But we're discussing the early pitch here. It's not enough to say We'll do the same as product X but BETTER. That's totally generic and not a plan.
So there has to be SOME insight at least on what might constitute "better," and then how to translate that into market dominance.
More seriously - network effects exist when there are reasons to choose your product other than what you, as the company running it, have personally built. People shop on EBay not because it's a great site (it isn't), but because if you want to find someone selling a rare item, you are far more likely to find it there than anywhere else. People stay on Facebook (despite oftentimes hating them) because their friends are on Facebook; it has nothing to do with what Facebook does, and everything to do with the behavior of their friends. People use Whatsapp/Line/WeChat/Hangouts/iMessage based on what their friends are using, irrespective of the relative technical merits. People stay at AirBnB not because it's a great site (though IMHO, it is), but because they're far more likely to find a nice place at an out-of-the-way location than any other hotel search.
The office messaging market - for whatever reason - has started to exhibit these dynamics. People setup Slack channels not because it's a good service, but because their employees encourage them to setup Slack channels. They write plugins for Slack not because they want to, but because that's where the audience is. They stay on Slack because that's where all the plugins are. Maybe these were intentional moves that Stewart Butterfield carefully planned out, or maybe they were just accidents of history. From an investor's POV, they don't care: the fact is that people do ask their employers to setup Slack channels, and they do get on Slack because other groups they're part of are using it, and that dynamic will lead to a monopoly that the investor wants to be a part of.
People don't use the alternatives to Slack because they're not as good as Slack, on various axes, is the short version. Slack has made themself hard to compete with.
Self-hosting, unlimited history, unlimited team size. They have great deployment docs for various methods and it was pretty easy for me to containerize it for docker.
It has a slack compatible webhook integration so many slack integrations will "just work" with mattermost, if the integration will allow you to specify your own url.