Why I don’t like Slack

I’ve been using Slack on a daily basis for around 5 years now. If I started a new project today I’d almost certainly advocate using it. But I have very little enthusiasm for it, and I have deep reservations about the idea that it will (or should) replace email.

Five things I don’t like

Slack is misleading on price

One issue with Slack is that it only stores the 10,000 most recent messages. Even in a small team you can hit that remarkably quickly. Of course, this is only a problem on a free account, but it exposes a bit of misleading marketing. They want you to believe that:

  • Slack is free to use for light users
  • Slack is an improvement on email

but these two aren’t true simultaneously. The free account is no substitute for email, because it can’t do what even a modest email box can do: preserve the important in preference over the ephemeral. If I get an email with my employee number on day 1 with a new company, I may well still want that email years later, but Slack doesn’t let me make that choice.

Slack puts the burden of organisation on the sender

Email does force you to spend time arranging your mailbox to keep things organised. Slack saves you this time in exactly the way that having your legs amputated saves time shopping for shoes.

Everyone has a different method of organising their emails. But that isn’t a problem, because email doesn’t require everyone to agree on how to do it. Furthermore, if I receive an email from someone who doesn’t care to classify their message by topic or urgency that’s no big deal, I can just do it myself.

On the other hand, Slack forces everyone to agree on what the list of channels should be. If someone creates a confusing or duplicate channel, look forward to long discussions with them trying to convince them it’s not necessary. Of course, if you do choose to stick with just one channel you can’t usefully merge the history of them. Did someone send a message to the wrong channel? Too bad, it’s stuck there forever, and when you’re looking for information on that topic you’ll have to give up and search all channels.

Threading doesn’t seem to work for anybody, at all, ever

I’ve yet to meet anybody, myself included, who uses threading consistently and effectively. When people are new to Slack, they usually miss the existence of threaded conversations altogether.

Even once everyone’s realised the feature is there there’s no real consensus on when and how to use it. Will branching off a thread lead to comments looking less important? If I decide a conversation is dominating the channel and needs to be moved to a thread, I have to pick a single person’s comment to branch it from, and will that look like I’m singling them out? After I’ve branched my thread there are a few more comments in the main channel, should I reply to them there and further the confusion, or awkwardly refer to them from the thread?

Discussions get terminated by confusion rather than consensus

Long slow-moving conversations are almost impossible to have on Slack, because only one topic can really live in a channel at any one time. Context rapidly gets lost. You could argue that threading would be the solution for this, but this relies on the conversation having been kept to a single thread the entire time, which doesn’t happen.

Sometimes I have an announcement I want to make but hold off on doing so because I don’t want to move the topic of the channel on from an important issue we haven’t reached consensus on. This is a communication tool standing in my way, not one giving me options.

There’s no way to mark a conversation as dealt with

In email, it’s usually fairly clear which conversations are done and which are waiting for input from you. In Slack they just scroll off the screen. Worse, there’s no accepted way to re-raise the issue to check if the person you’re talking to is still thinking about it or needs a reminder.

Similarly, there’s no way to spot important topics when looking back over an archived chat. If you’re away for a couple of days, good luck reading through a hundred messages and spotting the three that are important.

One thing I do like

Sending a Slack message to someone fits a nice gap in urgency between sending an email and making a phone call. Rightly or wrongly a rough consensus has formed that email can be ignored for an hour or two, but a phone call has to be picked up immediately. When someone’s online Slack messages are typically turned round in a matter of minutes, which neither email nor phone contact gives you.

This a level of urgency that seems very common in the type of work I’m usually doing, and I’m grateful for it as a tool.

The problem with “the new email”

Email has been around for a very long time, and many of its early design decisions show through as limitations today. Formatting is inconsistent and doesn’t mix well with quoting. Encryption is a pain. Attachments can be clunky. Quoting is useful in small doses but the default of dumping the entire previous email in below is unhelpful.

The protocols involved in email are hard to extend, and non-standard extensions leave us stuck with a lowest common denominator problem. It’s natural to argue that we need a complete break in order to build on better foundations.

Furthermore, the more annoying aspects of email are social. I’m old enough to remember when email was unfamiliar enough that the term “netiquette” was bandied about to encourage people to be mindful of their behaviour, for all the good it ever did. 2+ decades after email’s widespread adoption our ways are set and we’re stuck with people copying too widely, quoting too much content, endless inflation of the urgency markers in the subject line, and so on.

Again, maybe we need a clean break.

But not Slack.

Thoughts on Twitter, ten years on

I often write draft posts with ideas that I intend to come back and fill in later. Given how long a break I took from blogging, this has led to some serendipitous time capsules. Here’s an unedited note I wrote myself in 2011:

By making it impossible to publish thoughts in their proper context, Twitter will only strengthen the effect of cliques and groupthink online, since the only people who can understand the idea shorn of context are those who belong to an important subgroup.

I don’t suppose that I was the only person at the time worried that Twitter might drive polarisation, but I’m still quite pleased with how well this prediction has held up.

I can’t find good figures, but I feel like blaming Twitter for societal problems didn’t really ramp up until 2016. In 2011 the economy might not have been great, but there was still a fair amount of faith in our political institutions. Barack Obama was still in his first term and had just passed a health care reform of historic scale. The Arab Spring was still showing promise. Twitter was just 5 years old and I remember people lauding it for its role in coordinating the protests in Tahrir Square.

I don’t think my prediction was dead-on, though. For a start, I doubt that when I wrote “cliques and groupthink” I had any idea of the seriousness of the effect this might have on the real world. I also feel like I was overstating the importance of understanding the content against the more relevant issue of emotional valence. On the Twitter of today Republicans and Democrats know perfectly well what each other are saying, they just find it distasteful.

In fact, I don’t think it’s that simple. There are plenty of terms like “white supremacy” or “defund the police” where people on different sides of the argument talk past each other because they have different understandings of what these concepts entail. Not having to compress one’s thoughts into 140 (or even 280) characters wouldn’t fix this, but it couldn’t hurt.

A bigger miss in my 2011 way of thinking was overlooking that Facebook would be a far bigger issue than Twitter. The issue with Facebook isn’t to do with the lack of information within the post, but of the context the post is displayed in. In other words, my worry was that there would be too little information in a Twitter post, but on Facebook the problem is that we’re drowning in so many posts that we only pay attention to (or rather the Facebook algorithm only surfaces) the stuff that riles us up.

I’ve been complaining about Twitter for as long as I’ve had this blog, and since before most people were even on Twitter. In fact, my 2009 post about how the community is far more important than the technology that runs it was one of the thoughts that drove me to create a blog in the first place. Even my very first post contained a jab at Twitter and name-checked Neil Postman, who has lately become trendy among media commentators concerned about social media. Hopefully this raises my credibility when it comes to analysing the social media environment. But predicting something early isn’t the same as predicting it well, and I shouldn’t let my early skepticism alone determine how I see things today.

It’s a trite conclusion, but Twitter seems neither as bad as I had feared, nor as grand and transformative as its boosters imagined in the early days. To me, that’s a reminder that the relevant issue isn’t the capability of a technology per se, but the way in which society shapes and is shaped by the technology. That’s inherently a messy process with results that are hard to pin down.

DuckDuckGo is making progress

Recently I’ve noticed that both here and my other blog I’ve been seeing only 80% of my search engine traffic coming from Google, with DuckDuckGo taking a solid second place at 10-15%. This is actually a surprise to me, both that Google is as low as it is and DDG so high.

The audience here presumably skews more technical than the average site, so this doesn’t necessarily tell us much about what’s going on in the web at large. But if DDG is making headway with a technical audience that’s probably a good sign for the future.

I’m glad to see DDG getting some kind of toehold in the market. I don’t know how far I trust their claims about improved privacy, but any kind of competition has to be good for the search engine marketplace. I care a lot less about the details of what Google does than the fact that they are a monopoly, and that can never be good for the customer. This is the same reason I’ve stopped using Google Analytics.

Every time DDG comes up on Hacker News there seems to be a range of people saying that its results are too bad to be usable. I find this curious; I’ve been using DDG for my main search engine for about a decade now, and on the occasions where I retry a query on Google I rarely find much improvement.

Of course, DDG is bad at hyper-localised queries that require some knowledge of your location or previous activity to make sense of; that’s kind of inevitable with a search engine that doesn’t track you. To me that’s a sacrifice worth making.