Designing the VisualsSpeak ImageSet

We developed a number of prototype VisualsSpeak ImageSets and watched how people used them for two years. The development team used them in sessions and reported back to us. There were a number of other people we recruited to try them in various settings.

We recorded the categories people used, wrote down feedback, interviewed users and dug deep into the literature. Through all this I wrote enough to graduate with a Masters degree (Post Secondary, Adult & Continuing Education with a focus in Training & Development) and a couple of training certificates.

We talked to potential users about what challenges they needed to solve. If you came within speaking, phone, or email distance, we asked questions and listened to what you said. Every detail of the tool was considered, reconsidered, and tested out. We learned that no matter what we did, it wasn’t going to be perfect, so we had to figure out what was must have, nice to have, and might someday have.

Our goal was to create a tool that worked for as many people as possible. To find the sweet spot that made it usable in many situations. We started taking photographs based on our research. Tom did the bulk of the photography, with Christine shooting particular images she was looking for (OK, she grabbed the camera out of Tom’s hand when she got impatient). Antimo organized a weekend photoshoot where we had a whole series of models come in, and we had Tom and Antimo shooting the same situations. Every photographer sees differently. Peggy and Lew Pusch provided a number of the international shots. We took over 20,000 photographs to select the 200 we used in the ImageSet.

Photograph © Christine Martell

The first thing we looked at was image content. We had been tracking the categories people used, so we had some idea of the frequency of use. We knew we needed diverse images. We had tested with people from around the world but we still had to compromises to make. With only 200 images, we have tried to maximize the diversity of the images, but they still carry the bias of the cultural context of being produced by us.

What we see is influenced by who we are. It’s not possible to be fully neutral no matter how hard we try. The second factor we looked at is visual language elements. In testing, we discovered there are visual language elements that individuals will use over and over. No matter what question we ask, people will tend to create images with similar visual elements. We each have a visual language voice, most of us just aren’t aware of it. You see it with visual artists, often referred to as their style. So the images needed to contain a full range of those elements. Colors, lines, shapes, textures.

We knew from talking to people that having different size images was important. A large image says something different than a small one, and often would be used to convey relative importance. There are also technical requirements for commercial printing we had to meet. Some images didn’t have the quality we needed to print big. It was like working on a giant puzzle. We printed out several hundred test images, and used them in sessions. Watching, listening and refining more.