There's a reason nobody really does this: there's just no money in it. Nylas mail is a good example of this: it offered the ability to self-run it and pay them for the license, but nobody really did. I wonder what they would need to charge to make it feasible? $1000 a year? It makes me pretty sad, but it's just such a distraction for a company of Notion's size and stage.
On-prem enterprise software does make money. You still charge money for it either way. A few examples: Github Enterpise, JIRA, etc.
And notion doesn't need to see their users data to charge money for a service. They could engineer it to be E2E encrypted for the group and not have the ability to see their customers data even with a subpoena.
Enterprise is the key word here. Much more companies care about having control over their infrastructure. As for Notion, I guess even on Hacker News only a small share of users would care enough.
It's possibly something many of us should care more about, though.
Suppose your company handles a fair bit of confidential information, and some of that information ends up in Notion as part of your general project management workflow. Now you've got to worry about Notion experiencing a data breach, and whether or not that breach could get you in hot water with your clients, or even into legal trouble.
With on-prem options, you can gain a little extra confidence, because it can all live behind a firewall instead of being directly exposed to the Internet at large.
This started with the idea of a personal on-prem solution, which is probably not worth the cost to maintain for the company. I mean, how much is it worth to you to have an on-prem personal solution for a diary? Because that's the worry... people are afraid that their household budget or their database that tracks the content of various adult films or whatever is going to get exposed. That's mostly about embarrassment, not actual harm. That's not something worth hundreds of dollars a month to most people.
And really, if I want to write down my deep dark thoughts where no one will read them, I'll use paper.
Personal users want E2E encyrption, so it doesn't matter who is running the server. And most users do not have the skill to run their own server. The technically advanced can just use the enterprise version if they really care about personal on prem.
Maybe we should change the notion that it's difficult to deploy an on prem service.
Why can't it be as simple as containerizing it and presenting it exactly like your would install an app on your phone, with the same possibility to grant permissions? App store provides automatic OTA updates so you don't have to bother about security. And the app itself should not have internet connectivity, connectivity is provided through authenticated tunnel into the container so the app cannot phone home in the background, you can only connect into it using the tunnel with your own password.
Even if you have the skill, is it worth the time to do on-prem, considering, you know, a pen and paper?
(Full disclosure: I'm a total bullet journaling convert. Given a choice between giving up my bullet journal and giving up Google Calendar/Docs and Trello, I'd give up the online stuff in a heartbeat.)
> It's possibly something many of us should care more about, though.
Definitely, but I'd argue that even with containers etc., we don't yet have the capabilities to make what we know is right, as easy to operate as SaaS.
> Suppose your company handles a fair bit of confidential information, and some of that information ends up in Notion as part of your general project management workflow.
Many folks that post to HN seem to miss this even though they are probably very technically capable. Many IT people in small/medium organizations don't have the time to consider this. Sometimes relationships with big organizations force this reality and threats get addressed, but it's not the norm. Most IT managers are under pressure to cut costs, reduce staff, modify accounting (review vs capital budgets), coexist with shadow IT, etc. Cloud solutions seem fantastic to them, and they don't have the energy to fight with someone in marketing who introduces a rogue solution because they don't understand how the current Wiki works.
I think that's pretty common, so I don't disagree entirely. It can also be the case that you have increasing numbers of tech-savvy folks who feel that they can build their own tech solutions, but aren't actually equipped to build/select something that is equipped to serve the needs of their team or the organization.
This can result in the team being left holding the baby when certain individuals leave, data loss, security issues etc. In 15+ years of enterprise probably two-thirds of the time I've seen this issues and it has been predictable that it would work out this way, but IT didn't help themselves by pretending their solution was better. When IT tries to learn and adopts the tool, or incorporates the needs then things work out better. But, that's only a third of the time.
Well, then we should create open-source analog of notion. I actually was itching wondering what could I create in OSS realm. Probably will have a look at this task :)
That seems to be pretty hard. I just finished a long search for an Open Source one-page markdown editor (no split windows for preview and editing): the closest I could find was an editor called MarkText but even that doesn’t have half the functionality of Notion or Dropbox Paper. It’s hard to compete with fully funded and devoted teams.
Bitwarden offers self-hosting as one of their paid perks (Docker image and documentation included), in addition to a free tier. It may not be viable as the only business model, but I appreciate that they cater to the small number of people that will make use of it and I imagine the token of goodwill does something for the product as a whole.
Bitwarden is a niche application which handles very sensitive data. Value is created when it is hosted internally by a security sensitive organization.
When it comes to business and productivity applications the situation is very different. Many IT admins are tasked with moving to cloud (or other outsourced) solutions. When you see posts you have to realize you, or I, who value autonomy and control may not be the expected customer of these solutions.
To reduce the economics further:
* Startups like this have a product which may not have a complete market fit or reknown. They can't build something that's most likely to stick e.g. a required mail system because a couple of cloud products own the market. Therefore, they move up the productivity stack to team and document collaboration.
* Since they aren't a mature product and will likely change it considerably in the short-term, possibly "firing" customers as they search for a bigger market, they'll need their product to keep up. If they host it, then they can change it end-to-end.
* Any time you release a product that gets deployed on-prem it's virtually impossible to get people to update in a timely manner. Even if you decide to provide an on-prem version to a potentially huge customer, the enterprise sales cycle means that there is a risk your product will have changed before it gets deployed, but some person/process will prevent you get updated to the latest version as they advocate for a competitor.
* Worse, they like the old version and want to stay on it. How do you support these on-prem customers? They liked the value at your cloud pricing, but the cost to support on-prem customers can be much larger.
