July 26th, 2005
A database containing airports and their corresponding addresses. Not too difficult a thing to implement. Probably not something you would need to worry about the usability of, right?
Well, not quite. At least, not for Hertz.com.
Here’s the scenario: An experienced traveler comes to the site, hoping to reserve a car for pickup at the Seattle/Tacoma airport. They enter SEA into the “Renting City or Airport Code” field, which is exactly the right code for SeaTac airport.
To their surprise (and ours, I might add), our traveler was presented with a pop-up declaring that they had to choose which SeaTac airport they wanted, with three interesting choices: the conventional Seattle, Washington US, the interesting Seattle, Western Australia US, and the amazing Seattle, Wanganui US! Who knew you could have so many choices to find Seattle?
This is obviously a problem with Hertz’s database. A similar request for SFO (San Francisco) or BOS (Boston) only results in one choice of airport. (Though, it is a little strange that they also do a text match on the letters, so the pop-up for SFO also asks about Chelmsford, Massachusetts, Gosford, South Wales, and Amersfoot, Netherlands. I wonder how many people search for a city by typing in 3 letters from the middle of the name.)
This interesting database integrity problem requires the user to pause, figure out what is up, and confidently choose the right city. At least, that’s how it seems. Turns out, if they choose the SeaTac in Wanganui, they still get the real one. Whew!
But, it does highlight an important issue: are we, as usability evaluators, supposed to report on problems we find with the database? If so, this creates a whole new dimension to usability: data integrity.
As we can see, data integrity issues impact the user experience. But, it was just a coincidence that our user entered SEA into the form. Virtually any other city would have produced “expected” results, thereby causing us to miss the problem entirely. If we need to find these problems, how do we create test scenarios to ensure we uncover them?
I’d be curious to know if you are running into problems like this and, if you are, how you’ve enhanced your usability process to tackle this in a thorough and thoughtful way.Tweet