tl;dr: SaaS is the best open source business model.
N.b. Here I play devil’s advocate, steelmanning SaaS. I don’t do it well, the article needs attention.
SaaS enables clients to outsource IT. (Originally, they’d sell the backend to run on prem or run it for monthly fee, with everyone preferring the latter.)
Open source killed the mid-segment of fixed software sales (products like dBase, Turbo Pascal, Lightwave). Open source dev attention is generally unpaid (in spite of foundations and companies paying engineers to contribute.) The GPL always allowed resale and that won’t change. SaaS can be the middlemen-operators of yore. When you open source your product, you become a consultancy/host/maintenance company.
There are Extreme because this hypothetical client needs it. extreme usecases like resource intensive software with non-technical clients. Who configure everything there? (This is like selling a plane to someone without a license.)
Even with an turn-key on prem system, who will maintain the system? Ok, traditional IT consultancies which maintain software and equipment from Cisco, VMWare, Citrix etc. do exist. With planes there are pilots for hire. You could also sell lifetime licenses with support agreements, but non-technical customers won’t understand the delimination of responsibilities.
From the business side:
- investors like subscriptions because it’s consistent, recurring, provable revenue, leading to much higher valuations. One off sales makes this harder. (if interest rates are 0.5%, an annual recurring payment of $100 is worth $20,000; but if interest rates are 5%, like in the 90’s, it is only worth $2,000. With the fed increasing rates, maybe we’ll see a return to pay-once software?)
- monthly costs can fall into personal discretion but large one off costs require more company supervision to approve
Big providers (Stripe, Gmail etc.) can often burn/drop customers with no worry of repercussion. Countless stories of “Big X killed my company” appear. There is of course no customer service, when the shiny automatic service fails. This is the risk of doing business (With them as a vendor).
Easing, say commodifying competition in these spaces is one solution. If Gmail or Stripe had a dozen API compatible competitors, bad PR (from killing a business…) would launch a customer migration. Customer support would be a viable differentator. An world where anti-trust acted would see many SaaS vendors competing in these segments, as paid alternatives. Whether the costs etc. are healthier to the ecosystem as a whole (enabling more SaaS companies to thrive, while costing others more), is uncertain. Some bigtech antimarket effects do subsidize other activity. Certainly, the world would be more vibrant.
Contracts (SLAs) force SaaS providers to be direct, accessible and reactive.
Updated April 11, 2024: Opentofu poses a partial counterexample, where Hashicorp’s SaaS offering is scalped by a free-beer alternative emulating it on a 1:1 basis (in response to the move to business source.) If this keeps happening, we’ll have a crab bucket situation, dissuading new companies from open source, entirely. (I have no idea what that would do to today’s SaaS exosystem though.) Opentofu issues show a request before HashiCorp’s implementation appeared, making this particular case less clear.
Updated April 27, 2024: Looking into it deeper, we already have a clear picture. Besides Hashicorp’s move to business source (and IBM buyout), Redis, MongoDB, Cockroach, Sentry etc. moved to source available… When this happens, open forks appear.
But who should operate and generate revenue from these open forks?