September, 2006

Chapter Business

Our next meeting will be on October 19th at MIT, and November 14th is World Usability Day. We will be having a meeting on WUD in the next week and a half to solidity the activities/volunteers.

Chauncey would like to set a world record by doing 24 hours worth of 10 minute presentations. If we did 30-second talks, this might require all usability professionals in the northeast!

Job List
  • Staples is hiring an Information Architect for an 8-10 week contract.
  • RSA (now part of EMC) is looking for a senior level Interaction Designer.
  • Avaya is looking for a Senior Interaction Designer and an Interaction Design Manager.

Presentations

EQing as a quick method to get high level user requirements

Katie Juelich kicked off the evening by talking to us about using a brainstorming technique called "EQing". (You can learn more about EQing at " How to Come Up With Breakthrough Ideas ". Katie was able to use this technique to gather requirements from a group of power users (i.e. domain experts) at a conference, with very minimal time for preparation. There was an EQ card for each table, and stickies to represent where on the scale people at the table would rate requirements. Interestingly, the developer's main concern did not turn out to be a priority, and the top priority was "filled in" by users. There were several things that could have been done better (including more specifics of what the sliders meant) and having a usability person at each table to clarify instructions and listen to conversations, but overall it seemed a worthwhile exercise.

Q: Couldn't people just bump all the sliders to the top?

A: No, that wasn't allowed and people were generally good about it.

Q: Did each person at the table rank the requirements, or was it a group decision?

A: It was collaborative activity and everyone at the table had to agree. This led to some interesting discussions.

Using the Repertory Grid to Elicit User Experience Comparisons in the Customers Voice

Next Michael Hawley presented on using the Repertory Grid technique to learn about how customers perceived Staples as compared to their competitors (Wal-Mart and Target) for the "Back to School" user experience. They wanted some qualitative research to supplement their quantitative data. The RG technique is based on PCT (Personal Construct Theory) that supposes people organize their world based on attributes (constructs) that are opposites, and rate their experiences somewhere along the spectrum between "good" and "bad".

Using the technique, 20 sets of families first came up with as many attributes as possible (this is called "triading") and then rate them all on the spectrum. For example, the attributes might be "warm"/"cold", and so then they would rate where on the scale between "warm" and "cold" they felt a Staples store belonged. (Pictures of the stores were used to trigger people's memories of their experiences.) You could analyze the data a number of different ways, both qualitatively and quantitatively.

Q: So you were relying on people's memories of their experiences instead of actually showing them Back to School displays?

A: Yes, and that's OK, because they were testing "perception" not necessarily reality.

Q: Why did you test Target and Wal-Mart as competitors instead of Office Depot, etc.?

A: Because these are the toughest competitors for Staples in terms of the Back to School campaign. Back to School is like Christmas for them!

How to Run a Card Sorting Test with 200 People for $200

Next, Taylor Hayward showed us how he combined mturk.com with websort.net to get fast responses to an information architecture exercise. He described the combo as a "holy alliance for usability professionals" and easily explained why.

Mechanical Turk (mturk) is a part of Amazon.com that allows people to do perform specific tasks for a few cents each. So once you set up your card sorting exercise using websort.net, all you have to do is provide the link to the websort study in mturk (with the price that you will pay participants, something like $.10-$.20-yes, that's 10 to 20 CENTS!), and in a short time you have many responses. Taylor asked participants to provide their names, countries, and occupations, because he had a theory that participants would be college kids. : Once the study was completed, he downloaded and prettified the data from websort.net into a spreadsheet. Because they had several design options they were testing (i.e. it was a closed sort), they calculated an accuracy percentage for each option (and compared to their threshold value) to determine the best one.

Q: Can people do the sort over and over again?

A: Not easily. It's restricted to one person, but we suppose people could set up a whole other account (worth another $.20? Not likely.)

Q: Can you set up a screener?

A: Yes, mturk has a "qualifications" area.

Q: Did you get mostly students?

A: No. It seemed more like people at work wasting time. We got responses from lawyers, doctors, and other office professionals. 90-100% of the people were in the U.S.

Q: How many people? What was the total cost?

A: We got about 30-40 people per hour, and paid around $.10-$.20 each. So it was a very cost effective, large study.

Brainwriting: a fast, creative, egalitarian method for generating requirements

Kirsten Robinson taught us "brainwriting" by having us write down the names of our favorite usability books on small yellow cards, which were then passed to our right. (Well, the "passing" part was more like an "attempt" : ) Brainwriting is essentially brainstorming, but with the change of writing down ideas rather than expressing them verbally. Kirsten used this technique effectively several times. (Her slides are quite comprehensive so I won't repeat that information here.)

Q: Is brainwriting more effective than whiteboarding?

A: Yes. Whiteboarding requires someone to write down all the ideas (time consuming), and you lose the parallelism you get with brainwriting.

Q: Are there any limitations? It seems as though you might get more from having all people in the room hearing ideas rather than just the one to whom the card gets passed.

A: Yes, this is one possible limitation.

Q: How do you keep the session moving?

A: You can do it in a variety of ways: a specific amount of time, a specific number of things written, or just watch and see when people are stuck. You could also have all cards go in a pile at the center and have the next person randomly draw the card, or put ideas on stickies and then affinitize them later!

Note: Kirsten took our cards and will post the results from our own brainwriting session!

Brainstorming TIVO

Lynn Cherny wrapped up the evening by giving us a glimpse into the world she was in while at TiVo. She explained why TiVo was a great place to work because they "got" user experience. Everything in the organization was based on quality, and people from many different roles used their own product to find issues. Every Monday morning (after having done "homework") they would meet to discuss their findings in teams. TiVo designed with their customers in mind by doing surveys and noting the very different situation in which TiVo would be used (sitting on sofa, far back from TV, etc.). Of course they had to make their share of tradeoffs when working with big media companies and regulatory bodies, but they also paid careful attention to marketing their new product.

Q: What does TiVo stand for?

A: Lynn doesn't remember, but obviously the TV is capitalized for TV.