Guardian comment system changes: The perils of designing for all users, not just the vocal ones
The Guardian are currently trying out some changes to their commenting system. Like most changes to a major website, the backlash amongst some users is very, very vocal, and everybody gets to watch:
“It looks like Fisher Price were allowed to design it, and it kills linear conversation.”
“Too late; this reader is leaving. Taking my click throughs with me. Bellis [Julia Bellis, a product manager at the Guardian] you have my email, if you want real professionals to work on this get in touch but otherwise I outta here.”
“Who ever designed this system and decided it was a good idea should be fired.”
“This format is not merely ‘pants’ - this is pure, organically grown, ‘M&S tramp pants’, dipped lovingly in rancid urine, and smeared with faeces spattered forth from the rectum of someone suffering from dysentery.”
I don’t want to keep rocking up at the Guardian’s table like the Ghost of Banquo, but it does perhaps give me a chance to publish a little bit of context. Over the last decade I’ve been involved in a lot of site re-designs, re-launches and changes, and have often been one of the public faces in the comment threads afterwards engaging with the users about them.
One of the things that is incredibly hard to get across to the really active members of a community is that on a site the size of the BBC’s or the Guardian’s, or any news organisation for that matter, the people who leave comments every day are very much the statistical edge case. They may be the main users of the existing comment system, but they are not the main users of the site.
If you think about it, a site like the Guardian claims an audience of around 60 million uniques a month. The number leaving comments will be in the low tens of thousands, possibly less. The number leaving comments every day and “living” in the threads will be in the hundreds.
The design challenge isn’t always about super-serving the group of people who are participating, but designing for the majority of people who aren’t participating but might be encouraged to. So when you see a comment thread with two, three or four hundred comments saying this is dreadful, you don’t know what you are doing, you aren’t listening to your audience, you have to remember that this is a self-selecting section of a much larger total audience, and that the Guardian will have also canvassed that larger group for feedback.
Away from the regular users, that feedback will often be along the lines of not wanting to take part because they find the commenting system intimidating, and that the existing users feel like a clique with lots of in-jokes they don’t get. A design that works to be more inclusive, and tries harder to get those people engaged, is always going to be in tension with the requirements of the users who are already comfortable with the system.
The trouble is, of course, that the vocal minority are right. Sometime new features do ruin their experience and the way they interact. I’ve often thought that some of the more fun bits to get involved in at the Guardian site were where the threads were more like an open IRC channel that just happened to be underneath a Guardian article. But there aren’t that many of those, and there aren’t a huge number of people taking part, compared to the number of people who are visiting the site and who are possibly taking away a negative experience because of the existing comment format.
Those using the current system are concerned that their conversation and banter will get splintered and fragmented. What they can’t see is that there is a pent-up demand for being able to simply scroll past all of their puns and banter to get to the next substantive point.
One of the changes the Guardian is currently rolling out is the nesting of comments. I have to say that despite it being one of the most requested features on the site from users, I did once proclaim that the Guardian website would have threaded comments “over my dead body” because personally I hate them. But then I had my mind changed by the user data. The first tentative steps towards this was almost exactly a year ago with the introduction of a “response” button. At the time — guess what — the community on the whole didn’t like the change, and lots of people vowed that they would never use the button. In fact after a month or so it transpired that around 40% of comments on the site were a specific reply to another comment.
Likewise, in earlier trials on the site of threading, it didn’t work in a couple of places — one football blog post in particular was an absolute train-wreck — but after a few days running it consistently in a couple of areas of the site people adjusted to change, and the key metrics of participation and moderation costs moved in the right direction.
Inevitably in the comment thread about the Guardian changes, someone has said “if it ain’t broke don’t fix it”. What the vocal edge case can’t see is that it is broken for the 98%-or-so of users who never comment.
Still, what do I know?
According to the most vocal bit of feedback I ever received on my work at the Guardian, I’m just “some fuckwit in IT” who “got promoted and decided to change some thing.”