Wednesday, June 23, 2010

UI editing is a professional job

Once, while living in France, I decided it would be fun to go to a British pantomime put on at Christmas by a bunch of expats living in the area. Big mistake. The pantomime lasted for over three, excruciating hours. Like many amateur productions the wrong level of editing had been applied: the sets, costumes and dialog were great (apart from a few of the jokes), but none of the scenes had been cut.

The same sort of misguided editing happens with user interfaces all the time. The plague of any software product is the moment when non-experts in user interface design get to look at the UI. They instantly start copy-editing: they make small changes throughout the interface that locally are important (and might bring some consistency to the interface if carried through completely, or might wreak havoc instead). But they typically have neither the total overview of the product, nor the sheer will, to make the large changes necessary in a consistent fashion that will result in great UI.

This sort of ill-informed editing is a nightmare because...

1. Easy to see and change for personal preference

User interfaces are literally what people see when using your product, and so they are the first thing that people criticize. And a large part of their criticism is based on personal preference, not an informed opinion about what will make the product most successful.

For example, I like my terminal windows to look like this.

This is largely because when I started using terminals (after the line printer interface was replaced), they looked like that. So, it's a personal preference. But I'm a relatively old person when it comes to using computers. I wonder how many other people would want bright green on a black background. So, enforcing my choice as a default would likely be the wrong thing to do.

The same personal preference problem appears repeatedly when designing user interfaces, because it's hard for people to distinguish between something that's simply to their liking, and something that's likely to work for their users.

2. Informed only by single user experience

A related problem is that it's common for people to say 'well, that's how I would use the product'. Unless that person's goals are aligned with your customers', the way they are editing the interface is likely to be for their particular use (or flow) through the UI and not for the one your customers will actually follow.

The best way around this is to test with many users, by watching multiple users work through problems they need to solve with the user interface (even if the problems are simple things like "Add a person to the address book") you will be informed about their expectations and problems. A single user experience, and especially one from someone already familiar with the product, is likely to be off by some error factor.

3. No global view of the product itself

Also, it's easy to start editing part of a user interface without taking into account the effects of the edit on the overall product itself. Retaining consistency in both look and feel are vital, but even more important is maintaining consistency in how the application matches and guides user expectations.

Users will spend time exploring your application and it's as important to get the 'spelling' and 'syntax' of your application correct, as it is to get the 'grammar'. Badly designed applications often have different grammars in different places: at one point your application is flowing like an English sentence and the next you're speaking German. Maintaining this global view is very hard, and usually overlooked when people start editing a user interface.


One of the worst things you can hear when working on user interfaces is the classic disagreement breaker: let's make it an option. If you're brave enough the response should be: let's not. What happens is person X's feature is not liked by N% of the team (including the UX expert), but to appease person X the feature is added to Preferences (or worse Advanced Preferences). What you end up with is a cockpit user interface with every button under the sun. Except in the cockpit those buttons are actually needed!

Flickr user virtualpilot88

This, of course, is an area where Apple shines. Sure, my iPhone can't do X and Y, but it doesn't matter: I use my iPhone non-stop. I use all its functionality, and the functionality it has has been game changing for me.

The difficulty in building an "Apple" experience is that you need a single person who has the UX experience and the will power to drive through the design necessary, and, above all, to say "No, we're not doing that".

What you need is what my expat pantomime didn't have: a professional (UX) director.


chanux said...

I like green on black. And I like options.

Aditya said...

“... you need a single person who has the UX experience and the will power to drive through the design necessary, and, above all, to say "No, we're not doing that".”

While having a UI/X designer like that is definitely a good thing (I am one), many projects fail because that designer wasn't given the power or allowed to make the final decisions.

In a project's management hierarchy, the designer isn't high-up enough. It's the marketing and sales. When you put anything in front of marketing and sales, it's their way or the highway.

AR said...

Man I'm so glad that you get it.
I'm trying to break in to the UX industry because of the huge difference a good interface makes. It's all about the USER, not the designer, and many people forget that.
Be in touch I'd love to chat UX more with you. I've reposted this post to my Tumblr

Anonymous said...

You had me at green and black terminals.

Evan said...

@chanux No, you think you like options. You think you like options because what is provided as default is typically terrible. If the UX designer does his job right, then any option presented will only make things worse, and therefore will not be chosen, and therefore should not be included.

Add to this the dozens of other reasons for not including options, such as their inherent undiscoverability (hence the "or worse Advanced Preferences" line), that they will cause problems with people moving to a machine with the smallest of option differences, that they lead people to thinking about the tool rather than the content.

Options make you feel like you are in control, when really all you are doing is deciding between four terrible ways of doing something; and the very act of asking which you want is detrimental to your overall performance.

Anonymous said...

John, you clearly know what you're talking about! I have been involved in UX (since the days when we called it CHI) and I am constantly amazed at how few people get it. Even smart people I respect seem to go off into crazed misdirection when critiquing UI.

Thanks for talking sense!