If It Isn’t Broken, Don’t Fix It

By Pabini Gabriel-Petit

Published: March 10, 2014

“Over the last few years, I’ve been noticing a disturbing trend in UX design: changes in the design of successful software user interfaces that actually degrade rather than enhance the user experience.”

Over the last few years, I’ve been noticing a disturbing trend in UX design: changes in the design of successful software user interfaces that actually degrade rather than enhance the user experience. This seems to happen for a variety of reasons—for example, because of

  • designers conforming slavishly to current design trends such as minimalism or flat design rather than focusing on meeting users’ needs
  • companies’ leaders wanting their UX designers to create “cool” rather than usable user interfaces
  • UX teams not doing usability testing or other user research that would validate a new design approach rather than being committed to doing user-centered design
  • designers disregarding the power of users’ kinesthetic memory when rethinking application layouts rather than giving it the respect that it warrants
  • designers succumbing to the egotistical desire to put their personal stamp on the design of software user interfaces rather than recognizing and preserving the value that products have long provided to users
  • designers making changes for the sake of change alone rather than strategically driving change to deliver greater value to users
  • companies engaging in feature wars with their competitors—causing their software user interfaces to become bloated with unnecessary features—rather than striving to differentiate their offerings in the marketplace
  • companies crafting user experiences that selfishly further their business goals rather than deriving business value by meeting users’ needs better
  • companies releasing software whose quality is not up to snuff because they’ve rushed it to market without adequate testing and debugging
“Any user who pays hundreds or even thousands of dollars for a software upgrade rightly expects that the new version will be better than the old one. This expectation is all too often disappointed.”

When one considers design changes that have contributed to the devolution of software user experiences over time, it’s often difficult to determine which of these destructive forces may have been at work. What is clear is that the software industry is doing a disservice to its customers by allowing this to occur. Any user who pays hundreds or even thousands of dollars for a software upgrade rightly expects that the new version will be better than the old one. This expectation is all too often disappointed.

While refreshing an application’s or operating system’s visual design may occasionally detract from its usability, interaction design changes are typically those that cause serious problems for users. And many of the biggest software development companies are guilty of introducing changes that degrade the usability of their software. They can afford to do better, but once a company has got a monopoly on a particular market niche, it’s often all too tempting to prioritize profit margins over serving users’ needs. Let’s take a look at a few ill-considered design changes that have negatively impacted users.

OS X: Preventing Long-Established Interactions

“Preventing long-established interactions by introducing design changes and removing essential features is bound to frustrate long-time users.”

Preventing long-established interactions by introducing design changes and removing essential features is bound to frustrate long-time users. The design changes that Apple has made in recent versions of OS X that have bugged me the most are their removing two features that I relied on heavily:

  • scroll arrows on scroll bars
  • the Save As command on the File menu

The Loss of Scroll Arrows

In July 2011, Apple released OS X version 10.7, otherwise known as Lion. I immediately came to rue my decision to upgrade because Apple had removed the scroll arrows from scroll bars, which I’d habitually used. Figure 1 shows scroll arrows in Snow Leopard—the last version of Mac OS X to include them; Figure 2, the scroll bar settings in System Preferences.

Figure 1—Scroll bars with scroll arrows in Snow Leopard

Scroll bars with scroll arrows in Snow Leopard

From the Apple Human Interface Guidelines published on June 9, 2008

Figure 2—System Preferences for scroll bars in Snow Leopard

System Preferences for scroll bars in Snow Leopard

It didn’t take long for a storm of protests to erupt from irate users regarding the lack scroll arrows. For the following reasons, this is a change that Apple should never have made.

  • Without scroll arrows, it’s not as easy to to scroll to a precise location in a document—a capability that many users need.
  • While using the up and down arrow keys to scroll vertically works pretty well when browsing the Web, it’s necessary for users to move their hands from the mouse or trackpad to the keyboard, making the use of the up and down arrow keys less convenient than scrolling only with the mouse or trackpad.
  • Using the arrow keys to scroll in no way suffices when users are working in applications in which they create long documents, where pressing an up or down arrow key laboriously moves the insertion point from line to line, making it take forever to actually scroll.
  • Making the assumption that all users would happily adopt gestural interactions on a trackpad was a mistake. Apple shouldn’t have pushed their design agenda so hard. They could easily have allowed users to always show scroll arrows, just as they did for the scroll bars themselves, as shown in Figure 3. Not surprisingly, I’ve chosen to always show scroll bars because I prefer to be able to see where I am in a document and how large a document is becoming as I’m working.
  • Scroll arrows had been a feature of scroll bars since the launch of the Macintosh in January 1984—that is, for 27 years—making this change nothing less than a betrayal of users who had long ago formed their usage habits.
  • Taking away essential features is carrying minimalism too far.

Figure 3—Mountain Lions scroll bar settings in General System Preferences

Mountain Lion’s General System Preferences, including scroll bar settings

There are several very long discussion threads regarding this topic on Apple Support Communities that Apple has completely ignored. I don’t think that’s indicative of a good customer-service ethic. As one commenter wrote, “It really looks like Apple no longer cares about people who are trying to use Macs to do serious work.”

I’ve adapted to this change by using two-finger scrolling on my trackpad, but I find this interaction to be less precise than the use of scroll arrows. And, while I love using gestural, touch interactions on my iPhone and iPad, I’m not generally a fan of using gestural interactions on a trackpad. My long-established habit of two-handed use of a trackpad—using my right hand to acquire a target, then my left hand to click—just doesn’t work when most of the gestural interactions are turned on. And because gestures don’t work as well on my Mac as on my other devices, I’m not motivated to change my habitual way of working. Changing deeply ingrained habits is hard. However, I do try out using gestures in each new version of Mac OS X that comes out, in the hope that they’ll improve enough to make me want to use them. (While Adobe uses standard Mac OS X scroll bars in most of its Creative Suite applications, I’m grateful that they’ve retained the custom scroll bars with scroll arrows in Dreamweaver, one of the applications that I use most.)

Confusion Around New File-Saving Paradigms

“While Auto Save was a welcome addition …, it was coupled with another controversial design snafu: removing the Save As command from the File menu and replacing it with Duplicate.”

While Auto Save was a welcome addition to the applications that came bundled with OS X Lion such as TextEdit and Preview, it was coupled with another controversial design change: changing the Save command to Save a Version, which let a user save a different version of a saved document at any time, and removing the Save As command from the File menu, replacing it with Duplicate. Rather than letting a user either save a file under a different name or save it in a different place or both as Save As had, Duplicate creates a copy of a file, placing it the same place as the original, then lets a user change its name, as shown in Figures 4 and 5.

Figure 4File menu in Lion’s TextEdit

File menu in Lion's TextEdit

Figure 5—Changing a duplicated file’s name

Changing a duplicated file's name

At least, in this case, Apple realized its mistake. Though the company does not appear to have learned well enough from prior mistakes how best to correct such mistakes. In OS X version 10.8, Mountain Lion, the File menu has again morphed into something unfamiliar. This time into a proliferation of commands—Duplicate, Rename, and Move To—for which Save As alone used to suffice, as shown in Figure 6. Both the Save and Rename commands display a Save As dialog box, shown in Figure 7. Thus, the behavior of Save is inconsistent with that of the traditional Save command. The Move To command displays the rather useless dialog box shown in Figure 8. If a user clicks Other at the very bottom of the Where drop-down list box, shown in Figure 9, a Save As dialog box appears.

Figure 6File menu in Mountain Lion’s TextEdit

File menu in Mountain Lion’s TextEdit

Figure 7—Save As dialog box in Mountain Lion’s TextEdit

Save As dialog box in Mountain Lion’s TextEdit

Figure 8—Move To dialog box

Move To dialog box

Figure 9Where drop-down list box in the Move To dialog box

Where drop-down list box in the Move To dialog box

Fortunately, most other major software developers didn’t jump on Apple’s design bandwagon and have retained the Mac’s traditional Save and Save As commands on the File menu, shown in Figure 10.

Figure 10—Traditional File menu

Traditional File menu

Sadly, Apple has made many more design missteps since Steve Jobs became ill, then resigned from his position as CEO of Apple on August 24, 2011. His vision and attention to detail are sorely missed.

Adobe Creative Suite 6: Damage to Dreamweaver

“In CS6, Adobe has again almost completely neglected to make any improvements to Dreamweaver. Most of the changes that I’ve noticed are actually bugs that got introduced in CS6….”

In making a too-long-overdue effort to bring more consistency to interactions in the applications in Creative Suite 6 for Mac, Adobe decided to bring Photoshop’s user interface into greater conformance with that of Illustrator, disrupting the user experience for people who are primarily Photoshop users. In my opinion, Adobe should have realized it was past the point of no return and left well enough alone.

On the other hand, in CS6, Adobe has again almost completely neglected to make any improvements to Dreamweaver. Most of the changes that I’ve noticed are actually bugs that got introduced in CS6—for example:

  • Dreamweaver no longer remembers the window layout that I prefer to use or applies it to all documents that I create or open, and its defaults are exactly what I would not want. So, every time I create a Web document or open one that I haven’t opened recently, I have to do either of the following:
    • Click the Layout icon on the window title bar to display the Layout menu, then click Code and Design to display a split-screen view. Click Split Vertically to deselect it, and finally, click Design View on Top to deselect it, as shown in Figure 11. That’s four clicks times how many ever documents I open—a lot of clicks that I didn’t have to make when Dreamweaver remembered the settings that I always use.
    • Alternatively, click View on the menu bar, then click Code and Design on the menu. Click View on the menu bar, then click Split Vertically to deselect it. Click View on the menu bar, then click Design View on Top to deselect it. That’s six clicks for every document!
  • It’s no longer possible to specify a title for a link that I create using the Hyperlink dialog box. Yes, the field is still there, but if I specify a title, the link doesn’t get created at all! So I have to write the title code manually for each link.
  • Dreamweaver no longer allows me to find and replace strings of extra spaces that get introduced into the HTML code in text that I’ve copied from Word, then pasted into a Web document. (In the Mac version of Dreamweaver, there is no import feature for Word documents as there is in the Windows version. And content from Word documents that were originally created on a PC don’t come across well. In addition to throwing in lots of extra spaces in the HTML code, it omits most paragraph tags and uses line breaks instead.) So, if I want my HTML code to be tidy, I have to select each instance manually, then cut it.

Figure 11—Dreamweaver’s Layout menu

Dreamweaver's Layout menu

How Adobe managed to introduce so many bugs into Dreamweaver without making any improvements to the application is a mystery to me.

Google Gone Wrong

“It’s very disconcerting to have the search box whipped away from it’s traditional location—centered on the page—to the upper-left corner of the page as soon as a user starts typing a search string.”

Some examples of design changes that Google has made recently that have not been well received by users include the following:

  • Google Search—It’s very disconcerting to have the search box whipped away from it’s traditional location—centered on the page—to the upper-left corner of the page as soon as a user starts typing a search string. And the search page loads much more slowly than it used to, introducing a very noticeable delay before a user can even start typing a search string. Between the unnecessary animation to move the search box and displaying the search results on the same page, this is now a much heavier page. While displaying the search results on the same page is a desirable feature, this could have been done with more finesse.
  • Google Maps—Until recently, Google Maps was a joy to use, and its user interface had evolved coherently throughout the addition of many powerful new features. But the most recent version of the Google Maps Web application combines seemingly arbitrary layout changes—very little is where it used to be or is in the same location both on the desktop and on an iPad in landscape view; minimalism that has run amok, resulting in a more difficult to decipher user interface; unstable behaviors; cryptic icons; and poor assumptions regarding what users would want to do next result in a user interface that no longer meets users’ needs as well as it once did. Do its designers really think users would generally want to immediately zoom in on their destination rather than get an overview of their route first? That considered, was it really a good idea to minimalize the zoom feature for the desktop Web application, making zooming a much more laborious process?

In Conclusion

“When redesigning software, the key to great design is first distinguishing between what’s working well for users and what isn’t … and delivering a coherent, holistic design solution.”

Examples of bad design changes that have degraded the usability of software user interfaces are far too numerous. Software companies used to do a much better job of serving their users’ needs than they do today. Of course, there’s room for improvement in just about any software user interface, and applications that evolve gracefully continually provide greater value to users. But, all too often, design changes lead to user frustration. For example, when users’ kinesthetic memory makes them reach for a control in a particular location on a screen and it’s not there anymore, this frustrates users and reduces their efficiency.

When redesigning software, the key to great design is first distinguishing between what’s working well for users and what isn’t; then preserving the good—an,d in some cases, just the familiar—remedying the problems, adding new features without detracting from existing features, and delivering a coherent, holistic design solution.


The second point about “cool” interfaces irks me the most. Unfortunately, shiny wins over usability with upper management.

Anyone that does these things is not a real UX Designer. I’ve seen all of these things happen when someone without the training, experience, or sensibilities gives themselves the title.

Some corporations, even large ones, don’t have a good job description for UX Designers and therefore don’t recognize the difference.

Still doesn’t explain or excuse Apple and Adobe though!

To JK Hudson

Thanks for contributing another reason why these things unfortunately happen: people who are not qualified to do UX design are responsible for UX design. And I suppose a reason that’s corollary to that one is that some companies aren’t willing to pay the price to hire highly skilled, very experienced UX professionals and choose to hire people with little experience instead. Inadequate experience would likely result in poor judgment in making design decisions, and it would be difficult to ensure forward progress if one didn’t know the history behind past design decisions.

You’ve hit the nail on the head—and without even mentioning Windows 8!

The article is interesting, although I do not fully agree and, in particular, not for scroll bars.

Having seen computer interfaces evolve in the last 30 years, I have also seen resistance people have to change.

In the case of scroll bars in OSX, it is a little more complex than stated. Apple is evolving the way we interact with elements on a screen. We are shifting from point and click to just touch.

If you operate a MacBook with a keyboard and a mouse, you are in the point-and-click universe, and in there, you badly need scroll bars in order to move a viewport over a document.

But if you operate it with a keyboard and a touchpad, suddenly you are much closer to operating an iPad than a Mac.

First you change the scrolling gesture to natural scroll, in order to move a document inside a viewport and, once you do this change, you suddenly realize you do not need a scroll bar anymore, just an indicator showing you where in the page you are.

New scroll indicators don’t work well, IMHO, with a mouse or with touchpad scroll simulating a mouse.

Once I shifted perspective from moving the viewport to moving the document, I was perfectly fine with scroll bars’ disappearance, which BTW, I always found as the ugliest and most intrusive components in a computer UI as they are there all the time, even when you do not need them. They perturb reading and appreciation for images and pages.

Hi Pabini—interesting article. I also have lived this kind of thing from the long-term user point of view. But as designers, I am sure many of us have also lived the razor’s edge between long-term user base and desired new adoptees, who just won’t put up with the old way.

At our house, my husband, an avid weather watcher, refused to upgrade his Web interface to a well-liked weather site, bookmarking and continuing to use the older one—until one day I showed him something on the new site that hadn’t existed on the one he loved. He contemplated that for a while, then switched.

Maybe sometimes we just need to have some time to get used to the changes. And maybe sometimes the company has just goofed!

Thanks for your comments Sean, Luca, and Alice!

Luca, I’m really surprised that it’s taken this long for someone to disagree with me, since I was being deliberately provocative. :-) I agree with a lot of what you’ve said, but I think you’ve simplified the issue in some respects as well. People interact with computers in their own ways—perhaps not as you’d expect. For example, I haven’t used a mouse for so long I can’t remember when I stopped. Maybe 20 years ago? I’ve been using either a trackpad or a Wacom. And I really prefer natural scroll. I think it would be disconcerting to go back and forth between iPad and iPhone and a Mac with natural scroll turned off. But I still miss the precision of scroll arrows. Responsiveness to gestural interactions on a Mac is still a bit too herky-jerky for me to be completely happy with them.

Luca and AliceI hope you and other readers appreciate my restraint in waiting so long to write about the issues that were the article’s main examples. Complaints about the removal of scroll arrows are still pouring in on the Apple Support site, so clearly, this isn’t one of those cases where Apple’s decision is going to sit well with everyone. And Apple need not have made a binary choice on this matter. It’s desktop software, so turning scroll arrows on or off could have been a preference. If users are happy using their current tools in the ways they prefer, why force them to change their habits? It’s much easier to accept a new paradigm like touch on a new device that’s optimized for touch. In most respects, Apple did a good job of providing users with choices about touch interactions in OS X. Just not in this one case that turned out to be hugely important to a lot of people.

Alice, I’m actually surprised that your husband was able to keep using an older version of a Web application. On the Web, that’s rarely possible. For UX designers working on Web applications, it’s a big decision to move forward to a radically different user experience. I certainly wouldn’t want to discourage anyone from doing that, but it is something to be approached with a certain amount of humility and caution, and the goal should be a significantly better user experience. When we’re innovating, those are the times when we really should do generative user research and usability testing.

This article is arguing for companies to simply optimise until they reach their local maximums. And then what—sit there and wait until they get superceded as the landscape of technology and their users shifts under them?

Moving on from a local maximum can mean arriving on the new path at a lower level than where you used to be—which may irritate people —but the potential gains should much greater.

UX designers should be striving to move to paths that allow higher and higher maximums, but this takes vision not curmudgeonly conservatism.

Hello Ed

That’s not what I’m arguing at all. I strongly believe that companies must disrupt their own products to move the market forward before another more innovative company does so and drives them out of business.

Sometimes, when users end up with something that doesn’t quite work, it’s because the vision hasn’t moved far enough from the status quo. For example, in the case of the MacBook Pro, why not offer a model that replaces the trackpad altogether with a touch device rather than creating an awkward mashup of two different types of interaction devices?

If we’re going to disrupt users’ work habits, their having to change their ways should be worth the effort. Large, established, wealthy companies like those I’ve mentioned can afford to do the usability testing to ensure that their incremental innovations are successful before they foist them on an unsuspecting public. When companies revert their changes, as Apple did with their save interactions, they’ve obviously recognized their error, but later than they should have.

Ed said, “UX designers should be striving to move to paths that allow higher and higher maximums, but this takes vision not curmudgeonly conservatism.”

This can be an excuse to not do research and just try something because someone at the company thinks it’s cool. And, this happens a lot. As long as changes are research- and reason-driven and put out in a strategic way, no problem. You are framing things in a black-and-white context. User experience is as gray as it gets. Usability and intuitive design will always matter even when people are interacting only with their thoughts.

Well stated, Tim. I totally agree with you. Both vision and validation are essential and always will be.

Great article…

A point that companies and designers need to keep in mind is the context of use. Some things may present better on a tablet versus a keyboard and mouse interaction.

But what we must do is be knowledgeable about how and where our users will be interacting with the products, sites, and services that are being deployed.

I have found many sites have become unusable if one is working on a traditional screen, keyboard, and mouse setup because they have been optimized for tablets and the touch interaction model. Google made me sad when they decided to keep stuff about options in the way of the map info.

Many users do not have time to invest in learning contrived interaction metaphors to get their work done. If one needs to sit at a screen and keyboard, designers and developers need to be able to sense where that user is and how they need their interaction presented.

Please remember the context of use for what the user expects when designing and developing.

Thanks for your comment, SM. Glad you enjoyed the article. You’ve made an excellent point about how important it is to consider a user’s context of use and optimize our designs accordingly. I agree that this becomes an especially sensitive issue in relation to applications that people use to accomplish their work. When people are working, their tools shouldn’t distract them.

What you’ve described is the essence of responsive design. Modern technology provides more information about users’ contexts of use than we’ve ever had before. We should leverage that information to the max!

Thanks so much for the article. I’m totally with you with the file browser dialog.

You mention that removing the scrolling arrows was a bad decision and that you liked the arrows. Is this your opinion or is there any research that objectively proves that it’s a worse approach?

Thanks, Adrian

Another example is the Windows Start button, which seems to be back in one of the ugliest ways possible. No matter that it was a result of an internal power struggle at Microsoft. It appears the company is in panic mode, trying to fix the UX mistakes it made earlier. This new UI seems to be a forced marriage between the old and the new interfaces.

The users who had gotten used to the new tiled UI are no going to get irritated.

Surely, the user evaluation has not been done well enough, and this would have serious consequences on the sale of tablets on Windows platforms.

There were definitely better UI solutions to bring back the Start button.

Join the Discussion

Asterisks (*) indicate required information.