Jakob Nielsen wrote a nice set of Error Message Guidelines back in 2001. I think his guidelines still hold up, so let’s see how HTML Dog’s 404 page measures against them.
“Explicit indication that something has gone wrong.”
HTML Dog’s 404 message begins decently enough. “404! Page Not Found!” The exclamation points are a little extra, but at least it puts the relevant info upfront.
Then the screaming begins.
I can easily imagine certain older members of my family legitimately becoming worried if they read “Ahhhh! Panic!” Those words are so unusual and alarming. Accusatory, even. As if to say, “Look at what you’ve done.” People who don’t understand what a 404 message is or how webpages are retrieved over the internet might read something like “The authorities have been notified” and start to wonder if they’ve inadvertently committed some sort of cybercrime. The message declares that someone ought to be “sorry.” It’s all a bit… much.
“Human-readable language, instead of obscure codes or abbreviations”
I think the issue here is that HTML Dog is trying to be too “human-readable.” Normally error messages are indeed full of obscure codes and abbreviations, so kudos to HTML Dog for avoiding impenetrable tech jargon. I might argue, however, that the Justin Bieber reference counts as obscure. Billions of YouTube views aside, “Sorry” is several years old at this point.
“Polite phrasing that doesn’t blame users or imply that they are either stupid or doing something wrong”
“The person responsible will be fired. And cast out into the wilderness,” might be “polite” in the sense that it’s technically talking about someone other than the user, but keep in mind that most visitors will haphazardly scan the words on the page, not start from the beginning. If the words and phrases lingering in one’s mind after a cursory scan include “fired,” “cast out into the wilderness,” “this upsetting circumstance,” “Ahhhh!” and “the authorities have been notified,” that’s not what I would call a pleasant “user experience.”
“Precise descriptions of exact problems, rather than vague generalities”
This 404 message is as precise as it can be. It is true that the site wouldn’t necessarily know the exact reason why the error was returned, so it’s nice that it suggests three possibilities. It’s also helpful that that the most important bits are bolded. That being said, all the extra copy muddies that precision and might even introduce some of the “panic” that it sought to dispel.
“Constructive advice on how to fix the problem.”
HTML Dog is admirable for including a “What Now?” heading on its 404 page. Suggested next steps for the user are often missing from error messages. However, the non-hyperlinked parts of the bullet points are anything but constructive. “References: Refer to stuff,” to take but one example, is neither informative nor funny. And is the bizarre aside about Wikipedia supposed to be a pun about “roofing” because the site is called HTML “Dog”? Who the hell knows, man.
Jakob Nielsen’s guidelines are worth a look if you haven’t read them in full. Error messages are never fun to encounter, but I’d rather read one that calmly explains the problem and suggests next steps than one that tries too hard to be cute.
I had a great time at the virtual World IA Day Toronto 2021 event that took place last Saturday, where I was asked to present a lightning talk about finding an information architecture job in the current market. Since all of the lightning talks were prerecorded, I can easily share with you the video of my ~8 minute presentation. The local WIAD team even manually captioned my video, which is beyond awesome. I can’t wait for next year, but in the meantime you can check out my talk below.
Actions have consequences, no matter what. But the modern day architectures of information in which we carry out many of our day-to-day actions have a tendency to obfuscate and distort our perceptions of those consequences, and indeed even the true nature of the actions themselves.
This morning I watched a CBS News interview of Michael T. van der Veen, impeachment defense lawyer for former President Donald Trump. I encourage you to watch the full interview below.
van der Veen begins the interview by somberly reiterating his lack of disbelief that politicians will say and do anything, including doctor evidence. At around 1:20 in the above clip, van der Veen’s emotion begins to show, as he is faced with his first dismissive comment from anchor Lana Zak.
At around the 2:30 mark in the above clip, Lana Zak responds, “Let’s follow up with the point you’re making right now about the house managers as you say doctoring evidence—”
van der Veen steps in. “They didn’t deny it. I put it in front of them three times.”
“What I’m telling you is that they doctored evidence. And I believe your question says, ‘well it’s only a Twitter check and changing a year of a date here.’ They switched the date of a Twitter a year to try to connect it to this case. That’s not a small thing, ma’am. The other thing they did is they put a check mark on something to make it look like it was a validated account when it wasn’t. And when they were caught, they didn’t say anything about it. They didn’t even try to come up with an excuse about it. And that’s not the way our prosecutors or our government officials should be conducting themselves. And the media shouldn’t be letting them get away with it, either.”
Anyone who is familiar with the “Inspect Element” option on their web browser can tell you how trivial it is to edit the HTML of a webpage, take a screenshot, and then pass off the result as a real historical document. Or to open up Photoshop, prepare an image of an unverified tweet on one layer, open up a PNG of the Twitter verification check mark on another layer, then click and drag the check mark image into place in order to complete the forgery. UX designers create “mockups” of their prototypes and ideas for user interface designs in such a manner all the time. Once you’ve gotten practice, it doesn’t take much effort to create images like this. Click, click, click, drag, click, done. No big deal.
However, when a falsified image is presented before the United States Senate under the guise of evidence, it is a crime. Thank God for brave men such as Michael T. van der Veen, who still have the courage to stand up and speak the truth. It can be taken as evidence of our culture’s ongoing and accelerating slide into what conservative Christian author Rod Dreher dubs “soft totalitarianism” that his brave defense of the truth was unfortunately not without personal, negative consequence to van der Veen and his family. Terrorists tracked down van der Veen’s address and vandalized his home. As the local Philadelphia CBS affiliate reported, “Someone spray-painted ‘traitor’ at the end of the driveway of his Chester County home with an arrow pointed toward the house.”
Committing a crime used to, well, feel like committing a crime. Back before commercial image editing software became ubiquitous, doctoring images would have required, at the very least, “scissors or ink.” The deed could not be done with the same number of clicks as any other trivial action that we do all day on the computer at our regular jobs.
Andrew Hinton, author of the phenomenal book Understanding Context: Environment, Language, and Information Architecture, has demonstrated how a simple change in context can lead to a huge change in behavior. Actions that may have once felt obviously, tangibly wrong now lack the palpable, material feedback that our human bodies perhaps relied upon to inform us of the true nature of our actions.
Hinton illustrated this in a 2013 talk at World Information Architecture Day in Ann Arbor, Michigan. His point was how the design of Twitter scrambles the “ecological information affordance for action.” The difference in actions that a user must take to send a direct message (a private, one-to-one correspondence) versus a regular Tweet (a public, one-to-many correspondence) is subtle and easily confused. The “real-world” equivalent—that is, a person whispering a private message into someone’s ear versus standing in front of an audience with a megaphone—would never be so easily confused. An action that a user would clearly not want to take under normal circumstances in the built environment becomes much less obviously so when the same action takes place in an information environment.
Fortunately for those of us who value truth and facts—wherever they may lead us—this subtle manipulation of user behavior can cut both ways. Just as information environments can lead users to underestimate the ethical severity of their actions, it can also lead those same users to underestimate what they would need to do to “cover up” those actions.
Perhaps the impeachment lawyers believed that because Twitter management had deleted Donald Trump’s Twitter account and erased all of his Tweets from their site that no one would be able to fact-check the lawyers when they presented their doctored evidence. It was this carelessness on their part that, depending on your politics, either disintegrated the integrity of their other arguments or exposed the weakness of their entire case. All I can say is thank God for Archive.org.
Twitter is not the only tech titan whose information architectures obfuscate the true nature of actions that in the built environment would be obviously and tangibly criminal. The climax of the Reddit-GameStop fiasco that took place late last month was when Robinhood, along with several other securities trading platforms, actually removed the “buy” button from their interface in order to manipulate users’ perfectly lawful and reasonable behavior.
In a purely “real-world” environment, wresting control of countless customers’ financial assets away from them—against their will—would entail, for starters, a massive, worldwide information blitz to tens of thousands of workers at banks and other financial institutions. All of these workers would have to be informed (and convinced) why retail investors’ actions that were perfectly legal and acceptable moments previously must now be forcibly prevented at all costs. Next, arguably, these institutions would need to load up on security personnel and other measures as innumerable aggrieved and justifiably irate customers would begin to inundate these institutions with their cries of foul play.
However, in the information environment, all that needs to be done to disenfranchise millions of customers all at once is to highlight the few lines of code that produce the “buy” button, hit backspace, and ask the UX writers to draft some condescending fluff about how they’re keeping users “safe” from their own ignorance and unsophistication.
What we need is a renewed vigilance on our part as inhabitants of information environments. To return to the lessons of Andrew Hinton’s talk, the Venn diagram of “online” and “the room” looks more and more like a circle with each passing day. The resulting danger is that we lose sight of the consequences of actions.
We must not downplay or trivialize, as CBS News anchor Lana Zak and the impeachment team did, actions that take place online or on a screen. Those actions are no less “real” than those that take place offline. A crime is no less criminal whether committing it entails hours with scissors and ink or whether it entails “merely” changing a 1 to a 0. Notice it or not, the consequences are felt. Even if they aren’t felt by you, believe me, they are felt by others.
This year marked my first (and apparently last) trip to the annual IA Summit. Chicago hosted the event this time around, and with the city being an hour’s train ride away from my home suburb, I naturally couldn’t miss it.
It’s hard to know where to begin with a recap of my experience at this conference. I was blown away by the care and attention to detail put into everything from the conference swag to the website, which was continually updated throughout each day of the conference. Even now, the site functions as a treasure trove of slides and other resources from the talks and workshops.
First thing before getting started on content – “Are you aligned on intent?”
We see an opportunity to…
With content for…
So that they can…
E.g. “We see an opportunity to increase new and recurring orders with content for single, Paleo-minded athletes so that they can feel confident that Origin Meals will help them eat to perform.” <—— This is your strategy statement
Tool – Content Prioritization Matrix
Focus (good for user and business) (spend 60% of your budget on this stuff)
Drive (good for business, users don’t care)
Guide (not beneficial to business, but users want it) (spend like 25% of your budget on this stuff)
Here’s how I filled out this document – see items on the right side for examples of “content items”
With that exercise… personas from user research help going into this, because some things might be focus for one person but drive, etc. for another. (For example “athlete testimonials”). And you could do this exercise for each different persona… and then overlay all the results and find the most important stuff. Or you can plan for cookies that deliver different content to a user.
Here’s how my group and I made it through this exercise
Here’s how my group and I would prioritize the content presentation for a page
Here’s how my group and I would specify the messaging framework for a page
If clients fight you on taxonomy terms on your sitemap, you can pull up analytics data to back up your word choice.
It’s not obvious to everyone that tons of content isn’t harmless. Convince clients that it’s worth getting rid of that clutter. (Has SEO benefits… could have server cost benefits based on how large the site is.)
Meghan recommends that content owners sign contracts saying that they’ll maintain it.
“I think most content problems are people problems.” – Meghan
Example of very structured content – a health site that uses different words for things based on how weight-conscious the user is
Things you could do now…
Prioritization: conduct an assessment of a sample of your content and apply the prioritization matrix. use what you learn to make a case for further content design work
Organization: Identify the top 3 to 5 reasons users come to your site and document paths through your content you’d like to see users take. Then review the content to ensure it supports the desired pathways
Presentation: Pick priority content pages based on the prioritization and organization findings and hold a core model workshop with content owners, SMEs, the UX team, etc. Use workshop outcomes to recommend changes to how content is currently presented
Specifications: work with the appropriate people to develop a messaging framework to help content creators create on-message, on-strategy content
I could practice this on my own by coming up with more ideas for websites (e.g. a charity! a shoe store!) and then doing the prioritization exercise and other exercises as well.
I returned to the city on Thursday, not for any more workshops but for a “First Timer’s Dinner” with Peter Morville. A group of about 6 of us ate together at a sushi restaurant just a few steps away from the conference venue.
To me, it was these sorts of clear opportunities to sit down and have an extended chat with some of the biggest names in our field that really made the IA Summit live up to its goal of being welcoming.
(I also managed to get Peter to sign my Japanese copy of Information Architecture.)
The Talks I Attended
Each attendee’s lanyard came with schedules for each of the three days of the conference. Below, I’ve highlighted the sessions that I attended.
Welcome to the 2018 IA Summit!
Opening Keynote – There is No Artificial Intelligence without Information Architects, Seth Earley
IA at the Helm: Leading with Information, Bob Boiko
Prototyping Information Architecture, Andy Fitzgerald
Information Arrangement: It’s the Metadata, Dalia Levine
Designing Our Futures, Erik Dahl
Architecting Information for an Open Source Citizenry, Rachel Knickmeyer and Greg Swindle
A Strategy for Ethical Design in the Attention Economy, Samvith Srinivas
Evening Keynote – Different is the New Normal: Why Everyone Benefits When We Design for Disability, Elise Roy
I concluded the night with a dinner with my sister (8pm was a little too long a wait for me), though I stopped by the happy hour afterwards.
(I got a late start Saturday given my impromptu stay in Chicago that night and my waking up late as a result of that.)
No Static: IA for Dynamic Information Environments, Duane Degler
Group Mentoring lunch with Dan Klyn and Abby Covert (one of the highlights of my Summit; see notes from that below)
Connected Content: The Future of Information, Carrie Hane
IA Lenses: A New Tool for Designing Digital Structures, Dan Brown
Folk Illusions: Embodied Cognition for Today’s World, Claiborne Rice
Evening Keynote – Postcards from the Edge, Jason Hobbs
Dinner with the Chicago UX Strategy meetup
IA and Ethics – 2018 Roundtable Redux, Surla, Rice, Resmini et al
Morning Coffee. (I believe this is when I picked up some discounted books that I’d had my eye on for a while)
Going Global: The Intersection of IA and UX, Blanch, Shew, and Sengers
Don’t Make Me Wait! User Perception of Time & Software Speed, Chris Kiess
Using Stuff I Learned at Previous IA Summits to Set Up a CMS as if Content Mattered, Kristin Rowley
Bringing Everyone into the Process, Whitney Quesenbery
Closing Keynote – Marsha Haverty
Notes from the Group Mentoring Lunch (with Dan Klyn and Abby Covert)
What is the IA deliverable?
Maybe the process is where the value comes
Facilitating decision making
The meeting is the deliverable
Making the complex clear
You only get [IA, etc] into the culture by making everyone feel good after doing X
IA is the process, not the thing
Show the client a picture of what they’ve always seen and never seen
Getting started in IA
Find someone who’s doing what it look like IA is
Even Dan wonders about these definitions (“6.5 out of 7 days I don’t know what content strategy is”)
What do hirers look for?
“How do they sweep the floor?”
It’s the “who you are” not “what you do”
Don’t bluff. How do you deal with not knowing?
You can ask at an interview, “What does my 4th week here look like?”
You want to be a human at this company, not a body.
Is there such a thing as an entry-level IA?
How much coding does an IA do?
“The more you know, the more you’ll do.”
You don’t need it but learning it is not a waste of your time
It’s valuable, for sure
That said, many IAs don’t know coding. They’ll use InVision or something to prototype their ideas
Key difference between content strategy and information architecture
Information architecture is the car, content strategy is the gas
Can an IA be good only for me? It’s subjective, right?
Yes, for example in the movie High Fidelity where the records are sorted by the owner’s personal experience. That’s perfect for him and only him.
Conference Talk Notes
And finally, here’s a selection of notes that I wrote in the fancy black journal they gave us:
IA —> organization to screen
UX —> screen to user
IAs turn business into information
“IAs main talent is naming. If nothing else, we’re coiners.”
The basic skill of IA: naming: “Figuring out what everything is and how it relates to each other.”
Code is what nuts and bolts were 100 years ago
Surround yourself with artifacts to acclimate yourself to the identity you want, e.g. putting running shoes by the bed so it’s the last thing you see before you sleep and the first thing when you wake up
Starting with a screen is problematic because screens are downstream from structure
“My job is to consume ambiguity and shit clarity.” (Dan Brown quoting Ken Fast)
That last one is pretty great.
Next time I want to actually stay at (or near) the venue; taking the train back and forth from Chicago to the suburbs every day was tiring. That said, I’m glad I went and I hope I can go to more of these summits in the future. I was bit less talkative than I wish I had been, but nevertheless I managed to meet interesting, smart people. And hanging around such people, I hope, will help me too to become one of these interesting, smart IAs.