Give nothing, expect nothing: GitLab’s the latest punching bag for entitled users
What do Docker, GitLab, and Red Hat have in common? Aside from various levels of participation in open source, they’ve all been punching bags over the past few years for non-paying users angry that they’ve taken some freebies off the table.
When Docker had the temerity to introduce limits for free users pulling containers from DockerHub, or requiring a subscription for large business users, lots of people started complaining and/or looking for a free alternative.
When CentOS moved from a direct clone of RHEL 8 to the CentOS Stream model, hordes of people raised virtual torches and pitchforks because they were no longer able to run a freebie RHEL clone. Many noted that this was bad for their business because of the uncertainty and having to either pay up or move to another RHEL clone. Few seemed to note that it’s bad for Red Hat’s business when people don’t pay for its flagship product, which in turn continues to fund development of RHEL. (There was also a pretty hefty side dish of IBM conspiracy to go with this, but that’s a topic for another day.)
Now, GitLab is making changes to its free tier and may zap inactive repositories or repos that are too large.
The common thread here: Lots of users expect to get things for free, forever, from for-profit companies that don’t answer to them. Those users contribute almost nothing1 to the bottom line for the for-profit companies, and actively drive up costs for them. Yet, somehow, with no skin in the game, they feel entitled to complain and badmouth the companies because they’re not getting as much value for their monthly contribution of nothing at all.
Expect to pay, now or later
I’ve restricted my examples to companies in the open source arena, but there’s plenty of other examples out there of companies that have provided “freemium” models and eventually start taking away features from the free tier.
It’s utterly predictable, and rational, for companies to do this. But, somehow, it’s usually a big kerfuffle when companies do this and users get all angrified because the tap on the freebies has been not even turned off but dialed back a bit.
Sometimes users have a point, when the companies haven’t offered a chance to pay in the first place. I’m looking squarely at Google here, the world champion of killing off services that it offered only for free, then decided weren’t strategic or profitable. Maybe Google Reader would never have made a profit if Google offered a subscription, but we’ll never know.
As they say, if you’re not paying for something, you’re the product. If you didn’t pay for RHEL, you have no stake to complain about changes to CentOS. Especially if CentOS runs part of the business you work for. Tell those cheapskates to pony up for RHEL subscriptions or get used to the bumpy ride — smoothing out the uncertainty is literally part of the value of open source products.
If you pay GitLab or Docker and they change the terms of service for the worse, you might2 legit have a grievance and they ought to listen to you. But paying for services is like voting: if you don’t do it, don’t complain. You got more than what you paid for, so expect nothing.
If you pay nothing for a service then expect disruption later on when features or the service go away.
Build it instead
The other complaints I hear about the businesses that do business around open source is around “lock in” and proprietary bits.
My answer to that? If you want all-singing, all-dancing, 100% open source, federated services that provide full export capabilities and all the goodies — build them yourself.
Not literally yourself, necessarily, but if we have groups of users who really want a GitLab/GitHub/SourceForge type service that is 100% unchained from lock-in, proprietary bits, and so forth — then those users should do the work, band together, and create an organization to do just that.
Create a business case, get the funding, stand up the infrastructure, and pay people to work on it rather than expecting for-profit companies to prioritize (what you see as) the public good over profit. Whether that’s how things should be or not, it is how they are and that isn’t going to change as long as the only movement in the direction of change is people hectoring for-profit companies to do better. Their “better” is more revenue and profit than last quarter. If they can incidentally do some good, great.
Companies are run by people who often do want to make the world better, but if that costs money, time or reduces profit, they have to justify it. And, usually, somewhere in the organization is someone tasked with looking at that equation and if it doesn’t balance out right, they’re empowered to say “no.”
If you build it, maybe they’ll come. Aside from all the risks inherent in starting a new venture that has to make money and pay people, there’s the dirty little secret that lots of people who say they want to pay for such things… won’t, in the end. If the ethical, pure open, business is a little harder to use or costs a bit more… many people won’t put their money where their mouth is. I’m not claiming innocence on this point, myself. I own an iPhone, not a Purism phone, for example.
But I’d love to work for such a company doing things for freedom, justice, and the open source way. You know, assuming they were paying market rates or pretty close to them.
Because, bottom line, most of us are trying to make a living. If you worked for Acme Corp that had to choose whether they’d continue giving away X, Y, and Z services or paying you a bonus or giving a decent raise next year, which one are you going to vote for? Handing out freebies or being paid for your work?
[1] I say “almost nothing” because companies see value in dominating an ecosystem. To the extent that the value of increasing the reach of an ecosystem outweighs the costs of giving things away, they’ll likely continue to do it. When the scale tips the other way and the cost of giving away the freebies outweighs the benefits, they’ll stop. Simple as that.
The freemium tier is not much different than writers and artists doing work “for exposure.” As they say, “people die of exposure.”
[2] I say “might” because costs aren’t static. If, say, storage costs go up then a company can only eat those costs for so long before they might have to charge more or reduce the storage tiers.
Clearly, you make a point here.
users pull images from DockerHub not containers
Not particularly relevant to the post, but thanks for the correction.
I think it is relevant. If you want to write about a technical topic like this no one will take it seriously if it includes such basic errors.
Thank you for your thoughts.
CentOS was a community project first (hence the name), it was then bought by Redhat, and then killed. This is much like Microsoft’s “Embrace, Extend, Extinguish” strategy and its (former) users are right to be upset by Redhat’s actions.
Further, I don’t understand the point you’re trying to make. Those companies are more than happy to provide a service to users for free in order to try to upsell their products, gain marketshare, and brand recognition. Then, when sales are starting to flatten, there’s some marketing genius who wants to squeeze more out of their existing users. In order to do this, they change their offerings, piss off their user base, and hope those users keep quiet and give them their money. Why wouldn’t they be allowed to complain? They served their purpose of being a number in a numbers game, and are now chewed out.
Lastly, it’s kind of weird that you use the CentOS/Redhat example plentiful in your article but fail to mention Rocky Linux.
The point is in the title: If you contribute nothing, expect nothing. You’re right – companies provide things for free to gain market share and stop when it’s not profitable. That’s rational. Unpleasant, but rational. What’s irrational is expecting it to play out any differently – especially after years and years of it happening.
Users should expect that from day one and when it happens, don’t be surprised. You can complain if you like, but it’s not going to change much.
Not sure why I’d mention Rocky here – there are several RHEL clones that took off after Stream was announced, but why would I need to mention them or Rocky specifically?
Google tried to get their free tier Google Apps users to start paying for the service. The service had been free for more than a decade and you now had to start paying more than $5 per month per user. There were a lot of complaints and now they’ve reversed their decision. The same goes for Gitlab. A lot of users complained and the company changed course. So you can say that complaining doesn’t change much but in reality, it does.
I expected you to mention Rocky Linux, or one of the other Redhat derivatives, because you make the case for that. Redhat killed CentOS and people did what you suggested, albeit in a not for profit way. Another clear example is Opnsense which was started in response to PfSense’s decision making.
It is wrong to first promise the world for free, give it, and then just take it away. Developers spend years developing services on top of it. Their entire effort is wasted when the underlying platform changes policies in such a drastic manner.
Gitlab is running itself into the ground. It just cannot compete with Github. Especially ever since it went public, it’s been doing so poorly. Executives continue to take their exorbitant paychecks while burning everything. They pay more to non-engineers than to engineers, which is a suicidal way to run a firm.
Gitlab execs respect money, but not competence. They do not know how to run a storage business. It is silly to rely on premium storage at a commercial public cloud for such a large venture. Hard drives are pretty cheap, and bandwidth too is cheap if you know where to look. If however you pay the likes of AWS S3 for everything, then even a single petabyte of storage would hurt you.
You have a point, of course. When I signed up for Gitlab I didn’t ask what Gitlab gets out of it. I thought that was their concern. I can wonder now. Market share? Mind share? Critical mass? It seems Github has been more successful in all of these realms. Now Gitlab are rethinking their strategy and after making a bargain with the (freeloading) public, they’re revising the deal. I’m not in the vocal minority complaining about this. I just chuckled to myself because I had just started putting my hobby projects on Gitlab because I wasn’t thrilled about Github mining my code for copilot. Jokes on me!
I also have a private Gitea instance running in a Docker container on my file server where I put things I don’t care to share. I suppose I could put all of my stuff there and not worry about the rules changing. I just have the vain notion that someone might find my code useful at some point.
NB: I’m not concerned about the container vs. instance distinction. This is not really a technical article and your meaning is clear.
I understand the economic perspective detailed here, especially when centering commercial OSS. A for-profit product most often needs a for-profit business model, and there is a lot of commercial OSS which elects to solve this for-profit challenge later instead of sooner. This could be considered as proving the old hypothesis of whether “build it and they will come” ends up being true for an open source product or not. If it is true, then the product can always change the rules once it has influence. If it is false, maybe development slows or ceases altogether.
However, I feel the long-tail impact of trust and accountability isn’t given its due weight in this perspective. In 2018-2019, Mozilla and Open Tech Strategies did interesting work in defining eleven different open source archetypes, or models, based on several real examples of open source products and communities. There are several different (profitable) approaches to open source, and not all open source products follow the same path. In some cases, trust is expendable once the company knows who is willing to pay for the product. Revisions to services and offerings for free users could just be “trimming the fat” and letting the company sustain the future of their product. However, I would posit that it can also dull the innovative edge of an open source project. The overall demographic of the user community may change, and there may be shifting attitudes and views towards the company and what they will or will not consider. It might be free users today, but it could be a different stratification of paid users tomorrow, who end up being viewed as the new “fat” to trim off. It can also invite disruption by a competitor (Docker and Red Hat Podman / OCI? 🙂 ) who challenges the product not in features but in reputation and perceived accountability.
Eventually, this discussion bends to a philosophical one instead of economic once you interrogate the assumption of what sustainable profit is. Should profit be unlimited and without bound? The open source observer in me says there will never be consensus and I’m doubtful a singular consensus would be a good thing for open source anyways. I suspect you would also agree with me on this point.
My view here is to not discount the innovative and at times valuable role that a grown community (of both free and paid peoples and groups) can play in the product lifecycle. Once a company establishes influence and mindshare with an open source product, the company also develops rapport and accountability by their behaviors and actions. The user base will form expectations and beliefs about the sustainable practices (economic) and values (philosophical) of the company or vendor. I think among veteran open source practitioners, we often become jaded by the Free Software pragmatism that often stems from the social values embedded from the history of the movement. While pure pragmatism is surely frustrating at times, I think it does come from a sound foundation, and one that is not incompatible with economic sustainability. While some voices (perhaps quite vocal ones at times) will protest and bemoan the changing of rules, there can be a deeper harm that comes from violating that trust and accountability of the community. And I would hedge these are voices that may be quiet or perhaps even silent at times. And unless a company or its community is very good at listening, these can come back as a haunt in the form of ill will and mistrust. Most companies and open source practitioners I know have a difficult time in coming up with metrics or models that account for these points.
There is no one-size-fits-all approach. And even the users who contribute $0.00 may not always be worthless participants and observers. My 2¢.
Free markets always find a way to equalize value and compensation. How people choose to spend their time and skill is their own choice. If they wish to work for free for a while, that is up to them. If later they change their mind so be it, it is entirely up to them. Value transfer is always an exchange between two consenting parties.
my 2cents