A Critical Response to Mattermost’s Recent Changes

1

“If advice were worth something, it would be free.”

The recent shift in Mattermost’s strategy - especially with the introduction of the Entry Edition - feels like a betrayal of the open source values that built its community. By stripping away essential features like GitLab authentication, LDAP, SAML, and now imposing a 10,000-message limit per channel, Mattermost is no longer empowering teams - it’s cornering them.

  • Feature Removal: A Step Backward
    GitLab OAuth2 authentication was a natural fit for developer teams. Removing it from the free edition is not just inconvenient - it’s exclusionary.
  • LDAP and SAML support, critical for secure enterprise environments, are now locked behind paywalls.
  • The 10K message cap is artificial and unnecessary. It doesn’t reflect a technical limitation - it’s a deliberate barrier designed to frustrate users into upgrading.

These aren’t enhancements. They’re regressions.

Forced Migration: A Risky Strategy

Mattermost’s approach seems clear: restrict the free version so severely that users feel forced to pay. But this tactic risks alienating the very community that helped Mattermost grow. Open source thrives on trust, transparency, and shared ownership - not coercion.

Instead of building loyalty, this strategy may push users toward truly open alternatives. It’s a short-term gain that could lead to long-term loss.

A Better Path Forward

If Mattermost wants to monetize sustainably, there are smarter, community-friendly options:

  • Offer paid services: hosting, installation, consulting, and enterprise support.
  • Keep the core platform fully open, with all essential features intact.
  • Build a plugin marketplace for premium integrations.
  • Focus on value-added services, not artificial limitations.

Open source isn’t a liability - it’s a launchpad. A robust free edition strengthens the ecosystem, attracts contributors, and builds goodwill. Monetization should come from services, not from stripping away functionality.

Final Thoughts

Mattermost’s recent decisions risk turning a beloved open source platform into just another gated enterprise tool. If the goal is to grow, the strategy should be to empower, not restrict. The community deserves better - and Mattermost has the opportunity to do better.

7 Likes
2

I wouldn’t mind pushing everyone I know on a free/opensource plan to signup for a paid license, I have in the past, but every time I’ve got pushback saying the pricing doesn’t align with their organization or it’s simply too expensive.

A couple of years ago Mattermost had a few different plans like E0/E1.. (If I remember correctly..) which had a entry level price at $3.xx. Looking back in hindsight that plan/pricing would’ve been perfect today, but we were in different times back then (afaik this was pre-COVID/2020) where remote work/team-chat wasn’t really a thing.

Introducing back a plan like that ($3.99/user/month~) upto XYZ number of users per license, I feel would be a good way forward for Mattermost to encourage smaller companies to signup for a paid product to begin with.

Right now the $10/user/month paid annually, is a decent/steep price for many parts of the world and even in the west for smaller organizations it is a steep entry point for sure.

I understand support etc costs money too, but there must be a way out with community supported plans or “basic support” plans or partner companies who can offer paid/purchased support for these low-cost entry plans which would result in a larger benefit for mattermost.

tl;dr: The products great, getting people to pay the steep entry price is a issue.

3 Likes
3

I have no problems paying for products I use when the price is reasonable, but paying 400$ a month for something I’m self-hosting is extortionate, let alone if I had even more users.

The resources used are my own.

I wouldn’t even say the recent changes are that great. To wit, when I was still on entry before migrating back to Teams, I had the mattermost agents AI plugin causing the client to crash literally every second message. This certainly does not endear me to want the enterprise edition. There’s also been breaking changes that have caused some plugins I previously used to cease functioning.

To put it in the (in)famous words of Gabe Newell of Valve Software: “Piracy is a service problem.” And the lack of uptake to enterprise edition is a service problem too, in my view. Give small userbases a way to pay you some money and many of them happily would, I know I would have in the previous version days - I’m more reticent now, with how Mattermost is treating me, and treating us.

Give us ways to license self-hosted instances that aren’t charging us for our own resources and are reasonably priced and I think you’ll see some uptake, I think.

4 Likes
4

I agree, Mattermost needs hobbyist/small business pricing. I too like to pay for what I use (hell, I licensed winrar) but my MM instance is literally used for old friends staying in touch by posting memes and family pics. I’m not going to pay $1k+/yr for that.

I work in tech and understand the need to generate revenue, but these recent changes are absolute BS. There’s got to be a middle ground.

2 Likes
5

I think what I’d suggest is a DIY Self-hosting plan where you don’t get support (or maybe you can have a higher tier for support), where you just pay a flat X/month for the software and its licensed and we’re all happy. An X/user model makes sense when its in the cloud and its their resources those users expend, but it makes none when its my cpu cycles, my disk space, my recovery plan. And I’d go so forth as to say that you probably dont want to offer paid support to most self-hosting plans if you can avoid it because they’re all going to be 19032943209320932 different environments and you’re going to need significantly higher grade support techs to handle that.

2 Likes
6

Somehow (but not without problems) I can try to understand SSO removal from free version. However what happened with Omnibus (and/or system packages) is really confusing. Basically we’ve got two messages almost in the same time:

  • You will not be able to (apt) update your current installation
  • You have to patch a critical bug

That’s pretty much unfair…

7

It is interesting that Omnibus was supposed to be the future. Last year I put the time into transitioning from the manual install version to that. I also needed to migrate from MySQL to Postgres so it was a significant undertaking.

A couple months ago, discovering that Omnibus is now deprecated, I moved to the apt version.

Now the apt version results in Entry Edition (which is absurdly limited) so I just moved back to the manual Team version :sweat_smile:

I hope they settle on a course of action in the future that isn’t just ‘buy Enterprise.’

1 Like
8

It is uncharitable but I am inclined to believe that this is intentional - they want to make it so inconvenient to use free versions that you shell out.

1 Like
9

Thank you, were a small company in Uganda and just started installing mattersmost, got it working etc. thought I’d look through the forums, and what I’m seeing is disturbing to say the least, we’re looking to control our costs not get destroyed by them, I don’t mind paying but for a selfhosted version 10$ per user do behave. It’s cheaper over the year to develop our own.

10

Hi,

adding my 2c

Disclaimer:

I’m not working for Mattermost, but provide the German translations as a contribution to the product.

As far as I read concerning the limitations of the Entry edition, the 10.000 message restriction is CHANNEL based and not all messages, as stated here (We might need clarification on that):

Editions and Offerings - Mattermost documentation

LDAP / SAML was never supported without a payed license (there were some workarounds mimicking the GitLab OAuth API, true for this, I used this myself).
You’re referring to this feature as “critical for secure enterprise environments”, and I’m totally with you with one exception: Enterprises should pay for enterprise features.

Nobody is forced to use the licensed version, regardless if it’s Entry, Professional, Enterprise or Enterprise Advanced. The free Team version might suite most users better: It’s open as Open Source should be. Lacking features might be developed by some enthusiasts and, if not, one has to live with it or find a better fitting product. The user limit of 250 is a suggestion & AFAIK not a hard limit.

If we really want Mattermost to become a better Open Source tool, it’s on us. Either we use what we get & IMHO for small teams up to 50 users Entry is really a great option with all features available. Beside Mattermost, I’ve to use Teams & boy, how do I hate it: Slow, not reliable, confusing, integrations are a major pain… you name it.

We licensed Mattermost for some internal small team with great integrations to ERP & logistics in place. Also a very well working Atlassian integration. AI agents using Ollama API are running smooth & fast. Anyway, with the new Entry version we will switch back to a non licensed version as the price tag for us is now too high for Enterprise features (17 USD / month / user) (Advance would be 38 USD, but this is even more overkill for us).

Mattermost clearly focuses on major enterprise or government installations at a large scale with some unique selling points for these environments in terns of secure collaboration. If this is the way for them, OK: We can, but we do not have to follow.

11

I would accept this argument if they hadn’t been removing features from Teams edition to incentivise people to go onto Enterprise edition, but they have.

At this point, Matrix or just going back to the stone ages of IRC both have more to offer than Mattermost Teams.