Scott Adams: The "Less" Feature

There's a delicate balance between "simple" and "too simple."

I've considered buying a cell phone that only acts as a phone. No text messages, just some buttons and a "send" key. But some people only text. So I need a phone that can text. And I use the phone's clock instead of a watch. Oh, one more thing, I also use the alarm clock when I nap away from home. And I use the address list quite a bit, because I've only memorized 4 phone numbers.

Unfortunately for me, it would be stupid to design this phone! It reached a tipping point where it makes more sense to use a real mobile OS. Then they can add apps and settings and make it appeal to more people. These demands are very inelastic. I need an alarm, and others need a tip calculator, and others need to take notes, and some people need a calendar.

The real irony is that I would love to select my phone features à la carte past the point of just adding and deleting apps, but that defeats the original problem - too much choice.

Adams, as usual, doesn't let facts and his own hypocrisy get in the way of a catchy blog post.

His choosing an airline ticket took a long time precisely because he WASN'T satisfied with "less". If he really wanted less information or fewer options and choices, he could have just used Orbitz without checking JetBlue and the United site, Or he could have just picked a single carrier and taken whatever flight they had available. But he wanted more -- he wanted to know about every possible detail and every available carrier. If he wants to know who to blame for the long time it took him to pick his ticket, he need only look in the mirror.

Coincidentally, I used Orbitz for the first time just two days ago after an unpleasant experience with a human travel agent who I felt was trying to push me into options and prices I did not want. By comparison, Orbitz was much more efficient and offered a good balance between complexity and simplicity IMHO.

Get a travel agent and never worry about travel again. Fixed overhead to not worry about the problem.

I still think the most-complained-about reduction in benefits when I worked at MSFT was when they changed over from a corporate travel agency to this hideous book-yourself model. The agency used to also give you the options for your personal travel, if you asked nicely.

Of course, it was just a rank and file issue. Once I was high enough in the org to have an administrative assistant, I stopped having to book my own travel again. Then, I understood why none of the people who travel frequently complained when the agency was removed. After all, who cares if it takes your developers three hours on the company clock to book their travel? owait...

This is why you should only be an optimizer in few, very important areas, and should be a sufficer in the rest: You get paralyzed by this sort of choice.

Get a cheap enough flight, get a good enough understanding of the one feature on the watch you care for (and write it down if possible) and arrange a movie night with your friends without caring if they'd like a slightly different variation just a little better (I mean, hell, they are always able to arrange it if they would).

he could not only pay someone to make this problem go away, he could probably also earn more money in the time he's not researching this than it would cost to have someone else do it.

in fact i find it difficult to believe he doesn't already have a personal assistant that could handle things like this.

The movie example reminded me of this: http://www.youtube.com/watch?v=jKF1u-TOXuE

From a high level it's easy to look at a device or a system and say "All these options are making this really hard. This thing should do less." But less is just a direction, and it's only useful to go that way until you can't do something you need to do.

The iPad isn't a success because it just does less than everything else. It succeeds because it does what it's supposed to do incredibly well. Users have certain expectations when they pick up an iPad and the iPad generally lets the user do what they expect and is intuitive in the process.

In the case of travel websites it often seems that no one has taken the time to consider what users will actually do on the site, instead just lazily giving users every possible option. For example, if a booking site has found me a $300 nonstop ticket it's probably not necessary to show me the 3-stop, $1700 option. Yet they all do because nobody has taken the time to filter down the results to the practical options (the user could create a filter, but that's just another option to deal with). I would much rather see just the 5 flight options that I'm likely to buy instead of 30 options with 5 potential winners buried in there somewhere.

If you're building a tool, the user will have to give some input. The key is to give the user what they want with the minimum possible amount of information.

There has been a lot of discussion about the "paradox of choice" recently, and I think there's something useful in the concept, as it applies to an individual consumer attempting to find the best solution for his own particular needs.

But the producer designs their product offerings to appeal to the widest possible market. There may not be much variation within a particular customer's preferences, but there is a great deal of variation between the preferences of different customers.

Eliminating choices might make things simpler for some customers, but it will also shrink your market by alienating those customers whose preferences differ from your new standard offering. The less choice you offer, the less mass-market and more niche-focused your product becomes.

Overall, that means that there are two possible outcomes: (a) if there are high barriers to entry, many customers will have to settle for a less-than-ideal product, or (b) if there are low barriers to entry, new competitors will emerge to appeal to the market's gaps, eventually leading to a similar level of complexity in product choice.

The question here isn't whether a high level of product choice is a good thing; not offering choice is only a possibility for monopolists.

The real question is one of usability: how do you offer a large number of options in a clean and intuitive way? The solution lies in UI design, not marketing strategy.

Choosing a netbook feels very similar. You'd think that choosing a model (T410 if you're curious) gets you there. Nope. Even with the relatively small number of configurable parameters (CPU, graphic, RAM, HD size and type) they managed to make it a confusing mess. There are codes like NT7EUPB, NT7ETPB2C, NT7EXPB2 and no definitive source that could answer a simple(?) question: "What are the options and how much do they cost?"

Toll Brothers researched the impact of choice on their profit margins. What they found was that given more choices, customers were not only less happy but also cost more to please. There were greater logistical requirements to get part A to place B, time spent selling people on home feature C, and then additional needs for people with skill X to install said items.

The amount of options is ok if there exists some easy way to handle them (and to hide the options away if you don't care that much about details).

For example, if there would be a service which would take care about all those traveling stuff (best/cheapest flight, way to/from the airport, best time) and print all the information I really need to know, like what Scott has written down for himself, that would be a better solution instead of just restricting all the further options. Not sure though if we will ever see such a service. Most attempts to do so failed because they miss too much information (mostly about the local travel possibilities and information).

The problem is that the (current) notion of a "feature" lends itself to poor design. If it's a "feature", then there needs to be an entry on the drop-down menu, or an icon, or a name for the particular new thing that you can do.

What Apple seems to do right is to avoid thinking in terms of "features" and instead (strictly) in terms of use cases: tools should ideally be sufficiently versatile -- and simple -- that they can be applied to do many things without each of those things having a button. Hammers, e.g., have little in the way of "features", but are still pretty darn capable.

The less is more mantra may not be desirable for everyone.

From one point of view (say, a seller's) the solution of "reduced number of options" may be a good way to influence a quick and (potentially satisfying) decision (like Dan Ariely also says here - http://bigthink.com/ideas/20749.)

But I as a consumer would definitely want to explore all the options before making a decision. One good fix here is to just understand that one can't have everything and learning to prioritize the needs.

This is the exact problem that Joel talks about (http://www.joelonsoftware.com/items/2010/08/19.html) [HN discussion: http://news.ycombinator.com/item?id=1618011].

Like the iPhone mute button example, Scott's watch doesn't necessarily need to be simpler, it needs to be easier to use (i.e., elegant).

Here's a solution to the travel problem:

Let the user input their hourly income. And the dates they want to be there, and their actual home and destination addresses. Then search nearby airports and departure dates based on total cost-of-trip. Time spent driving, flying, layovers, departing early, returning late, all ring up at their hourly income with the cost of tickets. Use a maps API for driving times to/from airports. Wouldn't that be better?

Totally agree. Less is a feature, and can be exploited too.

Too much choice is a bad thing. See http://en.wikipedia.org/wiki/Paradox_of_choice

Relating to software, I often see too much choice with regards to pricing and features.

More features do not make a better product, just as much as more pricing schemes/packages/tiers will not lead to more sales.

The problem in this case is decision making: over-optimizing flight options even though the gain is outweighed by the time spent doing so. Should we really simplify everything because options might trigger someones OCD to optimize even if those options are really useful to some people? Let there be options but not bloat, and people can use what they need.

There's an app for that - HipMunk might solve this problem (A-B flight route planning).

The benefit of many options is that the user can do the simplification himself.

The problem isn't too much choice. If the service were able to say, "Here is an all-around good choice. You could save a maximum of 13% of the price of the trip by shoving your schedule around, etc."--summarize the choices and evaluate them the way a competent professional human would do it--then Adams could choose quickly and rest easy knowing he got a fairly good deal. This is a hard thing to make a computer do--maybe it's an easy problem and Orbitz are just idiots, but I doubt it. Perhaps they could do a moderately good job with a moderate amount of work.

Point is, the problem is that the computer finds a bunch of choices and doesn't know which one is an all-around good default choice, what differences are important (and thus should be presented to the user) and which ones are irrelevant (and should be summarized or ignored). This can manifest itself as a symptom in various ways. The computer could just spit out all possible choices, which is apparently what it does right now; then the user sees "too many choices". The computer could pick 15 choices (the first 15 cheapest, or alphabetically first, or the ones with indexes equal to integers from a random number generator) and only display them; then users would probably see "a few really bad choices, and I know there are others, but the damn computer pretends they don't exist!" Would users be happier with that? I think leaving the choices there is the most graceful way to fail.

For the other things--the watch, the movie theaters--the problem is that they're trying to cater to people with different desires, with the same product.

The watch--they could make a watch that tracked running distance and displayed the time and did nothing else, and a separate watch that let you record laps (and displayed the average time!) but had no GPS, plus a third watch with all features and a fourth watch that only displayed the time. This would require some diverse manufacturing equipment (dunno how much of a pain that would be; perhaps not much), 4x the shelf space at stores that sell watches (this would definitely be a pain), 4x as many entries in watch catalogs, and so on. Maybe this would be worth it, if enough people would be willing to "pay 50% more" for it. But, of course, this means a bunch more choices as well, just at a different time.

They could also just make one kind of watch and throw all their other customers out the window. This is like the "solution" of throwing out all but 15 arbitrary choices on Orbitz. I doubt if Adams would be pleased if they chose to make only the "it records laps, but no GPS and no tracking running distance" model. Maybe they could run a service: "You tell us what features you want on your watch, and we'll send you a watch that has only those features (and a correspondingly easy interface to them)." Dunno, has anyone tried that? Anyway, the problem is how to provide good products (where simplicity of use is a good feature) that are useful to many people who want different features on the product, when the product's nature (having only a few buttons) makes it hard to provide simplicity along with all the other features.

"Too many choices" is the symptom. The problem is something else, and providing too many choices is what someone thought was the best way to punt on that problem. If you ask for fewer choices, either you're asking for some nontrivial work--to solve the underlying problem in a way no one's thought of before--or you're asking for a different way to punt, likely a worse way.

oh teh irony of an article about "less" a being wall of text / tldr.