Yesterday I stumbled across Improvements in Windows Explorer on the Building Windows 8 blog. I found this post interesting because while it is superficially impressive, I see some fundamental UX design mistakes that I suspect will result in the redesign falling short of expectations. Students of my UX Design Essentials course and readers of my blog should recognize them immediately.
OK, what did you find? Let’s proceed…
First, for the good part. Here’s what I like with the Windows Explorer redesign:
I’ll add to this list as I discover more things, but this is it so far.
Guess what? Windows Explorer is getting a ribbon. What a surprise! Didn’t see that one coming. To be a bit snarky, I think Improvements in Windows Explorer could be accurately retitled How we used data to justify a ribbon for a simple utility that probably doesn’t need one.
Don’t get me wrong, I love the ribbon and think it is a fantastic UI innovation. But from The Politics of Ribbons:
The ribbon has one significant strike against it: it is the heaviest commanding solution there is. A good UX design principle is to use the simplest, lightest weight solution that does the job well, so applying this principle suggests that—given its weight—a ribbon should be among last commanding choices, not the first.
In UX Design Essentials, I identify the classic UX design process mistakes. Here’s #3:
Falling in love with one (usually first) solution.
It’s a good bet that the ribbon was not only the first choice, but the only serious choice considered. Using a ribbon was likely a politically driven decision justified with data. Look at the data…we just had to use a ribbon!
A good design process is about making choices on behalf of target users to create a product that satisfies their goals. But you can’t make a good choice if you have only one option to choose from. Focusing on a single solution short-circuits the design process.
To convince me that this is a great design, I want to see an exploration of simpler, lighter weight alternatives and a demonstration that the ribbon approach is in fact better. And if you practice effective prototyping, you can’t argue that would take too much time and effort.
Windows Explorer is a utility, meaning that it is typically used to support tasks initiated in other programs. It’s what Alan Cooper describes as a transient application, not a sovereign one.
This is a significant detail because it affects how people use the program and how you should make design decisions. Here are some implications:
Bottom line is that users want utilities to perform tasks as quickly and simply as possible, then move their attention elsewhere. A complex utility, with lots of stuff you really don’t need, undermines this goal.
Doesn’t this feel UI way too complex for what should be a simple utility? Do you really want to wade through all this stuff to find a command? Do all these commands need to be visible on the screen all the time? Don’t you think there must be a simpler alternative that does the job better?
With over 200 commands, Windows Explorer looks like it was designed by a feature hoarder. If your top 10 of 200+ commands are 82% of usage, time to start cutting back.
BTW: Here’s a useful ribbon guideline that I think is appropriate:
The highlighter test is a simple, quick way to evaluate how effectively a UI is using screen space. Gather your top tasks, perform those tasks using your design, then highlight the UI elements that are potentially useful for those tasks. If almost everything is highlighted, you’ve done a great job. If not, you’ve got some work to do.
Here’s my highlighter test for the Windows Explorer ribbon:
Note that I highlighted only 3 of 19 visible ribbon commands. Crikey! The reason is that I use shortcut keys, context menus, or direct manipulation (usually drag and drop) for all the other commands. Yes, the data shows that Paste, Properties, and Copy are top commands, but that data also shows that 85% of usage comes from context menus and shortcut keys. I don’t expect the ribbon to change that much.
One benefit of traditional menu bars that deserves some appreciation is that you can fill them with every command you’ve got and nobody is going to care much. Not true with a ribbon.
The key to user-centered design is to identify your target users, understand them well, and make good decisions on their behalf. Improvements in Windows Explorer mentions power users quite often, but doesn’t mention typical users specifically. Looks like the Windows 8 team is targeting power users. Nothing wrong with that if it’s appropriate.
A value proposition states the reason your target users will want to buy and use your product—especially when compared to the alternatives. Their post suggests that power users are using after-market add-ons like TeraCopy, QTTabBar, DMEXBar, StExBar (which BTW the ribbon-based Windows Explorer won’t support anymore). Power users also use command line and Windows Power Shell.
The “when compared to the alternatives” part of the value proposition is a killer here. While their post makes a strong case that the redesigned Windows Explorer has many features power users will like, they didn’t make the case that power users are going to care. Will a Windows Power Shell user switch to Windows Explorer because of the ribbon and a few new commands? I doubt it. And if not, are power users really the right target?
Good value proposition work would have revealed this all-important problem.
But wait, there’s more. Let’s start to create a persona for power users with a few key characteristics:
BTW: I wouldn’t expect “Desire to respect Explorer’s heritage” to make this list. Your design priorities should clearly align with your target user.
Now assume that you are a power user and want to copy the file you just selected. Which approach are you more likely to take:
For this power user persona, I just don’t see #2 happening here. I know that the Windows team no longer uses personas. Perhaps they should.
At this point, you might be thinking: OK then, what would you do instead? Glad you asked!
Note how the Windows 8 team is focused on data, features, and feature discoverability. I mentioned in the introduction that using data is a much better way to make decisions than personal opinion. But there are several traps using data, the most important being that using data is often used as a substitute for thinking.
Data tells you “how much.” Knowing “how much” is important, but knowing “why” is far more important still. Raw data doesn’t tell you the users’ goals or what they are trying to do. It only tells you what they did and how often.
For example, the Windows data shows that Properties is the second most frequently used command. Why? Why do suppose users want properties so often? Is looking at properties a common task or goal? When you see a file, do you immediately think “I really want to see its properties”? No! The Properties command is a means to an end, but never an end it itself. The user is trying to perform some other task and checking Properties appears to be required in order to do it.
While a good design (that mindlessly follows data) would make the Properties command more prominent, a great design (that leads with data) would strive to eliminate the need for the command in the first place!
How do you do that? With scenarios! A scenario describes a specific target user trying to achieve a specific goal or task in a specific environment. Their post mentions Windows Explorer supports many different scenarios like viewing photos, playing videos, and playing music—then completely ignores that fact when justifying design decisions.
Remember: Windows Explorer is a utility that helps users manage files—a goal that most users rarely have!
A better approach is to design for the goals that users really do have. Design the best possible experience for viewing photos! The best experience for playing videos…for playing music…for protecting users data…for finding stuff. Supporting these scenarios should drive the design decisions, not raw data. So, for example, if you view a folder of photos, the top-level commands should be focused on the top photo tasks—and nothing else! And if done properly, excise tasks like displaying Properties go away because they are no longer needed.
Unlike task-based design, scenario-based design considers the user’s context—a great way to eliminate unnecessary complexity. In any given context, the user isn’t likely to need 200+ commands, but more like 5.
If you watch the video at the end of their post, Alex Simons show some tasks that look like scenarios. But those feel fake to me. Why? Remember that Windows Explorer is a utility and scenarios are tasks done in a specific environment. Real Windows Explorer scenarios should rarely start or end in program itself. These aren’t “end-to-end” scenarios but rather what I call “technology vignettes”—scenarios fragments artificially limited to the component you are working on, but not tasks that users really do. Designs that hold up well with “technology vignettes” routinely fall down when evaluated with real scenarios—especially for utilities. They aren’t “end to end” but “middle to middle.”
I really want things to work properly. I’ve found several basic features in Windows Explorer that just aren’t quite right yet, but in the interest of time I’ll limit myself to four:
These details weren’t mentioned in their post. Hopefully the Windows 8 team is addressing them, but I fear that focusing on the ribbon will distract from nailing the basics. I’ve seen that happen before.
If you do only one thing… When evaluating a design, take a step back and ask yourself what is driving your decisions. Technology? Features? Politics? Schedules, deadlines, and budgets? Data? While practical realities, these shouldn’t be the driver. Instead your design should be driven by providing value, satisfying user goals and tasks, delighting users by doing a few things exceptionally well. If you are using data, don’t let it become a substitute for thinking. Take the time to understand what the data is really telling you.
Want to learn more? Be sure to check out the UX Design Edge training courses. Every technique I’ve mentioned here is presented in UX Design Essentials.
Disagree? Leave a comment!