Kevin Cheng  

Choosing to Give Choices

June 3rd, 2005 by Kevin Cheng :: see related comic

I like to equate just about everything to some video game experience. Remember the choose your own adventure books? You would get to a decision point and have to choose which path to take, and each choice led you down to a different paragraph/page number. The choices were limited but boy was it fun. Now look at the modern day role playing games (RPGs). You can do anything you want! You can walk anywhere in the world, just about whenever you want to! It’s incredible really, just how much world and monsters I can see and _not have any fun_.

More choices does not equal a better experience. In fact, quite often, it simply confuses the issue and unnecessarily complicates matters. Ever go out with a group of people and try to decide where to go for a meal? It’s always easier to just give them a choice of a few to choose from, isn’t it?

Continuing on last week’s Mac talk, I often cite Apple as a great example of taking the route of limiting choices. Consider the purchase of a computer. If I was looking at an Apple, I basically have two decisions to make. Do I want a desktop or a laptop? Do I want high end or low end? Between those, I have four core choices (G# desktops, iMac, Powerbook, iBook). Sure there’s some variations internally but not a lot.

Now look at PCs and the choices they afford you. What motherboard am I going to use? How much RAM? What type of RAM? What brand of RAM? Does that even work with the motherboard I chose? What about processor? Hard disk? Video card? Media readers? The list goes on and on. Sure, most people aren’t putting together their own systems but even if you went to the poster child for computer configurators - Dell - you’ll still get three pages of options before you can complete a purchase.

##Too Much
It takes a certain amount of guts to be able to dictate the choices for a user. In my mind, when you’re faced with a dilemma of how something should be done, “letting the user choose” is often a cop-out. Giving people the option to do _everything_ is like designing by [focus group][1]; you assume people actually know what they want.

Case in point:

Granted, MS Word’s audience is so broad and varied, it’s impossible to cater properly to all of them but how many actually use _any_ of these options? I think one could go so far as to set what menu they want the Save command under, what icon to use to represent it and what hotkey to assign to it. In essence, a user has the option of redesigning much of the interface. Heck, why hire designers at all?

##Too Little
Obviously, limiting choices is the better way to go then. Apple believes this so strongly in fact, that they take it to whole new levels:

Yes, that screenshot is courtesy of Tom “[I really really want to like my Mac][2]” Chi. In case it’s not clear, that’s the desktop background selection for OS X (10.3 I think). You can choose a whopping 10, OS X colour scheme friendly colours for your background. Want to have a BLACK background? Better open Photoshop and create a gif to tile. While I’m on the subject, I love the icon next to “Solid Colors” that gives the impression one would get to choose their colour from a colour wheel. What does Microsoft, purveyor of user choice, offer? Why, _16 million colours_ of course. Hm. Maybe limiting choice isn’t such a good thing after all.

##Choices Online
These patterns aren’t only showing up in desktop applications though. Look at [Yahoo! Movies][3] and how one rates a movie on that site. It’s a _13 point scale_. How does one differentiate between a C and a C+? A- and B+?! More importantly, are people more or less likely to bother giving a rating at all when the scales are that granular?

If you look at [Rotten Tomatoes][4], they aggregate all the movie critics they can find into one score. Each critic either likes a movie (fresh) or doesn’t (rotten). It doesn’t matter if they gave it 100% rating or 65% rating, it’s still just one single fresh vote on If someone had proposed such an idea prior to this site’s existence, I’m willing to bet a great number of people would feel a system that disregards all these ratings to be inadequate. Yet the ratings seem to have been fairly consistent indicators (but that of course is subjective).

##Choosing to Give Choices
As I’ve said, it takes guts to cut down the number of choices supplied to the users, but that doesn’t mean it’s always good to take away choice either. Sadly, much like so many other aspects of our field, the answer on when to give choice and how much choice “depends”. More often than not, however, designers are currently not even in the mindset to consider limiting choices and are too quick to hand over all the controls to the user so if nothing else, next time take a step back and think really long and hard about whether your audience really _needs_ that choice.

[1]: “Losing Focus”
[2]: “Beep Beep Beep”
[3]: “Yahoo! Movies”
[4]: “Rotten Tomatoes”

27 Responses to “Choosing to Give Choices”
Usabilist wrote:

Nice point about the choice of Macs. I “grew up” on PCs and I’ve got so accustomed with the choice of every component of the PC so I did not realize that advantage of Macs until I read your atricle!

spinner8 wrote:

True, true, true. And, I wouldn’t have guessed that I’d hear about “Choose your own adventure” books today. Thanks for a great point about technology and a nostalgia blast at the same time.

Who wrote:

Limiting choices would definitely speed up the decision process, kind of like set menu’s for lunch. Saves time and headaches alright but I think it takes away the fun.

I recalled that Apple almost died when it changed from its open for all Apple II to Macs. While PCs thrived because of the plentiful choices available in hardware configuration and application software.

mattui wrote:

Oh, I live this problem every day. I work for a software company that’s run by Engineers. Engineers don’t listen to Designers unless you’re saying yes to “Can I get rounded corners on that box?”. I’m just dreading the day that it’s “Can I get rounded corners on that box with a customizable radius set by the user?”

Wade wrote:

To me the difference between Mac and PC software development cultures often boils down to this: When faced with a decision whether or not to add a control/parameter to an interface, PC-oriented software companies ask “Why not?”, while Mac-oriented software companies ask “Why?”

Jacques Troux wrote:

At a previous company, the UI design was handled by people with little UI design experience who didn’t want to admit that they had little experience. The worst upshot of that was whenever the group couldn’t agree on how feature X should work, the immediate “solution” was “show a message and ask the user what to do.”
At one point I actually counted the calls to MessageBox and there were on the order of 200 in the whole code base (which was pretty big, but still… 200).

Phlash wrote:

Good points here - but not much mention of the ‘Advanced…’ button, which I sometimes loathe (because I can’t find the options I need without clicking it), and mostly love because it makes the user choice easier to start with, without actually forcing them to not have particular features.

A similar idea I’ve seen used in the freeware Video LAN client (amongst others) is a toggle between normal and expert mode for all options, again trying to reduce the clutter and make choice easier for most people.

Finally there is the Mozilla approach - use a text editor to hack the .js/.xml/.whatever configuration files directly.. ummn….

PS: I usually refer to ‘Advanced..’ buttons as ‘Scare me..’ buttons, since that’s the typical effect when clicking one :)

X wrote:

Murphy’s Law of Interface Simplification:
No matter what your sampling method, the first choice you remove will be the most missed.

Stu wrote:

There’s a good book about this called Paradox of Choice by Barry Schwartz. It’s a good, quick read. One example he uses that I find interesting is a jelly display at a store. When you could sample any one of 24 jellies something like 5% of people would buy a jelly, but when you could only sample six 30% of people would buy a jar of jelly.

Once different options have been added it’s hard to take them out because, to everyone that is not a designer, it looks like you’re removing value from the program.

Apple’s background color selection is horrible. I’ve had to make that black background in Photoshop. Being presented with a choice of 16 colors might be more intimidating that choosing your own color. If you’re picking your own color you probably already have on in mind, so you’re just trying to get close enough to that color. That’s much easier than picking a favorite.

Michael Zuschlag wrote:

Option proliferation is an insidious disease because it’s easy to see the benefit of adding an option while the cost (just one more checkbox in a dialog) seems insignificant. The result is you end up with a huge confusing mass of options that confounds the users’ attempts to actually *use* any of them, defeating the purpose of the having the options in the first place. As a first approximation of simplification, I’d advise having only options that have been shown in usability testing to have a performance advantage for one set of users and a performance detriment to another. That the users (or developers) cannot agree on what they like is not a good reason for having an option. On the other hand, I cannot deny that users get satisfaction out of customizing the superficial aesthetics too.
Maybe it’s a good idea to keep functional options separate from aesthetic options. Maybe for the aesthetics you can have just 8 to 12 “canned” themes to please most users, with the capacity to override each element (perhaps directly on the thumbnail representation) for those more fussy ones after a very specific look.
I wonder how successfully an application can guess the likely options a user might want from the context and make these options easy to see (e.g., by putting them in a task pane or adaptable menu). For example, if a user changes the printer in the Print dialog, the application presents the option to make the new printer the default.
BTW, it’s true that MS Word gives you the power to put the Save menu item wherever you want, but actually I regard this to be one of its better executed customization features: it’s relatively easy to find and use whenever you need it and doesn’t get in the way of other things you’re trying to do (unlike the Options dialog, where you can search indefinitely among other irrelevant options for an option that may or may not be there). No, I haven’t moved my Save menu item. Ironically, I used Word’s menu/toolbar customization capability to simplify the UI –I removed nearly half the default toolbar items whose clutter were slowing me down from getting to the toolbar items I really needed.

sassyfrassylassy wrote:

Ah the quandary of choice. Something that I have pondered long before viewing life as a UI person. One of my favorite songs of all time sums it up for me. “There are too many products to choose from. We can never use them all.” It also has the line “Psychology gave me a reason to give up hope.” - This is Logical Thinking by LMNOP

Joshua Kaufman wrote:

Unless the business rules dictate that a system must have a certain number of choices, my general suggestion is that fewer choices is almost always better. The observation by Wade is spot on, and the Mac direction is much of the reason why their products are more straightforward.

I never noticed that you couldn’t select whatever color you wanted on the Desktop & Screensaver preferences in 10.3, but they have corrected this in 10.4. There’s now a color chooser just below the name of the color, which brings up the OS X standard color dialog.

Dave wrote:

I think an important distinction in this discussion would be to talk about the different needs in designing enterprise software vs. non-enterprise software. Enterprise software requires a lot of options b/c every environment the software is used in is a completely new context, that is ungeneralizable and so the options are required to reduce required customizations, b/c they cost more money than configurations.

In the PC vs. Mac world, I believe this is why Mac’s aren’t used in the Enterprise nearly as much as they could be. They limit their options to the end-user so that IT staff don’t feel they can make the environment the way they want it/need it to be.

Darren b wrote:

Isn’t this yet another good reason for iterative development of your product? How many times have you asked an audience what they wanted, built a prototype and then through evaluation, found out what they really needed! (Agree with X though - reality can bite!)

Carolyn Snyder wrote:

I’ll second Stu’s recommendation of Barry Schwartz’s book The Paradox of Choice - this book was an ephiphany for me. One key point that Schwartz repeatedly makes (backed by research) is how people become paralyzed by too many choices, and make a poor (or no) choice as a result. Example: When companies offer too many investment choices in their 401(k) retirement plans, participation drops.

Part of the problem is that users/customers believe - and say - that more choices are better, without realizing there’s a threshold where their utility will start to decline. It’s the designer’s job to seek this elusive threshold, hold the line there, and help educate people how to get the most out of what they’ve been given. The ethical dilemma (which Schwartz also discusses) is that influencing people’s decisions can seem like Big Brotherism. But the set of choices we give people is already affecting their efficiency and happiness, so we’d better stand ready to shoulder that burden. As Kevin says, it takes guts to do this.

Designers and developers contribute to feature creep too, often innocently by thinking of something that wouldn’t be hard to implement and offering this idea as a potential feature. As a usability consultant, one of my favorite questions is, “What problem are you trying to solve?” Unless there’s a clear answer, it’s usually best to say no. A variation, especially when the idea has been suggested by a customer, is to ask them to explain how it will help them - there may be an alternate solution already available.

RoskeHF wrote:

One approach to the choice paradox is to attempt an analysis based on how people generally THINK or approach the task of making a choice. In the case of colors a “natural” integrated user task model may guide us to a design model that might have the user select one from among 10 basics and then allow to lighten/darken/ or mix that with some recommended color values for that one choice.
For “INTEREST” or “PLAY” value I would also add a choice of “YOU DECIDE” which randomly selects and/or changes the display based on the weather forecast or stockmarket!
Why keep it dull, boring, or predicatble when it comes to color? As designers we must help people to play this means we should “SELL” any feature - user motivation will be raised by “freshness” and an “interestingness” value of the selection feature.

Jaime Guerrero wrote:

I often feel the “too many decisions” issue when dining out– large menus with too many delicious choices. The appeal of prix-fix menus is that the chef has reduced the choices to one or two that are guaranteed to be fabulous. Being a vegetarian actually helps a lot because you can mechanically whittle down the choices to a subset, and choose from them. This becomes a crutch, however, at vegetarian restaurants where all options are game.

This phenomenon also occurs at typical large American supermarkets. I read recently that studies have shown that most shoppers buy the exact same items every visit, despite the enormous variety available. This also explains the insane manipulation and marketing of consumer products (store display positioning, shelf space issues, brand-extensions) that was described in the book The Total Package.

Sometimes I think folks in the old Soviet Union had it easier, as far as consumer decision-making. When there is only one choice of bread or car, that’s what you’ll be buying! :-)

In USSR, the lack of consumer good choices was source of unrest. Some would argue that the abundance of choice in the West is a salve, providing choices and the illusion of power over the lives of people who are really powerless in the more significant aspects of their lives (work, poverty, car-dependent suburbia, etc.

I wonder how many software manufacturers add Options as a salve to placate the desire for user control over software that users are otherwise stuck with and powerless to replace to really control? Consider certain entrenched but really suboptimal operating systems, hardware platforms, or styles of interaction (the WIMP paradigm, the Save File To Disk storage model, etc.)? :-)

José Vale wrote:

I work in a software company in the comercial division and belive in this. Its very hard to decide what we need to have or no. If we do one thing and we didnt give the end user the option to decide to use or not the product is bad even if the end user problably dont use the option or only need to configure thar option for one time only. We have a term where that says ” We are guilt to have a dog and guilt for dont have a dog”. The customer wants to have the feeleing of controling and the products ara made for the customers. So is very dificult to decide that.

Bob Salmon wrote:

Seeing everything as being like a computer game? Well, it’s different to seeing everything as making love to a beautiful woman, I suppose.

One cause of this is an aging product. You create version 1 and some people are happy and others not with area X. So you could change X from red to blue, but some people need red to fit in with their business processes but others equally need it to be blue. Do you cheese off the reds by forcing them to be blue, or appear to ignore the blues by keeping it red?

Or, do you give them an option - if so, which is the default? (My opening gambit in this is the existing behaviour i.e. red in this case.) Or, do you create an extension point and force all users to create a red plug-in or blue plug-in (which you can maybe sweeten by supplying a standard red plug-in).

Maybe it isn’t a choice in X being red or blue, but it’s a new feature that’s quite big and maybe requires configuration etc. The existing customers who don’t want the new feature don’t want to set up configuration data they’ll never use, skip screens/tabs etc. they’ll never use etc, so a turn-on-this-feature option can help existing customers.

I work on a product where just an upgrade takes customers at least 6 months and a non-trivial amount of money, so the poor existing customers already have a hard enough time, particularly as we like them to keep up to date so we don’t need to maintain and support loads of releases.

(In case you’re wondering, it’s not because it’s badly designed or written, it’s because it’s very large. Just imagine what it would be like to upgrade to a new version of SAP in a factory. There’s more to life than single-user PC apps, don’t you know ;-) )

Sherrod Segraves wrote:

As a software developer, I’ve had to learn the hard way that it’s best to present a simple interface with only a few standard choices.

This even applies when making code modules that will be used by other programmers. I think that’s why the Microsoft Visual Studio / C# combination have been so successful. Both the programming language and the programming environment have been designed with usability in mind.

On the other hand, choices should not be completely removed. Instead, hide them on an advanced screen or something.

I hate fighting apps that don’t give me the options I want. I’m the sort of user who has used many of the options in the “too much” example of a dialog box. I’d be angry if it was completely taken away.

In the past, I’ve even resorted to manually editing registry entries or other application settings to get what I wanted.

I’ll admit, I’m a technical user. Designing with me as your primary audience would be a mistake.

But there are a lot of technical users like me out there. Give us our advanced options screen so we won’t curse you behind your back.

Ian Stalvies wrote:

I think a few people have cracked it here with the “Advanced” option - 80% of people are covered with the simplified menu/choices/etc, but the remaining 20% still have somewhere to go.

Apply the concept elsewhere … a well designed search facility on a website, a “talk to an operator” facility on an automated phone system, and so on. Should keep MOST people happy?

Tracy W wrote:

The first time I read about the 6/24 items of jam experiment was well before the book came out. That evening I was meeting a friend and arranged to meet her by Starbucks. While waiting for her I counted Starbucks menu options and came up with 35, not counting such general options as using soy milk (there were 3 such options, thus leading to a total of 35^3 combinations). This made me extremely doubtful about how relevant the jam experiment is to the real world - if people did really prefer fewer choices then I would expect Starbucks to offer fewer.

But, look at Starbucks more closely. Its 35 different items are broken down into sub-groups of about 6 each. And there was a distinction between coffee and non-coffee drinks. The menu looked like it was set-up to lead the customer along the lines of: do you want coffee or non-coffee? If coffee, what of these six types? And of that type, which variant? And would you like an extra with that? So rather than 35 items, the customer faced 4 choices from 6-7 or less items. Some of which, especially the coffee/non-coffee decision, would have been decided probably before the customer entered the store.

Similarly at the supermarket. I don’t choose between however many different types of cereal it is, my choices are pre-narrowed to among the different types of toasted museli. There isn’t enough range in that grouping - I can’t find one with both fruit and without yoghurt-coated raisins.

And as for investment funds - the last superannuation package I considered offered only 1 index fund! Which probably explained why its fees were so high.

So in summary, I think the negative consequences of choice are over-rated. Companies often structure choices so people work their way down a list, and people often structure decisions themselves to a much smaller group than the psychologist sees at the supermarket. It would have been interesting to see the results of the jam experiment if the experimentors had tried grouping their jams in some way, e.g. berry jams, marmalades, etc.

John wrote:


Upside: I can set things up EXACTLY the way I like them.

Downside: I fumble about like an idiot on any computer except my own.

Robert Speirs wrote:

How about Subway and other sandwich shops? They ask you what you want, you say, “Turkey Southwest pepper” or whatever. Then the sandwicher takes up the roll and says, “And what would you like on that?” And you have to say, “Turkey and Southwest Sauce and hot peppers”, when you really want to say, “Whatever will make it like the Southwest Turkey sandwich you’ve got up on the menu, moron!” So they pretend to give you a few discrete choices, but then feel compelled to give you unlimited variability. They’re the sandwich experts. Shouldn’t they know what will taste good?

Lu wrote:

Gee wiz…while we are still on the topic of food and games. Anybody out there
play yet?
Learn about food aid + help UN WFP work towards a world without hunger. go on…

“Interface Your Fears” y’all ;)

Andre Wijono wrote:

First off, Great Article!

I’m currently faced with a problem similar to this.
A client of mine has about 40 different services to offer.

They wanted this website to be more of an “enticing” page. Many Call to Actions. The purpose is for the user to contact or to call them.

From the Audience Perspective, these will be targeted toward the housewives and for people who have limited education in their life, and we will provide the skills for them to have to get a new career quickly and easily.

Now, from having those information, how much information should we display?

I agree with with Jaime Guerrero about his analogy with the Restaurant Menu. I feel the same way. When I am in a new restaurant, displayed with pages and pages of varieties, I’m left with confusion on what to decide.

With a smaller selection, I will be more at ease with my selection. With the varieties of selections, I’m left with puzzlement, confusion, and afraid if I chose the wrong dish.

This is not what kind of feel that I want the user to have.

From the 40+courses, I divided them into three sections. On the homepage, Each sections will display 4 selections, with a dropdown that will display more. On the top of the selection, there is the Title, and I made the Title clickable that will bring the user to ALL of the courses under that current section.

From the beginning, I wanted to make this a simple homepage. Client started to add more Call to Actions, Testimonials, News, Articles, … it made the page more text heavy.

And now, one of the Client’s co-worker think it is too heavy. Wanted a simpler design (like I previously had) - but that means, removing all the sections they decided to have.

So now, I’m currently deciding.. THE BALANCE of the content. a Content, where it will not confuse the end user, but Enough to entice and make them click through, and explore the pages.


Great Article!

Robby Slaughter wrote:

The Word “Options” dialog is like a great archaelogical find—it tells a story about the people who created it. Everything you see on there is something that enough people on the design team violently felt should not be mandatory (or violently telegraphed from actual users) to warrant a way to shut it off.

Over time, this dialog has become ridiculously complex as a result. So what’s the solution? Instead of (or at least, in addition to) grouping optional features together on one dialog, create individual interfaces that handle each one *in context*.

This is much harder, both to visualize and to implement, so I’ll show where it’s already in place. Go into a modern version of Word and type “1) blah blah” followed by the return key. Word will autonumber for you (unless you’ve turned it off) and display a little lightning bolt icon right next to the number. Click on this and you will have the choice to Undo the autonumbering, turn it off permanently, or go to the AutoCorrect Options dialog.

If every “feature” in Word had a similar model, I think the application would have just as many choices, but would be much easier to control.

Leave a Reply

OK/Cancel is a comic strip collaboration co-written and co-illustrated by Kevin Cheng and Tom Chi. Our subject matter focuses on interfaces, good and bad and the people behind the industry of building interfaces - usability specialists, interaction designers, human-computer interaction (HCI) experts, industrial designers, etc. (Who Links Here) ?