50% slower but still a better experience?
Recently a colleague of mine who owns an iPhone showed me how you set an alarm on the iPhone. He was really liking it and so did I at first sight. I pulled out my own Nokia phone and compared it with the iPhone. My feeling was that the iPhone solution ‘felt’ better but that the Nokia solution may actually be better in terms of usability and effectiveness. Nonetheless, the iPhone seems to be a nice compromise between usability and fun that makes the end-result still interesting or even better than the Nokia solution. So is it a good thing to sacrifice usability over fun? Is that what ‘User Experience’ is all about?
Here are the two interfaces:
Let’s examine this case a bit further…
The iPhone solution
The iPhone uses a rotating wheel solution that we all know from using them in old fashioned cypher locks. In order to set the time you need to position the 3 wheels. If you have selected 24 hour system you don’t get the AM/PM wheel though. I did some timing tests with three of my colleagues that owned an iPhone. I gave them 10 alarm times to set and they needed on average 10.2 seconds to set an alarm. That is including the ‘add alarm’ function. The main reason I could see for this long task time is that you have to do several strokes to get to the right number and then set the one you need, and then do that twice. If you only need to put your alarm half an hour later than normal, the time can be shortened and if you are lucky you only need to use one wheel.
The Nokia solution
The Nokia solution is more straightforward. You simply type in the time using the keypad. The time it will take you is always exactly the same about of time since it is always 4 key presses. Well, nearly the same because of the keyboard layout. I ran a similar test with three other colleagues who were familiar with the Nokia interface. After setting the same 10 alarm times as the iPhone users, my colleagues only needed 6.7 seconds on average per alarm time. That is 3.5 seconds faster than on an iPhone. Actually the difference is even larger because the confirmation message is displayed for about 1.5 seconds and doesn’t allow you to enter the next time. That means the task completion time on the iPhone is about 50% slower! And that is a lot….
Usability versus User Experience?
Mathematically speaking the Nokia solution is far superior in task time completing. Nonetheless, the iPhone solution provides a nice User Experience and seems totally satisfactory for iPhone users. Nobody is complaining about the iPhone solution and they are even praising it! So can a better experience still outweigh a 50% slower task completion time?
According to ‘official’ definitions of usability task completion is an important factor but it is certainly not the only factor. Learnability, memorability, the number of errors made and the subjective satisfaction are other elements of usability. So is the ’subjective satisfaction’ what brings us to the User Experience part of setting an alarm? It seems so and it is tempting to conclude that task completion times can sometimes be sacrificed in order to improve the overall user experience.
When is slower better for the User Experience?
The difference in task completion times in this example is huge but the task is very simple and you don’t need to do it many times a day, sometimes not even in weeks if you always want to wake up at the same time. My explanation is that in this case the following aspects make it ok to sacrifice task completion time:
- The task is very infrequent, perhaps not even once a day or once a week.
- The total task time is still relatively short and differences in short task completion times do not matter much if they are not frequent.
- Users tend to appreciate fun more than efficiency for short infrequent tasks.
This leads us to the even more interesting question: when does the balance tip over to task completion time again? Probably when the task is more frequent but how frequent should it be exactly? I don’t know. Perhaps somebody should do more testing?
March 13th, 2008 at 1:42 pm
It is true, satisfaction is subjective, although generalization and usability heuristics and principles allows us to model for ‘global users’. Again the frequency of performing the task in ‘Usability versus User Experience’ should be considered as key factor. People love chocolate but if they are forced to eat each day several times… the ’satisfaction’ level will drop with the increasing rate of ‘eating’ it! Of course let’s not talk about ‘tooth pain’ and dentist bills!