Docker’s cash conundrum is becoming a bet on a very different future
Tiers of a clown – or the early days of a better service model?
Opinion Docker has a problem: too few whales. Its user demographic is hugely skewed to free plan subscribers, while heavier users (meaning corporate customers) are too thin on the ground to generate enough money through their paid-for tiers.
As Docker's financial whizzes can read a spreadsheet as well as anyone, the vendor has been adjusting who gets owt for nowt for some time. It's a really interesting problem: the company's growth is a death march if the increased operating costs from happy punters outrun the revenue growth, but as in classic freemium models you rely on the bottom tiers to feed a critical mass of adoption to create a viable ecosystem of experienced users and Docker-compliant workflow tooling.
The company's latest wheeze is to charge its bigger customers for a management tool, as well as more expensive tiers and some cogent roadmaps for intrinsic security and compliance management - which is grand, but not here yet. The new pricing strategy is where it's at.
Quite subtle stuff. It's also some way advanced from the "free for noncommercial use" simple approach. That can work well, especially when you factor in paid-for support, but not where you have substantial per-user operating costs. If the only things your free user asks of you are a download of a software image and the occasional update, then that's very different from day-in, day-out service usage.
Progressive pricing may seem dangerously like socialism to an industry which has managed to avoid admitting to itself that open source is the biggest successful non-authoritarian communist experiment on the planet
That makes Docker's problem worth watching. You must now pay for Docker Desktop if you're a company with more than $10m revenue or 250 employees, which one presumes was set based on a careful analysis of who's using it already and how important it is to them. There are two principles at work here: an organisation of that size isn't going to be impacted by paying for the software and, intriguingly, it's fair to ask them to do so.
It is this underlying concept of fairness which bears most examination. For if it is fair to ask an organisation on the plumper side of the SME spectrum to cough up so much per seat, is it also fair to ask yet more from those larger still?
This is the opposite way round to how things normally work in commerce. If you're big, you drive a harder bargain. You get unlimited enterprise licensing that lets your per-seat cost be a lot lower than for your poorer siblings. Being rich is very cheap, it's being poor that's expensive.
What Docker is edging into is the concept of progressive pricing. Like progressive taxation, it works on the principle that there are common goods that benefit everyone but which not everyone can afford to pay for equally. Progressive taxation is widely seen to be fair, except by the very rich, their lobbyists and their media companies, because while nobody needs a Ferrari, everyone needs transport – so have your Ferrari if you like, but pay for some railway track too.
How much is Docker a common good? That's another interesting question. Tools for building software are often free or low cost because the industry instinctively understands that this is how good people make good things. As we are learning, ecosystems and economies can be dangerously incompatible. Open source thrives because once a contribution is written, it stays useful with no further input. Companies that provide services need more than coders; they need cash.
Progressive pricing may seem dangerously like socialism to an industry which has managed to avoid admitting to itself that open source is the biggest successful non-authoritarian communist experiment on the planet. It may seem like something no big company, chronically averse to taxation, would tolerate – but these people already give Oracle and Microsoft money, so let's not overestimate their prissiness.
Most importantly, we need a good sustainable model that combines the robustness of open source in the market and the growing importance of service-based technology, which threatens to re-establish rentier gatekeeping as the primary model of access to IT.
If we don't want that to be Google and Facebook burning our privacy in the furnace, we have to find a way to restore fairness to the system, and fairness means a progressive approach. How it would work if widely adopted is another discussion: a generally agreed licensing model? A governance-oriented industry association? Everyone fends for themselves?
These aren't new ideas. Economics students will know the "tragedy of the commons" concept where unfettered access to a limited resource goes awry – and how it so often works out quite well when good rules evolve.
So watch Docker carefully, and don't rush to judgment on a company that has to live the experiment. If whaling is to work, it has to be sustainable - and it's in all our interests that it does. ®