Co-worker: Lauren Elbaum
Date: Oct, 2017 - Dec, 2017

* Part of class assignment


Paper Author


Background Research,
User Interview,
User Observation,
Contextual Inquiry





The Internet has a plethora of information, so it is difficult for users to find the precise piece they seek. As an alternative method, search tags were invented to help people filter their search results. The main characteristic of tags is that is allows people to search by combinations from different hierarchies. It is more useful when the user has multifaceted criteria for search. E-commerce, where the users want to know diverse aspects of product, is the leading the field in using the tag search system. Although these systems have the potential to provide great value to users, many people give up searching before they reach their informational goal. For e-commerce, users' search interruption is critical, since it leads directly to a decrease in sales. Zappos, the biggest online shoe store, has put a lot of effort into preventing such interruption by offering a robust tag search system. Zappos provides a specified bunch of tags in 15 different categories, including size, price, styles of heel and toe, occasion, performance, etc. The specified categories give the users an opportunity to think about other features they are looking for, perhaps even features they had not thought to search before. Therefore, the users can easily narrow the search result and be close to what they are looking for. However, the users do not use all of the tags, and those search assistance of tags does not always lead to purchase. In this paper, we would like to discuss the user's search experience during online shopping and their usage behavior of tag search system; how many tags people use, when and why they decide to stop adding more tags, and how they behave after using tags, by analyzing Zappos.


Research Questions




Which search features enable users to find the shoe they want?



Which information helps them to decide which shoes to buy or put in the cart?



How efficient is using Zappos for finding a shoe?



Familiar, Closed Ended Task

You’re planning to go to friend’s birthday party and you want to purchase a new shoes for the occasion. Since you already decided on the clothes, you have a specific types of shoes you want to wear in your mind: Black, closed toe shoes with an ankle strap and a 1-inch heel shoes. You’re going to search for shoes to buy for the party. Which shoes will you purchase?


Familiar, Open Ended Task

The winter is coming, but you don’t have any boots for winter. So, you decide to purchase winter boots. You’re going to search for boots you want to buy. Which shoes will you


Unfamiliar, Closed Ended Task

Your 8-year-old niece’s birthday is next Saturday. She really wants pink sneakers. Since you really love your niece, you’ll spare no expense to get her the shoes she wants. You’re searching for shoes you want to buy for her. Which shoes will you purchase?


Unfamiliar, Open Ended Task

You started to work at a construction site. For the safety, you should wear a steel toe work boots. You’re going to search for shoes you want to buy. Which shoes will you purchase?


The experiment ran for three days, 11/29-12/01 in North Quadrangle and Munger Residence. Four participants completed the task before noon, and one participants completed at night, 5:30 pm. There was no time limit for the tasks, but each task took around 5-10 minutes, and the overall experiment time took 60-80 minutes. We used our two laptops, one was for the subjects to perform the tasks and the other was for us to monitor the subjects’ search behavior. Using the ‘screen share’ function in Skype, we could see how the subjects completed tasks. We used Chrome to access Zappos.

The experiment was done in the following order.
1. Explain the experiment and the tasks.
2. Ask the subjects to fill out the pre-search questionnaire.
3. Start the task, and monitor the subject's search behavior via Skype.
4. Write down subjects’ every search behavior into the observation form.
5. Allow the subjects to complete the tasks and tell us they are done.
6. Ask the subjects to fill out the post-search questionnaire about their task.
7. Repeat steps 3 to 6 for all four search tasks.
8. Analyze their search behavior, based on the observation form.
9. Interview the participants about their search behavior.

To minimize the order effect, the task order was randomized (Table1). We showed onlyone task at a time, and hid the other tasks. During each task, when the subjects found shoesthey liked, we asked the subjects to save as many pairs of shoes into the shopping cart as theywanted, and let them pick one final pair of shoes among them. The shopping cart items werescreen captured based on subject, and were deleted for the next experiment. We did not requirethe subjects to think aloud, but some subjects did think aloud voluntarily. Task start and endtime were decided by the subject’s announcement, and the task time was recorded bysmartphone. Each search behavior was written down in sequential order in the observationform.