Testing Zulip
-
We installed this today because we are looking for a Rocket.chat replacement because Rocket is no longer going to do notifications for free, so easier to just move to something else. Zulip came up as an option, so we are investigating. Wondering if anyone else has tested it out?
We got it up and running on Ubuntu 20.04 without a problem. It's installation process is clean and simple. It's pretty nice.
But once you log in, my first impression is that the "instant messaging as email conversations" concept is way too close to Google Wave (which obviously was an idiotic idea) and isn't efficient, and more importantly, I can't figure out how end users are going to be able to figure this out. For an IT team, sure, generally smart enough and putting in the effort to figure out what is going on. But for regular end users? It feels like it might be insurmountable for them to try to use it.
Much like Teams. We see customers struggling to actually use it. We don't want to force something inappropriate down our user's throats. And my first impression is that Zulip would be just that. Great idea, nice code, good for probably small, very "on top of it" teams. But for general end users, it seems just too hard.
Going to leave the server up for a little while to keep playing. But we have a Mattermost up too, and that just seems way, way more intuitive.
Plus, 90% of conversations that we see are just people talking, and the "topic threading" concept seems inappropriate, like it wouldn't make any sense for the use cases that we use.
If we were using it for a support test trying to coordinate ticket responses, it might be awesome. Make a "topic" per ticket or something. Brilliant in that case.
-
I have not tried Zulip before but Mattermost has always been my top opensource recommendation when people are fed up with paid alternatives.
-
@jmoore said in Testing Zulip:
I have not tried Zulip before but Mattermost has always been my top opensource recommendation when people are fed up with paid alternatives.
We have a Mattermost server for testing, and it's always very solid. I like it a lot.
-
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
-
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
-
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Edit: Or go SaaS with them and pay $2 a month and have no limit for notifications at all (and not have to manage anything). That seems like a crazy deal.
Edit2: Sorry missed the hundreds. You have hundreds of employees?
-
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Edit: Or go SaaS with them and pay $2 a month and have no limit for notifications at all (and not have to manage anything). That seems like a crazy deal.
Edit2: Sorry missed the hundreds. You have hundreds of employees?
Hundreds of NTG employees? No -
But NTG client employees number in the hundreds - and they use it as a point of contact other than phone or mail
-
@gjacobse said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Edit: Or go SaaS with them and pay $2 a month and have no limit for notifications at all (and not have to manage anything). That seems like a crazy deal.
Edit2: Sorry missed the hundreds. You have hundreds of employees?
Hundreds of NTG employees? No -
But NTG client employees number in the hundreds - and they use it as a point of contact other than phone or mail
Then it's a no brainer. The $2 should be built into the support cost. The employee cost for it should be minimal.
-
For fun, let's say Scott as the CIO of NTG makes a base salary of $100,000. Pretty low, but an even number to work with. That's $80 per hour the company is paying to employ him. Let's also say it takes two days of work to set up a new solution. Actual setup, testing, etc. Over however long it's probably 16 hours of work, even if it's more than that in days. We'll be generous and say NTG has 40 employees.
At $2 a month, you're at $80 a month for their hosted SaaS solution. No notification limits, no managing servers, etc. For the $80 a month to catch up to the 16 hours of work testing and moving to another solution, it would take 16 months. And that's just the initial set up cost, not including server maintenance.
Customer cost doesn't come into this because they should be billed for the service. They should be currently billed for this because the maintenance of the server isn't zero, so they should be paying for some aspect of it.
It's a no brainer that the SaaS solution is cheaper. Plus you're also contributing to a project and company that you like.
-
@stacksofplates said in Testing Zulip:
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Hundreds of users.
-
@stacksofplates said in Testing Zulip:
Edit2: Sorry missed the hundreds. You have hundreds of employees?
No, users. Rocket isn't our internal messaging system. That's Zoho Cliq.
-
@stacksofplates said in Testing Zulip:
@gjacobse said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Edit: Or go SaaS with them and pay $2 a month and have no limit for notifications at all (and not have to manage anything). That seems like a crazy deal.
Edit2: Sorry missed the hundreds. You have hundreds of employees?
Hundreds of NTG employees? No -
But NTG client employees number in the hundreds - and they use it as a point of contact other than phone or mail
Then it's a no brainer. The $2 should be built into the support cost. The employee cost for it should be minimal.
That's not viable MSP math. You start doing that and suddenly every tool costs you tens of thousands of dollars and you spend so much supporting a person that no one can afford you.
There is a reason why $3/mo RMM tools don't make sense, it's too costly on an end point / person basis. Individual endpoints or people should be cheap. If the ENTIRE cost was $2 well heck, yeah. But once you start down this path you pay $2 here, $2 there and suddenly you're spending more per user on tooling than most companies charge for support.
The idea that you just bill the customer for you not keeping the cost down sounds great, but in reality, being the company that delivers more value for less money goes a long way.
-
@stacksofplates said in Testing Zulip:
Customer cost doesn't come into this because they should be billed for the service. They should be currently billed for this because the maintenance of the server isn't zero, so they should be paying for some aspect of it.
This is a customer only system. So ONLY they come into it, really.
-
@stacksofplates said in Testing Zulip:
For fun, let's say Scott as the CIO of NTG makes a base salary of $100,000. Pretty low, but an even number to work with. That's $80 per hour the company is paying to employ him. Let's also say it takes two days of work to set up a new solution. Actual setup, testing, etc. Over however long it's probably 16 hours of work, even if it's more than that in days. We'll be generous and say NTG has 40 employees.
Just to make it clear how impactful this is. We had a product in use by a portion of our customer base. It was only $3. Which yes, is 50% more than we are talking about here, but the same logic.
At that price, we were able to afford to hire a full time engineer just to build and support a solution to replace it and were able to provide a vastly superior solution to the customers, at lower cost. The entire engineering budget for it was covered in the elimination of the $3/mo/end point cost. All of it. With lots of profit overhead. Lots.
Now that's having a full time engineer for something complex. Imagine how little engineering time is needed for something like Rocket, Mattermost, etc. That $2/mo is a cripplying expense that we'd be insane to even entertain. Even at $.50/user we can do it way cheaper in house. More like $.005/user.
-
@scottalanmiller what's your time worth?
-
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@gjacobse said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
Rocket is no longer going to do notifications for free,
It shows you get 5 thousand for free per month?
Yeah, doesn't seem like nearly enough for us. We've hundreds of users and constant chatter. Seems like we could go through that in a few days.
How many users do you have? Seems like a good investment to pay the $3 a month to support Rocket.Chat.
Edit: Or go SaaS with them and pay $2 a month and have no limit for notifications at all (and not have to manage anything). That seems like a crazy deal.
Edit2: Sorry missed the hundreds. You have hundreds of employees?
Hundreds of NTG employees? No -
But NTG client employees number in the hundreds - and they use it as a point of contact other than phone or mail
Then it's a no brainer. The $2 should be built into the support cost. The employee cost for it should be minimal.
That's not viable MSP math. You start doing that and suddenly every tool costs you tens of thousands of dollars and you spend so much supporting a person that no one can afford you.
There is a reason why $3/mo RMM tools don't make sense, it's too costly on an end point / person basis. Individual endpoints or people should be cheap. If the ENTIRE cost was $2 well heck, yeah. But once you start down this path you pay $2 here, $2 there and suddenly you're spending more per user on tooling than most companies charge for support.
The idea that you just bill the customer for you not keeping the cost down sounds great, but in reality, being the company that delivers more value for less money goes a long way.
Wait so instead of a support portal you create a user for anyone from the customer who wants to talk to you guys?
-
@scottalanmiller said in Testing Zulip:
At that price, we were able to afford to hire a full time engineer just to build and support a solution to replace it and were able to provide a vastly superior solution to the customers, at lower cost
You hired an engineer for $600 a month?
I'm guessing because "hundreds" doesn't tell me anything. So I'm going with 200 people.
-
This post is deleted! -
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
At that price, we were able to afford to hire a full time engineer just to build and support a solution to replace it and were able to provide a vastly superior solution to the customers, at lower cost
You hired an engineer for $600 a month?
I'm guessing because "hundreds" doesn't tell me anything. So I'm going with 200 people.
While it’s been two or so years since I was an active NTG staffer, one client would have had 200 employees in just one state, and they where in all 50.
Rocket was a good process as it allowed for whole company separation and still include notifications like down time. But also allowed for the personalization of direct contact that wasn’t a flood to NTG
-
@gjacobse said in Testing Zulip:
@stacksofplates said in Testing Zulip:
@scottalanmiller said in Testing Zulip:
At that price, we were able to afford to hire a full time engineer just to build and support a solution to replace it and were able to provide a vastly superior solution to the customers, at lower cost
You hired an engineer for $600 a month?
I'm guessing because "hundreds" doesn't tell me anything. So I'm going with 200 people.
While it’s been two or so years since I was an active NTG staffer, one client would have had 200 employees in just one state, and they where in all 50.
Rocket was a good process as it allowed for whole company separation and still include notifications like down time. But also allowed for the personalization of direct contact that wasn’t a flood to NTG
Wait a minute. So NTG is offering Rocket.Chat as a service to customers for internal communication along with using it as a means of notifications for SLOs?
