UI Design Newsletter – April/May, 2009

In this research study the question was: Will decisions about which camera to buy on a website be influenced by the way the cameras were sorted on the web page?

Specifically, the researchers Cai and Xu (2008, reference below) considered the perception of quality and price points, and whether presenting the product information in descending order, ascending order, or random order made a difference to the users' decision making processes.

The experiment had two phases. First, the researchers had 30 people rate the importance of different camera attributes. The question in this first phase was "What attributes of a camera do people consider when deciding which camera to buy?" The attributes that emerged as being most important were:

• # of megapixels
• strength of optical zoom
• screen resolution
• amount of zoom available digitally after the picture is taken

In the second phase of the experiment, the researchers designed a simulated e-commerce site selling cameras. They used the above attributes in describing each camera.

Sixty-two people shopped at the simulated site. (These were different people than the 30 who had rated camera attributes mentioned above). The shoppers were shown a picture of the camera, and information about the important attributes, as well as the price. Each person was asked to choose which camera they would buy.

The researchers varied how the cameras were sorted. The sorts were descending, ascending, and random, based on the above attributes. (Note, the cameras appeared at the simulated site in a certain order. The users were not able to re-sort in a different order).

In addition to seeing which camera people picked, the researchers also measured how important the various attributes were, and how important price was.

Here are the results

When the cameras were shown in descending order (highest quality at the top), then the importance of quality was ranked the highest. Price was not deemed that important.

If the cameras were shown in ascending or random order, then quality was not deemed to be that important, and price was a more important consideration.

The average price of the camera the participants picked when the cameras were shown in random order was $672. The average price when a descending quality order was used was $801, which is a 19% increase.

The authors comment that many websites allow users to choose the sort order or sort by price. This study shows that the best sort order, at least for selling high end items like cameras, might be in descending order by quality. They hypothesize that this might be related to the Fear of Loss principle. Recently there has been research on the idea that the human brain is averse to losing value or attributes.The researchers hypothesize that when the best quality cameras are shown at the top of the list, then the purchaser may be reluctant to give up those quality attributes as the list continues.

Summary

If you are selling a high end product online, sort by quality/attributes in descending order.

References

Cai, Shun and Xu, Yunjie (Calvin), Designing Product Lists for E-commerce: The Effects of Sorting on Consumer Decision Making, Intl. Journal of Human-Computer Interaction, 24(7), 700-721, 2008

I'm pleased to announce a special offer for CUAs only. I've succeeded in patenting an algorithm that can make you a usability hero. Interested in learning more?

Learn more here

Recently a student in one of our courses asked, "In our Agile shop, we always use quick informal workshops to prototype our UI designs. But now we have to tackle a really complex application, so we're thinking we need to run a larger, more formal workshop. Any guidance about how to run a large Agile UCD (user-centered design) workshop?"

Agile teams usually work in short "sprints" (a.k.a. "iterations"). They take a small piece of an application, and quickly and iteratively research, design, code, and release it, all in a short amount of time (usually weeks). The design activity often involves rapid lo-fidelity prototyping with users in a very brief and informal workshop.

For complex applications, longer workshops are needed, and a bit more structure to keep them on track. But the design can still be chunked into short sprints. For example, a series of one-week sprints could each address a chunk of the design, with each sprint using 1 day for preparation, 2-3 days for the workshop, and 1 day for follow-up. A coding sprint could follow these workshops. For example:

Workshop #1. Identify user stories (scenarios/tasks) for the new application.
a. Preparation: Review user research, and create strawman stories.
b. Workshop: Review and revise stories. Outline screen flows.
c. After workshop: Distribute user stories and flows for comments.

Workshop #2. Design top-level navigation and main application pages.
a. Preparation: Create strawman wireframe designs of navigation and of the main "work" pages.
b. Workshop: Review and revise designs.
c. After workshop: Distribute designs for comments.

Workshop #3. Design several key detailed pages.
a. Preparation: Create draft detailed page designs of key pages.
b. Workshop: Review and revise designs.
c. After workshop: Distribute designs for comments.

Here are some best practices for running a large workshop in an Agile environment:

1. Talk individually with workshop participants in advance. Get to know each of them. Explain the workshop's objectives, agenda, and process. Listen to their ideas and concerns.

2. At the beginning of the first workshop, review usability principles and design goals with participants.

3. Consider using an experienced facilitator to lead the workshops, keep everyone on track, and help participants come to closure. Make sure this person is familiar with your project, the user stories, the application, etc.

4. Consider having a "scribe" take notes and record final results. They should document only enough to support sharing the workshop results with your Agile teammates.

5. Prioritize the user stories and other requirements, and address only the top ones.

6. Use pre-approved, page-design templates to help workshop participants create consistent and usable designs faster, in a "cookie-cutter" manner. The participants will choose an appropriate template, e.g., Notebook Tabs, and then adapt it to their needs.

7. Ensure that a usability expert attends the workshop. This person will help the participants create usable designs.

8. Prepare to run "guerilla" usability tests during a break in the workshop. Discuss test results with workshop participants, and then iterate the design.

9. Schedule the workshop to run only 4-5 hours. Leave time for usability testing, doing further user research, and addressing outstanding issues.

10. Don't try to get the design perfect. Do "just enough" to support the subsequent sprints, which will refine the design.

By now you've likely heard about PET design, and if you haven't, you should.

In short, HFI has spent the last few years uncovering and expanding upon a body of research around persuasive design. We've done so to propel the "traditional" usability work we do to a new level. That's why when we launched PET design, our CEO, Eric Schaffer proclaimed "Usability is no longer enough."

Eric made that statement to draw attention to these new additions to our thinking. Some of you misunderstood the message and even took offense. But what Eric meant was, what I refer to as "classic" usability, is as important as ever. For example, as it relates to websites, understanding a site's users and the tasks they want to perform, and then optimizing the site to enable those tasks, is critical – to the user as well as the business. The reason why it is "no longer enough" however, is that there are additional business objectives that need to be considered when designing websites for your customers.

Let's take task optimization for granted for a moment, assuming that classic usability is table stakes. As a marketer, the conversation gets really interesting as we begin to discuss how to get my customers to not only do what they want to do on my site, but to get them to do what I want them to do.

I used to run websites in the financial services industry, so let me use an example many of you are familiar with – online banking. One of the most popular features is for customers to review account balances. Obviously, the easier that is to do, the better. The last thing a customer wants to do is to have to click 6 times to view a balance. From the marketers' perspective however, while viewing a balance, it would be advantageous to let the customer know that the bank has products and services that are relevant to him or her. It would also be ideal to encourage customers to sign up for online statements so that the bank can save on postage – and so on. Surely, the banks are all over this. But how persuasive are they in doing so? And how do their methods for engaging customers work in tandem with their usability efforts? (This video provides a terrific example of how PET design can improve the persuasiveness of an existing website.)

This is why we developed PET design. To make sure that each side of the customer-business equation is addressed – to increase customer satisfaction as well as increasing our clients' KPIs (key performance indicators.)

Take a moment to consider how PET design can enhance your website, intranet, or application. While you're at it, you can find out more about PET design here. You can also read some blogs and forum posts about PET design on HFI Connect.

I'd love to hear your feedback on PET design, our revised newsletter format, or any of the other content published by HFI. Feel free to contact me at mark.cohen@humanfactors.com.

See you next month!