Wednesday, January 6, 2010

Usability, after almost 30 years...

This morning, I spent over an hour to make a simple one night single room hotel reservation. The hotel's site, like many others, has a "one strike, then you are out" policy for payments: if you get the security code of your credit card or anything else wrong the first time, you can start the whole reservation process from scratch, including all reservation and address details. This policy, combined with my credit card provider's great idea of sending out a new credit card with a new security code without a clear transition between the two (thus, making the security code ambiguous), led to five attempts with two different credit cards until I was finally allowed to pay. Maybe the current "y2k+10" problem in parts of Europe helped. But such help was not needed, since the same kind of usability problems occur over and over again on web sites. DB, the German Railways, for example, give us two separate username/password combinations for two kinds of logins, transferring to the wrong entry form after failing once, which makes you fail twice and then your access is blocked.

Frustrated, I decided to relax with some scientific matters and went on the web site of "the world's premier venue for informatics" to check the materials I had uploaded for a seminar. I found them misplaced (under the general seminar materials rather than those of a participant), having no idea how this had happened. When I corrected the error, I forgot that one has to save a part of the form separately, despite a single "upload" command at the end of the page. This gave me the opportunity to write the whole abstract of my talk again (back button failed!)... Then, the site told me it would not accept slides as uploads, but accepted them anyway. Before quitting, I decided to read some slides from the general seminar materials - only to find that they crash my slide software (Keynote).

Everybody may have similar stories to tell and we are wasting hours of our lives per week. Here are some more experiences over the past month or so:
- A query was not successful and I have to modify it. Why, in so many cases, am I not given the opportunity to change the first query?
- Many travel and other service web sites offer quick search windows with default assumptions. A "more choices" (or similar) button allows for changes to any defaults. When I type in my data and then switch to more choices, why do I have to re-enter everything?
- Why are tabs so often not used to move between form fields? German Railways had the superb idea of using them for moving to another form!
- Why do the end date fields for date ranges rarely update themselves based on your start date choices? etc., etc., etc.

I know this has all been said, published, investigated, punished, and said again, many times before. Why is progress on usability so damn slow? Do we really have to live with these kinds of annoyances, almost 30 years after the first HCI conference and Xerox Star?!

Write to the web site owners (the management, not the techies) each time you get annoyed! I still hope that more and more companies will eventually realize that usability affects their income, directly!

More usability, and less "customer satisfaction surveys", PLEASE!!

5 comments:

  1. A footnote on this entry: a usability bug in blogspot had me first publish the entry in my other blog, erroneously...
    Werner

    ReplyDelete
  2. progress tax we are told ... Now let me go finish my Bamboo bicycle

    ReplyDelete
  3. The DB website can be a real pain in the neck sometimes. I keep on wondering why I have to click another button to get to know whether special prices are still available for a train?

    Another great example: the other day, I decided to give DHL's online prepay service a shot, which lets you pay and print the label for a package online, and the you can just drop it off at a Packstation. Of course you can only print the label once – bad luck if the stupid applet decides to send it to a printer that is currently not connected to your computer…

    ReplyDelete
  4. That is indeed a hot topic:
    10min ago I tried to put tickets for a concert in Düsseldorf (The Bar at Buena Vista) on www.fansale.de the reseller service of eventim. A christmas present we can't attend.
    After filling in a long form with descriptions, dates, 24digit-Ticket numbers and so on (15min)I had to choose a date, how long my offer should be online. I chose the day before the concert, so buyers could still get the tickets personally... After submitting I received this error message and a complete emtpy form!
    "Das Enddatum muss mindestens fünf Tage vor dem Veranstaltungsdatum liegen, damit eine rechtzeitige Zustellung gewährleistet werden kann." which means more or less: "The latest date for selling should be at least 5 days before the concert."
    Not only that the form was completely empty after this error, even the message is not clear: What does "mindestens" or "at least" mean here? In which direction should I count time here? ;)

    @Carsten: Same DHL-Problem happened to me, but at least their human-based hotline worked really good. I called, someone picked up immediatly and I received a new sticker within 2min...

    ReplyDelete
  5. I get annoyed at least as often as Werner and think of two reasons why the situation stays as it is:
    - people react to computers with a feeling that they should master it and if not, then it is their, and not the programers, fault. This has been the attitude since computer exists and saved the programmers, which are guilty of making software that is hardly usable.
    - the user interfaces are difficult to program in the usual languages and each is its own 'piece of art', which we have to learn. But who wants to learn how to deal with the DB or OeBB web page? who wants to learn yet another manuscript management system?
    A possible, perhaps the only, solution is in software tools which create usable and consistent GUIs from a higher level description. This would give at least consistency and avoid having to retype after an error.

    ReplyDelete