Case Studies - Paper Prototyping: The Fast and Easy Way to Define and Refine User Interfaces - page 30

E-commerce Web Site: Priceline.com
In the mid 1990s Walker Digital, a small Connecticut think tank, was preparing to launch a new online
service. Their name isn't exactly a household word, but the Web site they launched became one:
Priceline.com. Back then, the site was several months away from its initial launch, and they were still
trying to figure out exactly how this unique [ 1 ] method of selling airline tickets would work. There were
some pretty significant risks, both with the technology and with people's willingness to do business in
this new manner. The product team wanted to mitigate the risks by finding out whether average people
understood this service and were willing to use it to purchase airline tickets.
For those who aren't familiar with how Priceline works, you tell them when and where you want to fly
(U.S.-originating flights only) and how much you're willing to bid for an airline ticket between those cities
on those dates. You also give them your credit card number. Priceline checks with several major
airlines and lets you know if any of them is willing to sell you the ticket you want for the price you
named. If so, bang—you've bought a ticket (usually a nonrefundable one), which is charged to your
credit card. The site targets travelers who have flexibility in their plans; the customer can't select the
airline or flight time but may be able to get a price that's lower than advertised fares.
Initially, the product team had been focusing on the technical challenges (which were considerable) and
also on selling the concept to the major airlines. But they recognized that without consumer
acceptance, their clever new business model for selling airline tickets wouldn't get off the ground, so
they decided to conduct some usability tests. I worked with several members of the product team to
create a paper prototype of the site. All of the pages were drawn by hand because the site was still its
conceptual stages. People had some ideas about how the interface might work, but just about
everything was still open for debate.
Two members of the product team participated in 2 days of test sessions by "playing computer." We
recruited seven people who fit the profile of Priceline's target market and asked them to work with the
paper prototype to perform various ticket-purchasing scenarios. We found several important issues,
including three showstoppers:
Asking for an email address to enter the site met with complete resistance.
The site (including the company behind it) had to establish its credibility before people were willing
to transact.
Three days was too long to wait for an answer.
After the first two usability tests (four users total), we'd already observed so many serious issues that
three of us spent a few hours that evening completely revising the prototype. The next day's tests
confirmed that most of our changes were improvements. Equally important, we found some things that
the team didn't need to worry about:
People didn't need to see probabilities of getting a ticket; they were going to do their own
homework on pricing.
People understood the concept of submitting a binding offer. They had some questions about it,
but overall they understood what the site was offering and how it might benefit them.
Following are the details of what we learned from the paper prototype tests.
Email Address
Some of the developers had already predicted that users wouldn't be willing to provide personal
information up front, but the team was still arguing about it. Some of the Marketing folks insisted on
asking people for their name and email address before letting them into the site. (Don't laugh.
Remember, this was 1997, when the word e-commerce wasn't even in the dictionary [ 2 ] and before it
was commonly understood that people would rather have a root canal than give out their email address
for marketing purposes.) The feedback was unanimous. All seven users said they would either leave
the site or enter bogus information. This evidence was enough to convince the proponents of
aggressive data collection that it was a Bad Idea, and it was dropped.