quantities of repositories

Other shared forges such as framagit.org which have been in service for much longer ended up actively discouraging people from creating accounts because they had too many, to reduce the spam (account requests are not approved once a month in batch)

1 Like
#14

Laurent @ Easter-eggs was kind enough to listen to the pitch and provide feedback. No red flag :tada:

This was during a private conversation that was not recorded.

1 Like
#15

I wasn’t aware of that, thanks for sharing!

Codeberg limitations are sensible but complicated. I like the Sourcehut way where they manually review gratis requests. It shows that gratis isn’t sustainable but also that they care about their customers.

But I don’t have any experience providing gratis services at scale so I don’t know what it would cost us. I will trust your judgment on this. :slight_smile:

1 Like
#16

I need to redo the pitch based on detailed feedback. The key takeaways are:

  • Focus on Hostea being a showcase for its members
  • Clarify the members are Free Software companies happily making a profit providing services based on GitLab
  • Clarify that Hostea provides support (fix what is broken) and not services (spend time understanding what a given customer wants and being paid to work on it)
  • The Hostea offering is B2C and members don’t do B2C at all
#17

A post was split to a new topic: Gna! pitch for potential member organizations