Ravelry's first mistake was in doing an interface redesign in the first place.
Ideally you should never do a complete, ground-up redesign of an existing, well-loved site. There are reasons why people love the site!
Folks tend to get into complete redesigns because of reasons that don't actually matter.
Example 1 (Ravelry): You're doing a brand overhaul, changing your logo, and you want the site to match.
I promise you, you are the only person in the world who cares about brand cohesion.
Example 2 (metafilter & others): You're doing a new interface because google says you have to.
Again, your users don't care what google thinks. Why would they? Your search rank dropping is your problem, not theirs, and thrusting an unnecessary interface change on folks is just an attempt to take google's problem (which google made your problem) and make it into their problem.
And that's problematic.
(while we're on the subject, I just want to emphasize how much of a fool's errand it is to try to get into the top page of google results. You'll spend months and months doing it, and then someone in some sweatshop will rewrite your article in a weird stilted keyword-stuffed voice, add a table of contents in a blue box and big headings marked INTRODUCTION and CONCLUSION like they're back in high school, and knock you off the top with a worse article, because search engines generally suck)
Example 3: Your codebase is old and creaky and your site is still laid out in tables etc, and you just tried adding some new feature and it's a big pain in the bum because of all the creakiness, so you decided to recode it all and heck, why not give it a new coat of paint, since you're in there already.
Again, your users don't care! Who gives a heck if your site uses tables! That just means it'll probably still work in twenty years!
The only people who care about your HTML or will ever even look at it are other web designers, and as I'm sure you're aware by now, nobody cares what web designers think.
(except other web designers of course)
So these are all bad reasons to do a big ground-up site redesign. In general there aren't good reasons!
Example 4: you're redesigning your layout to make the header smaller, or take stuff from the top of the page and put it on the side instead, because computer monitors keep getting shorter.
This is actually legit. At least until we get the whole screen situation under control and start making computer monitors again instead of just small tellies, we're gonna have to deal with lousy displays, and putting stuff on the side instead of the top can help in that regard.
(but don't be tempted to put in a javascripty three-line menu thing. If something that used to take one click now takes two, that's a bad redesign)
Anyway, Improbable Island's retheme falls into category 3, rewriting because the old one was a big pain in the bum.
Because this ain't my first rodeo, the very first thing I did was to recreate the old theme, as near as "to the pixel" as I could, in modern HTML/CSS.
The second thing, because Ravelry is fresh in my mind, was to LOUDLY and PUBLICLY and ENTHUSIASTICALLY and REPEATEDLY open the beta testing period to *every community member.*
Seriously, whenever I'm in chat I'm asking folks how they like the new theme, getting feedback, and converting people over to the new theme.
Ravelry only opened their beta theme test to a handful of people. Don't be like ravelry.
Next, I made some very small improvements and ran them past the people using them. Most were small quality-of-life improvements, adding spacing on specific elements to avoid finger mistakes on mobile, that kind of thing.
Then, I made one big dramatic change, but tied it to a new feature (backgrounds that change depending on what you're doing in the game) that would be impossible in the old interface.
And people love it and are excited about it.
You *can* do an interface change humanely!
To do a successful interface overhaul, you have to be intimately familiar with the reasons why people love your site, and the things they love your site in spite of. Fix the second and leave the first well alone.
The things in the middle, the things people are indifferent about, can retroactively become people's favourite things if you mess with things too much.
Make your new skin a carbon copy of the old one, and slowly file off the rough bits. Do this live! Don't make a bunch of big changes on your test server then upload them to make a big change all at once, that freaks people out. Let them see the improvements happen as you make them and talk about them.
You want your members to feel that you're being slow and careful and considerate. To have them feel that way, first you have to actually BE slow and careful and considerate. Second you have to SHOW them that you're being slow and careful and considerate, by making changes gradually - testing locally and then uploading and asking for feedback on each one.
Let users be a part of the process and make it clear that this redesign is for THEM more than it is for you.
Barometers to see whether it's going well:
* Someone says "I just swapped back to the old skin and it looked WEIRD and BAD"
* People talk about interface *improvements* rather than interface *changes*
* People talk each other into trying the new skin
* People are curious about upcoming changes
* People ask you to fix other little rough edges while you're in there
Red flags that it's going poorly:
* People ask if you're going to keep supporting the old skin, and if so for how long
* People talk about interface *changes* or put scare-quotes around "improvements"
* People mourn the old skin
* People worry about things they like disappearing
* People ask you not to change specific things
Chucking this into the community management thread 'cause it reminds me of a thing that happened to me:
Our out-of-character chat area used to be called "Banter." We changed the name like five or six years ago. We had good reasons for doing it, not ones that I remember off the top of my head, but I do recall there was much discussion about it.
Players continued calling it Banter, and still do to this day. Everyone knows that it's called Banter.
So yeah posts are still gonna be called toots.
A thing I learned the hard way is that if you're gonna make a change, or even *talk* about making a change, do it late morning. Don't do it late at night just before you go to bed, because then when morning comes everyone's been talking about it all night.
That means that you spend half the day catching up with everything everyone said (and driving yourself bananas while you're at it), and it also means that the conversation has probably veered wildly into Speculation Town while you slept.
Speculation Town is a fun place to visit, but you wouldn't want to live there. As an example, once I talked about doing some visual improvements to Sheila's Shack of Shiny Shite (a weapons and armour shop). At some point someone went "Well, as long as she's still Australian, I'm cool with it." By the time I woke up, people were grumbling about my heavy-handedness at even THINKING about replacing their beloved Sheila with a vending machine.
(because posts in chat are trimmed to 100 pages and there were a lot of posts that night, I couldn't even rewind the chat to figure out how people went from "Alternating table row colours to make it easier to read" to "Scrapping a major character and retconning her into a vending machine." I still have no idea but it falls into that category of information where you can just say "Game of telephone" and that tells you everything you need, the details aren't really important)
As an admin you've really gotta look after your own mental health very carefully, 'cause there's a lot of work involved that boils down to doing other people's emotional labour for them.
By that I mean calming down people who can't calm themselves down, helping people to remember that the admin is a person too, and where necessary removing people who've gone too far down paranoia rabbitholes to still interact with the community in a non-harmful way.
This is pretty tiring.
Looking after a community of thousands isn't something that anyone can naturally be good at. Websites have only really been around a couple of decades, just aren't wired for it. We're wired for maybe one in thirty or forty of us being leader of a manageable little village where folks generally have similar needs. Nobody can have an innate talent for looking after tens of thousands of hugely different people, that's something you pretty much have to pick up over a decade or so.
Expecting your admin to be good at his or her or their job is a BIG ASK, honestly.
Expecting them to be good at it in a way that benefits you, specifically, all the time, is just flat-out nuts.
An admin's job involves occasionally annoying or inconveniencing some portions of the userbase, for the benefit of others. You're gonna have to do that. It's gonna suck.
(obviously I'm not talking about accessibility here. If someone's blind or has to type one-handed or needs different contrasts or anything else that's necessary through no fault of their own, obviously you should bend over backwards to make the site usable for them even though they're a minority of the userbase, not only because that's the right thing to do but because improving baseline accessibility ultimately helps everyone)
Online Community Management Thread Continued, a thing that happened today in which I wanted to talk about harmless misunderstandings.
In my game, you can make Mementos. These are items to which you assign a description and some text that appears when you "use" them. You can gift them to other players, show them off in your bio, attach stats to them, use/query/gift them in conjunction with little programs in player-owned buildings, that kind of thing.
Anyway, a long-time player in good standing made a memento that talks about someone called Kim Keller. The memento used the colour blue heavily. At the bottom, it said "Thank you LEO."
Being unfamiliar with the work, I DDG'd for Kim Keller, and turns out they're some republican arsewipe. This, plus the blue, plus the "thank you LEO" made me wonder if this was some sort of blue lives matter nonsense, but it seemed very out of character for this player, so I messaged them about it.
Turns out Kim Keller is a character in a series of graphic novels made by a guy who goes by the initials L.E.O. It had absolutely nothing to do with the police and it was all just a series of coincidences that led up to a harmless misunderstanding.
But if *I* had that misunderstanding, it's reasonable to assume that my other players might also have the same misunderstanding, and have cause to feel unwelcome or unsafe.
So we talked about it a little and we ended up erasing the memento, which I didn't feel good about (I would have preferred to just change it so that it didn't set off any potentially harmful misinterpretations).
The important part here is catching these things early, because if someone misinterprets this thing to mean blue lives matter, and talks about how that makes them uncomfortable, and the creator isn't around at the time, then that sort of convo can spiral.
This is the sort of thing that can easily lead to long and often unproductive conversations about what's allowed on the site, about censorship, about consideration of others' feelings, about culture wars in general, and then twelve hours later when the original author logs back on they're confronted with this absolute FIREHOSE of Discourse about a thing that they *didn't even mean to say.* They were literally talking about something completely different!
By then the conversation isn't even about the thing they made, it's about These Things In General, and the creator will either have horribly uncomfortable and complicated feelings about this whole Thing they've accidentally unleashed, or they'll take one look at it and go "Nope," erase the thing, post "I was actually talking about a graphic novel," and go hide for a little while until it all blows over.
Moral of the story is that it's _totally okay_ to apply your moderation privileges to things that are absolutely, 100% harmless, but which have the potential to be misunderstood to be something hurtful. Catch them early before they're seen and misunderstood, explain to the creator that you know they didn't mean it like that, and where possible work with the creator in question to resolve the issue and avoid misinterpretations.
Online Community Moderation Thread Part 93819: the same story going round and round forever
Been watching this Scott Cawthorn thing play out (the Five Nights At Freddy's guy, who it turns out was secretly donating the game's profits to far-right abusers while publicly donating to useful groups to give himself a veneer of respectability).
If you've been moderating for a while you'll see the same social mechanics play out over and over like clockwork, and you'll feel frustrated at this.
You'll be like, "This only last happened a couple of years ago, don't people learn?"
Remember, your site has new users since the last spin of the merry-go-round. Some of them are in their teens!
The FNAF community is handling their Milkshake Duck situation poorly because they trend so young that for many of them, *this is their first Milkshake Duck.* Their first ever! You remember how much you flailed and how much of an ass you made of yourself during *your* first Milkshake Duck, don't you?
You'll continually go through the same stories over and over until a portion of your userbase is old enough to recognize these cyclic events, and then you'll keep going through them because you have new users, but the old users will roll their eyes and go "Pfft, kids today."
The solution - I'll say it again! - is to list common harmful social mechanisms *right there on your website,* so people recognize and derail them on the fly.
Again, it's gotta be right there on your site, and *tailored to your audience,* and in a place that's easily accessible for reference. If you're a gerbil site, talk about the time Yummy Nuts Gerbil Food partnered up with the Handmade GerbilSkin Pillow guy and the absolute mess of how your community responded. Your community has some cultural memory but on the internet it's not worth Jack if it's not written down.
(armchair admin moment: Y'know what I would've done if I ran the FNAF subreddit during Milkshake Chica? I would've posted a notice and then shut that crap down. Gone read-only for a couple of days. Milkshake Duck events are a time for personal reflection and emotional processing, not for reading the hot takes of internet strangers. ("Hot take" is shorthand for "I haven't thought this through at all but I want to be the first to post it in case I get lots of redoots or uptweets or whatever."))
(surfing random waves of algorithmically-amplified emotion is not a healthy thing to do during a traumatic event. A Community Consensus will attempt to form within literal minutes of the news breaking, regardless of whether anyone's had time to sit with their emotions or not.
Of course shutting down a subreddit while it's at the absolute peak of its traffic is unthinkable, because corporate-owned internet wants us to think that numbers going up is good.)
(it's often a good idea to do the things that corporate-owned internet would find unthinkable. Corporate-owned internet does not have its users' interests in mind.)
One wrinkle to the "Shut it all down and let people think for a minute" approach is that for a non-zero portion of the userbase, this community IS how they think, how they process, how they identify themselves, and suddenly yanking that away and forcing them to sit with their own emotions would be cruel to them. I'm aware of that.
I said earlier in this thread, "You will have to deliberately hurt someone." That's important to keep in mind if you want to run a community website.
Online Community Moderation Thread continued, red flags for an early ban:
* using "god" or "hero" or "deity" or other wanky self-aggrandizing crap in their username. You'll wanna be up to speed with the semi-obscure names of almost-forgotten gods that undesirables tend to use, including spelling variants (we had a guy who named himself Asmodius who ended up in prison, you can probably guess why but it's actually worse than you think).
* framing commonly-held community standards as "unwritten rules" and demanding that every infraction of theirs results in a new Official Rule. This ends up with you writing new rules for one person - which means making everyone read the new rules (and go "huh? people were seriously doing that?") when really only one person is the problem, and that can easily be solved with a ban rather than pointlessly inconveniencing your entire userbase.