True, but there’s also something vital that does not show on the balance sheet (and can sink companies):
Excellent tools spread through the best kind of word of mouth. Excellent tools that are free spread even faster and further. There’s no way to know how many corporate plans were sold because a single team member relied on Slack outside of work, or how many pro user accounts exist because someone was visiting a Slack-fluent friend.
And because there’s no way to know; the cost-cutters look at the numbers and say “cut the free plan enough to stop the hemorrhage”. Worst part is the next quarter looks great. And maybe the one after that. But soon enough the income starts dropping and there’s no “clear reason why” and then it’s too late to recover it. It has to be built again by building trust from scratch.
It's very similar to Salesforce's Heroku move. They either just don't care about that sales funnel and they remove it completely or they will keep squeezing until they kill it.
I would say they are acutely aware of the sales funnel. They weren’t making any money from those users who were ‘happily using Slack’s free tier for years’ and neither was there any prospect of converting them into paid users in the coming years. Very little point in keeping users (and paying to maintain them) who are going to be perpetually at the top of the funnel.
I think it's complicated; the fact that I use a free slack instance for note taking / bots is one thing, but I'm on 2 company slacks right now (paid), 2 non-profits (1 paid 1 not) and a handful of open source communities. Am I a user who won't convert?
Everyone on a free slack instance is NOT using a competitor's product.
There's a tipping point where having to check multiple apps will push people towards consolidating in a single client if they possibly can, which (in the long term) may not be slack if it doesn't support "ad hoc" communities or personal use-cases at a reasonable price.
> I think it's complicated; the fact that I use a free slack instance for note taking / bots is one thing, but I'm on 2 company slacks right now (paid), 2 non-profits (1 paid 1 not) and a handful of open source communities. Am I a user who won't convert?
Everyone on a free slack instance is NOT using a competitor's product.
I’d suggest the answer to that is probably no - an open source community is by its very nature going to want costs and overheads down, and all but the largest non profits and NGOs are going to try to get by by stitching together free tools and services. The one that is paying in your example is the one in Slack’s target who is going to be most likely to pay. Does Slack care about a few open source communities potentially churning? Probably not. Does it care about lowering average Time to Revenue? Absolutely.
This. Companies should be charged, sure; but an upside of Slack is that I can use various tech community’s Slack, the union’s Slack, and the coworking’s Slack in the same app.
None of them can pay, and they should be sponsored by our usage of Slack for our main job.
They can all better use Discord, whose model works on a per-user (not per-user-per-community) billing model. So less good at separating money from corporations but better aligned with non-commercial organisations as you don't need to pay $2k/month as none-profit just to keep chat history.
> Everyone on a free slack instance is NOT using a competitor's product.
Arguably this is even worse from slaks POV, no? Everyone on free slack is on free slack because they aren't willing/able to pay and slack would indirectly benefit if those customers were freeloading (and incuring costs) for their competitors.
Not the OP but I think if free slack becomse unattractive enough people might start to look around. There are enough alternatives (self-hosted like Mattermost, Rocket.Chat or free tiers from other vendors e.g. Zulip, Discord). That presumably has two effects:
1. People are used to them and if the decision at the workplace comes up to choose a solution the other product is on the table. Or worse if Slack messes something up with the paid tier people might be more willing to switch to something else that they know already.
2. Every private group chat might involve people who never used something like it before. For these people the software they use first is the baseline. And they are more likly to recommend the solution they already know to a organistation that is willing/able to pay for a higher tier.
We can't assume someone who won't pay for Slack won't pay for the thing they move to. How many people do you think pay for Discord's paid features that never would have paid for the forums Discord killed? The almost $15 billion dollar valuation suggest a lot.
It is not that simple too. Tools used for reasons outside of work may not always be perceived as good tools for work. And tools used professionally are not always considered to be good for use for personal reasons. If someone have met Slack for the first time at their job, they may recommend it at their next job, because they associate it with work - this conversion channel is pretty clear and established. But the conversions work<->personal are not so obvious and may perform worse. I'm pretty sure Slack marketing team controls awareness part of the journey and they understand well their conversion funnel. Ditching some non-converting users is a bold move, but it may be worth it and it may fit well in whatever sales strategy Salesforce has in mind.
Well, if I wasn't clear enough: the fact that someone have used a tool for personal reasons and may recommend it as a good tool for work is not enough to justify spending on those personal use cases. You need this to be a healthy conversion funnel and valuable brand awareness channel, otherwise you are wasting money. To make a call you need data from marketing research and I would assume that Slack marketing team knows more about it than HN crowd, simply because they are in position to have this data and HN crowd is not. It can be the case that they are not acting professionally, but I would not assume unprofessional behavior by default.
someone hav[ing] used a tool for personal reasons and may recommend it as a good tool for work is not enough to justify spending on those personal use cases
I am sharing that I have an equal and opposite opinion, which is that the someone having used a tool for personal reasons and may recommend it as a good tool for work sometimes IS enough to justify spending on those personal use cases
You are right that data can inform the decision, of course, keeping in mind that the benefits include intangibles like goodwill, which goes further in the techie community than in a layman community imo
Salesforce doesn't even care about paying customers. Heroku has stagnated for years. Recent Heroku massive mistakes they've been making have been absolutely terribly handled by them.
They also own Tableau, which has stagnated as well and is pretty badly handled.
I think this, combined with Heroku, marks end of an era.
These tools built themselves on being free (as in beer) up to a certain extent, and now they're big enough, and trust their momentum, so they change their stance.
This will inevitably create some changes, both how they perceived and being developed, but we'll see...
However, web application development is much easier and scalable infrastructure is democratized now. Their moat may not be as robust as they imagine.
They are doing something similar to Tableau. They have been actively rolling out features only available to subscription based customers - ie if you own perpetual licenses and only paying x% yearly for maintenance you are starting to get less product than their subscription customers. It is nuts and is driving us to actively consider PowerBI which is like Teams is included with the enterprise license.
What? Those free users are practically providing free infrastructure testing. Those low-no profit margins are hiding a lot of stuff underneath. Intel and Cloudflare are famous for reaching scale by using cheap/free end of the market to expand and solidify their positions.
I fail to see how letting crooks mine cryptos on Heroku’s free boxes would lead to improved sales. Slack is a different story, but all hosting/CI/etc providers have been pulling back free offerings as they get abused for crypto mining
Salesforce doesn't care. There about to roll out slack-first infrastructure that turns slack into the only chat-based CRM. Their core users aren't on the free model. They're paying a small fortune for Salesforce and see Slack's cost as a drop in the bucket.
Makes me wonder how long before Discord rolls out their enterprise tier. There are tons of Gen Z and Millenials using Discord who are likely in positions where they can make those kinds of decisions for the company. And it's a great and intuitive product too.
As a moderate user of both, they seem almost indistinguishable to me. Slack probably has more integration with productivity tools and a better auditing/compliance story, but it’s not at all obvious to me that Discord couldn’t become a drop in replacement.
How would you rate mattermost and zulip compared to slack and discord? I've only used slack, mattermost and skype. Of those I prefer mattermost by a fair margin, but I'm still not sure I like it. It seems like the ui could be better. Something about the way both it and slack handle channels or groups feels clunky.
Zulip is very “channels first” and once you get used to it it works pretty well. I constantly run into issues uploading files > 10Mb to Zulip though, it just sits there “Waiting” forever. I have to put things like screencasts somewhere else. Otherwise it’s pretty good though
Guilded over discord in my opinion. More features like forums in a channel. Long term, I only see discord and their company culture turning it into twitch or other zealot infested dumps.
I'm surprised to not see more comments like this. I think chat clients are a product which is used so frequently that after adapting, users become blind to defficiencies. I consider myself pretty technically literate and have used probably a dozen chat clients over years. Discord is the only one where I couldn't understand how to do the basics straight after installing it. I don't use it much, and when I come back to it there's always an element of guesswork. (fwiw I'm not a fan of Slack's UX either)
Because on the flipside, there are users like me who found Discord far more intuitive than anything else. I don't think your 'technical literacy' overcomes a difference in expectations and culture when many others feel it works just fine. It's not the best UI, but it is acceptable given the status quo.
Neither you nor GP are articulating or giving examples. Start there and we can have a proper discussion, otherwise it'll be just another case of 'he-said-she-said'.
What chat software do you use? I find Discord to be among the best when it comes to UX. Not perfect (and it is still an Electron app) but when you have atrocities like MS Teams, it passes off as good.
In case you are wondering, I don't consider the UX for IRC "good" (in a forum with "hacker" in its name, of course IRC will come up). It is good because it is open, simple and lightweight, but when you want a history and image/file transfer, and most people do, it breaks down and you need ad-hoc solutions.
That’s surprising. In what way? I find it to be the most consistent and the most correct out of any chat software I’ve used, including IRC, Slack, Mattermost, and Microsoft Teams. They seem to just get a lot of stuff right.
The only thing that took me some time to adjust to was how overwhelming joining a new server can be. So many channels! Everything needs my attention! But after using it for a long time, I think this is actually a feature. I can mute channels and categories I’m uninterested in, but how many channels on Slack are never found due to poor discoverability?
We run both Slack and Discord at my work and the engineers vastly prefer Discord, for what it’s worth.
I think I could be considered as a high openness, hacker-ish type keen to try new things, and my avoidance of discord isn’t that it’s complicated (hard to use), it’s that it’s complicated (unnecessary busy). From the notifications being a lot of noise by default to threads being so much work to catch up on after the fact, it’s just heavy and dense and complicated.
Caveat: It seems much more geared towards users who commit significant brain-space to the subjects they join (for example: catching up on threads is easier if you’re paying attention to realtime notifications), whereas I have a lot of interests and really just want the gems whenever I “sit down” to work with a particular interest.
Nope, it's possible to control the notifications to a much better degree than any of the other options allows.
I could see Discord as made for power users where Slack and Teams would be aimed at casuals. But I do think that Discord does a pretty good job at being intuitive.
Discord seems like the best UX to me. Only things I can think of is some weird UX around new call features and search could be better compared to say gmail (although I generally have no problem finding things I want to find).
Discord is way better than Slack for B2C if you're reaching that demographic.
I'm running my start-up via Discord and it's grown to a 60,000-member community. Incredible engagement. Fantastic way to find customers and business use cases, too.
For "enterprise", we run a separate instance for staff with lots of useful Github, Clickup, and other integrations.
While I understand that approach EF going Discord-forward, it's one that's prevented me from joining many circles now as I've pledged to take my data privacy more seriously and proprietary apps aren't it. I hope you offer at least one alternative (but if you're not targeting open-source users or privacy advocates, there likely wouldn't be a difference).
One nice thing about IRC is it is sure to remain an alternative as services which purportedly lack alternatives build paywalls or become the center of scandals relating to mishandling user data.
Discord is just way more polished. Everything works reliably every time.
The voice channels are flawless, even when there's people on shitty internet connections.
The client is always responsive and snappy.
Their design of how the server/chat/voice/videos/users is laid out is is clearer, easier to see what's going on at a glance and it manages to do that while it fits more information on screen.
Their dark mode is much better.
Their code display/highlighting is better.
The user groups/roles which is more like tags is much better than others.
It's way easier to find where notifications come from.
I like that in Discord you can choose to spin off the conversation into a separate thread, or you can reply within the main chat and it quotes the comment you’re replying to (and the quote is clickable) in order to see context. You can’t do the latter in slack, and it’s easy to not realize that others have started a thread when it’s above the fold.
No, discord threads are siloed off into their own world. They're effectively more so temporary channels (and discord itself enforces the temporary nature - they get locked after disuse). Parent poster seems to like it, but for my use case I'd rather slack threads.
I wasn’t commenting about liking discord threads. My comment was focused on how you can “inline” reply in discord, while you can’t do that in slack (though someone mentioned that you can post your reply to the main channel in slack - a feature I’ve never noticed or seen used).
Yikes, here's hoping you mean "after a year" or something, because having to make a fake new thread just to comment on an existing locked thread would make me rage quit
Someone with the manage threads permission can unarchive it, but like the pinning thing that also comes with permission to delete it, so you can't just grant that to everyone in your server.
Then that squares with my mental model of discord being for the gaming community, because imagine one's surprise about coming in to work on a Monday to have something to add to a thread made on Friday but unable to contribute to it. Or, what's this "vacation" stuff I keep hearing about?
Roles and channel permissions. I can't assign a role to someone and grant them granular permissions in slack like I can discord. I'm either a guest, user, admin or owner.
Think of it as "tagging" users. Instead of a user fitting a single pre-defined category, you can add tags to a user.
To give an example, let's say you have both a Discord channel and a Slack channel with the same 10 people working on a project, 7 of which are software developers. To mention the developers, your Slack option is either @everyone (including 3 non-developers) or individually. On Discord you can create @developers and add only those seven, not bothering the other 3.
You can really go buckwild with that basic concept: only admin can tag @everyone, only @talkers can join a voice channel, only @projectmanagers can run some bot, only @collaborators can post to #project, @watchers can read #project but not post, while those that are neither @collaborators nor @watchers don't even see the channel... And in #otherproject you can have those same people in totally different groups and with different roles.
Way more customizable than by just having public/private channels and @everyone at your disposal, as Slack does.
Slack user groups are much less powerful than Discord roles. Slack user groups are _only_ a way of bulk notifying users. Discord roles control effectively both views of your server, and permissions (though why "Pin Messages" and "Delete Other User's Messages" are still combined into a single "Manage Messages" checkbox is a glaring flaw in that granularity).
And some other features like colours which are nice when you want mods to stand out for public servers, or teams for some sort of competitive event server.
Tagging users is also possible in slack. I don't know if permissions by tag is possible, but the basic concept of "@ these 7 people" has been there for a while.
That, and their paid-for subscriptions are natural extensions of the already super useful platform. Either more quantity (actions) or more useful features on top of an already fully useful feature set.
Users pay for convenience and for more, not just slightly better.
Slack's leadership likely think that Slack is already so ingrained in corporate settings that people will go for it by default, without word of mouth from people using a free plan. There is enough word of mouth from people who used it on a previous job. Sort of like Jira.
I started using Slack in 2015, and never in my life used a free Slack account.
There’s no way to know how many corporate plans were sold because a single team member relied on Slack outside of work, or how many pro user accounts exist because someone was visiting a Slack-fluent friend.
So why assume it’s positive ROI? The people who make purchasing decisions in the organizations Salesforce plays in don’t typically make them based off recommendations of people happily using a free tier on outside projects but rather those made by massive sales teams.
This is not my experience. But don’t take my word for it.
Why else do you think students get software or checking accounts for free or much lower cost? The net LTV of getting users accustomed to your product who later evangelize it for free is priceless.
Underneath both, they make those choices based on trust. If a massive sales teams gains their trust it means sales. If employees have gained their trust it means sales.
Not only that but there are decisionmakers in organizations that also use free tiers for stuff in their private lives. Piss one of them off enough and suddenly you can lose a big sale or existing customer.
So tinkering with free plans downwards can have unforeseen consequences.
And with the amount of data slack uses so low and costs of bandwidth and spinning rust approaching zero - the idea that free tier limitation are anything other than create a painpoint for people to upgrade is a joke.
I agree with the idea, at the same time I'm sure there's some income vs expense graph that shows expenses raising faster than income and at some point surpassing it.
I wonder when it will hit github. At the moment they offer free hosting for open source projects, including github actions. It seems pretty clear at some point they'll have to back off in some form or another. Maybe the "open source" part is enough of a filter?
In their FAQ [1], GitHub state: GitHub Copilot is powered by Codex, a generative pretrained AI model created by OpenAI. It has been trained on natural language text and source code from publicly available sources, including code in public repositories on GitHub.
Genuine question, has there been any evidence to substantiate your claim?
I think all of this is possible through sampling and customer interviews. Besides Slack has fairly high awareness. The marketing strategy changes as your product moves through the awareness curve.
After the Salesforce acquisition SF may believe they no longer need this funnel, since Slack now has access to all Salesforce customers and vice versa.
This concept of market saturation is a well known concept, however, you forgot one important part, the next generation. It is well known under marketeers that is it is hard to change habits. An 80 year old is not suddenly going to change toothpaste brand because they saw an amazing ad on television. However, a rebellious teenager/young adult will want to pick their own toothpaste brand, and are susceptible to marketing campaigns. Personally I have been using a different toothpaste brand because my teenage children want it, and they offered bulk discount, so not worth buying my usual brand in addition.
This is the same with Slack/Discord. If I can attract bright individual to work for me if I offer them Discord, I don't really mind switching to Discord.
There's something to me about this that smells like it leads to bad assumptions; especially on things that have unforseen effects. You could argue that the internet was saturated in the 80s based on being connected only by a few universities. I don't think this is a well known science by any means or a well known concept.
Excellent tools spread through the best kind of word of mouth. Excellent tools that are free spread even faster and further. There’s no way to know how many corporate plans were sold because a single team member relied on Slack outside of work, or how many pro user accounts exist because someone was visiting a Slack-fluent friend.
And because there’s no way to know; the cost-cutters look at the numbers and say “cut the free plan enough to stop the hemorrhage”. Worst part is the next quarter looks great. And maybe the one after that. But soon enough the income starts dropping and there’s no “clear reason why” and then it’s too late to recover it. It has to be built again by building trust from scratch.