He explains how two important aspects that are key to successful design teams:

• Direct exposure to user observation of team members for at least 2 hours in six weeks

Each team member has to be exposed directly to the users themselves. Teams that have dedicated user research professionals, who watch the users, then in turn, report the results through documents or videos, don’t deliver the same benefits. It’s from the direct exposure to the users that we see the improvements in the design.

• Involvement of strong influencers (execs, project managers, etc.) in direct observation of users

The tipping point came when we found teams where all these other folks were participating in the user research studies. No longer did they assert their own opinions of the design direction above what the research findings were telling the teams. Having the execs, stakeholders, and other non-design folks part of the exposure program produced a more user-focused process overall.

The question is – will we be willing to sacrifice invest in order to ensure this happens to make our designs more successful? It’s a common occurrence everywhere, and we as designers need to stand our ground.

Most of the research that’s been done around images is centered around the technology itself, rather than the behavior. Alas, there’s not much we know about why images are important to people apart from what we can already assume.

The study wasn’t meant to be exhaustive, but it does provide a framework for understanding why we use images on the Internet. Essentially, it comes down to the four categories below:

• Learning/Research
• Being Objects of Communication
• Connecting with Remote Experiences
• Supporting Other Goals

These four categories explain at a very high level why images are important to people, but this alone is not enough. Let’s take a deeper look:

Learning and Research

Images are highly valued for their visual qualities. Without words, you can gain so many insights just by looking at something. However, in my study, I noticed four very clear patterns of image use related to learning/research. And they were all driven by slightly different motivations:

Supporting an interest or hobby: Sometimes images are the best medium to showcase the things that we love so much, from sports to books to photos of interesting pigs.

To satisfy curiousity: Some users did just that – search for photos because they wanted to know what something looked like (what does pelt look like?).

For discovery of new facts: This is when we learn new things just by seeing them and making sense out of the visual information (e.g. learning that there are two types of meat mincers from product photos on Amazon).

For ideas: And this is when we want to get inspired by looking at things in lots of different ways (one participant spent 2 hours looking at birthday cakes on Flickr just to get ideas).

Objects of Communication

Images are not just used for learning. They are artifacts in and of themselves, and are used in many ways to support social interaction. In the study, participants not only used images as ways to communicate (e.g. smileys, showing a photo of something without using the words), but they were also used for social activities, like games (e.g. name-this-person).

We’re extremely natural at collaborating with each other using visual things, as opposed to just words alone. Images weren’t just used as talking pieces – sometimes, they were the message themselves! It was quite funny to see how people loved to use photos to replace words.

Maybe that’s why LOLcats is funny – it’s like having an invisible 3rd person giving an odd punchline.

Connecting to Remote Experiences

Our brains are hardwired to visual and auditory stimuli, so sometimes images (especially really large, high-resolution photos) act as windows to places and experiences we want to connect to -  our memories and our imaginations. The study showed how large photos were used to connect people to physical locations, past experiences, to connecting with friends and family, and even with personalities (e.g. celebrities – Michael Jackson died in about the same time the study was carried out).

Supporting other Goals

One of the most insightful things I’ve learnt is how people use images as sort of a swiss-army knife of the web. It’s used as a replacement for text, for getting past poor text search results, for sorting out navigation in a geographical space.

Alternative Answers: Images can be used to get around the limitations of text, language, and meaning. One participant was searching for an authentic Mexican dish, but Google search is awash with American renditions of the dish, which of course, wasn’t authentic. Google Images to the rescue – she found the recipe she wanted instantly because she knew what she was looking for.

Indexes: Have you ever tried searching for a music album that was a Japanese import, that was only distinguishable by the way the album cover looked? Some search results don’t reveal those things, and sometimes you just want the exact one you saw in the store. Thankfully, most e-commerce sites feature a screenshot of the product, and users do rely on that for efficiency and accuracy.

Maps and more: Sometimes, maps aren’t enough. Users are smart enough to look for visual cues like landmarks, photos of store entrances, the location of potential parking spaces (StreetView), and so on. Sometimes, if an image doesn’t provide enough information, look for another image.

More than meets the eye

I think we’ve only just begun to understand that behavior, while complex and sometimes idiosyncratic, doesn’t just happen without a reason. The purpose of this study isn’t about predicting behavior – it’s about understanding why something might be likely to happen, so that we can make the right design decisions to anticipate for those behaviors if they do happen.

It’s also about testing our own assumptions about how people use images on the Internet. The model presented here is only a starting point. I believe that as interactions become more complex and multi-modal, this may well change and evolve over time.

Note: here are the slides for the presentation I gave at NordiCHI 2010.

When I finally succeeded in recruiting a participant a few days ago, I was elated. Okay, so it was someone I met on the London IA Ning group, but nevermind – she had still some experience in a domain I was keen on doing some research in.

I based my interview method roughly on the interview method outlined in Holtblatt’s Rapid Contextual Design book, which I was also reading for this month’s UX book club. Unfortunately, the book describes an interview method that’s suited more for researching existing corporate processes, which didn’t quite fit the work I was doing – understanding a completely new domain for which there wasn’t an existing application for. Thus, I focused the interview strategy on understanding the subject area – shared ownership.

Here’s an outline of the questions I roughly had in my head (I didn’t prepare a script, because I wanted it to be more open and conversational):

• How did you get into car sharing?
• What motivated you to do it?
• Can you explain the process of what’s involved? (activities)
• What tools/devices/media did you use/consume in your tasks and activities?
• Can you tell me the various social aspects about the whole thing, if any?

Actually, this is a pretty bad list of questions. When I was going through the data, I realized that I had missed out some very important questions, like -

• were there any problems you experienced in any way? how? why?
• can you explain to me in detail (about a particular activity)?
• how often did you schedule the use of a car? why?
• asking questions about the user’s technical abilities and expectations
• a storyline of a typical scenario of a common activity
• what did you enjoy most about using the service? why?
• what activity was particularly easy/user-friendly to perform? why?

I could probably go on, but the thing was – I used this interview as an opportunity to fail, because I needed to learn from my failures and reveal my blind spots. Of course, I didn’t try my best to fail – I tried my best to do the interview, but I needed to know how I could improve.

This became more apparent when I went through and talked my colleague through the things I learned during the interview. This is where more than one person in a research process can be extremely useful. One of the key things we looked for were gaps in the research – things we failed to address during the interview or research sessions. This helps planning for subsequent interviews and research.

I learnt in this exercise that interviews in user research isn’t just about learning about users (which is like ethnography). In fact, it’s more about problem solving, except that the user may not even know he or she is experiencing a problem. When problems aren’t obvious, it can seem very hard to convince yourself that there are problems to solve, but problems always exist. There’s always something you can do to make a user’s life better.