pages/blog/cloud-foss.md (view raw)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 |
--- template: slug: cloud-foss title: Cloud (F)OSS is a good model subtitle: On building (mostly) open source startups date: 2022-02-07 --- Of late, I've been thinking a lot about what makes a startup work, and I think open sourcing your product -- or a good portion of it -- is a great approach. To be clear, I'm only talking about SaaS platforms, and not any other kind of product. Hence, I'm dubbing this as "Cloud FOSS". The title of this post was initially "Cloud FOSS is the way", but I quickly realised that I know next to nothing about actually building companies and I'm really just talking out of my rear end. Nevertheless it's still pretty fun to try and reason about why the open source model is great for startups, so reason we shall. Broadly speaking, there are two kinds of "Cloud FOSS" companies: - the open source and cloud versions are identical, entirely free software (à la [sourcehut](https://sourcehut.org)) - the cloud version has exclusive "premium" features that aren't present in the open source version -- a.k.a the open core model; it seems to be gaining a lot of popularity in recent times with the likes of Gitlab etc. Let's dissect each approach and see what drives them. ## the all-FOSS absolutist model As mentioned above, sourcehut is a great -- if not the only -- example of a company built this way. Unless of course, I'm gravely mistaken (wouldn't be the first time!), and you know of another, or are building one yourself -- please [email me](mailto:x@icyphox.sh); I'll be happy to mention it here. **Update**: I was indeed mistaken. Here are a few companies built similarly: - [jmp.chat](https://jmp.chat/) -- free-as-in-freedom US/Canadian phone numbers - [Frappe](https://frappe.io/) -- Indian company building a suite of free software products - [Plausible Analytics](https://plausible.io) -- free software analytics platform. The self-hosted version has a less frequent (LTS) release schedule. For those unaware, sourcehut is an entirely free software company/startup that's building a software development platform; a suite of tools and services like git/hg hosting, CI, mailing lists, issue tracking etc. All of these can be self-hosted, with [plenty of docs](https://man.sr.ht/installation.md) to get started doing so. Or, you can of course, pay for the hosted service at [sr.ht](https://sr.ht), their flagship instance. Granted, this one's probably quite hard to pull off, especially if you're VC backed or have investors of any kind. When your product is free software, you can't really bake in analytics and other creepy user-tracking shit that's common these days; what pages perform better, what buttons do users click more often, the likes. And naturally, you don't really have metrics to show your investors that your latest feature du jour is doing great (or not). Thankfully for us, sourcehut is completely bootstrapped. They've [written about](https://sourcehut.org/blog/2022-01-09-how-does-our-business-work/) their business model, and it's beautifully simple: they make money from users subscribing to their service -- the hosted version of the sourcehut software -- and from other free software consultation gigs. That's it. And they're quite profitable. Very cool sourcehut. ## the less absolutist open core model The open core model, according to Wikipedia's definition: > primarily involves offering a "core" or feature-limited version of a > software product as free and open-source software, while offering > "commercial" versions or add-ons as proprietary software. Typically in SaaS offerings, the "commercial" version is a hosted version of the open source software, _plus_ the usual suspects like analytics, project management, user management, and a seat count for different subscription tiers. Some startups may simply continue to operate this way, with feature parity between the open source and cloud versions (minus the meta "features"); however, this may not always be the case. Oftentimes, in accordance with the definition above, the open source version may be _feature-limited_, with some parts of the core feature set only available on the cloud version. In most cases, this probably doesn't matter to the average self-hoster, since the paywalled features will most likely be irrelevant to them -- usually catering to larger deployments (support for say, CockroachDB), or enterprise-level deployments (Active Directory support, SSO), etc. The business model is quite straightforward -- get initial traction via the self-hosted open source software, while building out your cloud offering. The community built around the self-hosted software are your initial pool of potential paying customers. Generally, these are either early-stage startups (seed to series A) who don't have a large enough team to manage additional infrastructure, or large enterprises who are happy to pay for well, the enterprise features. These enterprise deals may even transition into large on-prem contracts. One thing to note about most open core projects is they'll most likely enforce a CLA or a Contributor License Agreement. This usually says something to the effect of "while you own the copyright to your code, you grant us the rights to make money off it, or even relicense it should we wish to do so". It's [probably not a good idea](https://drewdevault.com/2018/10/05/Dont-sign-a-CLA.html) to sign one. ## why cloud FOSS works Cloud FOSS as a basis for building a company probably works because of a bunch of reasons. As a startup, getting an initial foothold in the market can be hard, and offering a fully open source version of your platform can simplify that a whole lot. Setup a GitHub organization, do a Show HN and Bob's your uncle. Obviously, it isn't _that_ easy, but you get the idea. Having an initial userbase from open source can be very useful. You essentially get free insight into what features work and what don't, and what features your users want. You also get bug reports, issues from various deployment scenarios, environments, operating systems, etc. A goldmine of information to help drive product development decisions. Further, having a community built around your product helps too. High quality contributors (sometimes), an audience for PR events, the ability to conduct user surveys and most importantly, a hiring pool. With that concludes my ideas on why Cloud FOSS is a good way to build a startup. Again, I must reiterate that I literally have no idea what I'm talking about and whatever I posit is merely a result of "Oh I've seen it work this way before". I'm happy to hear what you think. <hr> This post was inspired by a conversation with [Prithu](https://prithu.dev). |