Contextual Design often is compared to marketing studies (typically surveys and focus groups) that collect random samples and hundreds of data points. And years ago, while testing for usability, people in the industry were not comfortable with test results from small numbers of users. However, after 15 years of collecting data, the industry has found through experience that small numbers add up to a detailed picture of work practice that supports design. And we’re not just looking at usability any more; we’re engaged in market characterization at the level of work practice.

Small Numbers Support Modeling

At InContext, we build lasting consolidated data with anywhere from 15-30 field interviews and from that can characterize markets of millions of people. We’ve seen consolidated data gathered from 8-10 people identify a large percentage of the key issues that eventually will be identified in the market.

For example, we routinely consolidate data after just 10 interviews and then grow the models and affinity from there. By doing the affinity early in the process, gaps revealed in the data can then guide further data collection. The affinity also represents the key areas and distinctions that can grow in detail as more data is added. Similarly, although additional data adds depth and detail to all the work models, the basic structure that is central to the project focus is identifiable early on. This allows the later interviews to be increasingly focused and ensures a comprehensive picture of the customer is created. This means that a little bit of data goes a long way. And because this data is also stable over time, it becomes a key corporate resource.

Focus on Similarities Leads to Successful Design

People get hung up on numbers because they tend to focus on variations, on people’s differences rather than their basic similarities. But focusing on people’s differences leads to products designed with infinite customizations. It eventually de-structures a system with ever more customizable and slightly different options, or requires costly upfront customizations at installation.

On the other hand, if people are so different from each other, how could any markets exist? Clearly there is enough similarity of practice between people for the idea of shrink-wrap software and off the shelf electronic products to make sense. Consolidation helps people see and find the structure in work practice, and this drives successful design.

We’re All Different, But We’re All Alike

Everybody looks different; humans have great variation. People are of different ethnic groups, child rearing practices and cultures. Everyone chooses different clothes, hobbies, careers and life styles. So at one level we are all unique. But at the same time we’re all alike. For example, from a clothing manufacturer’s point of view we are human beings with one head, two arms, two legs and so on. So we can buy clothes off the racks in department stores, and then tweak them — we hem the dress to make it fit. Structurally, the variation between people is small and the structure of our bodies is common. So while we tend to focus on our differences, given the product focus of clothing manufacturing we are the same. If we focused on variation, it would distract us from how to structure clothes. But focusing on variation is good to provide different looks and materials, which can then be repeated in themes. This is true of any product development.

There are Only So Many Ways to Do Work

Any product and system design is really a very narrow focus on the human experience. Within one kind of work, there is only so much variation possible. The roles we play, the intents and goals we have, and the way we do things is common. Variation, once you start looking for structural elements like roles instead of job titles, is small. We have found, for example, that there are only two to four strategies for any primary task in a work practice. And if those are the key strategies for that work practice, the question is not ‘which should we support?’ but ‘how do we support them all?’ We do not need frequency data to tell us that these strategies matter; these are the basic elements of the practice. Designs that target a certain domain of work are created for a certain set of people doing a fixed set of work tasks, situated in the larger work culture, using the same set of tools, trying to achieve the same kind of goals. Under these constrained conditions, the work practice will have similar patterns. If this weren’t the case, there would be custom systems for every single person. Instead people use the same software systems with some preferences and options. And after you have collected data from 3-6 people doing the same.

Geographical Differences are Mostly About Standards

Even geographical differences in work practice are small. They are mainly about law or law-like standards, not culture. For example, in Europe if somebody travels for their job, they are compensated using per diem rates. The actual rates and inclusions vary from country to country, but they are always paid per diem. In contrast, in the U.S. they may be compensated per diem or they may need to show their receipts. In any event, since there are only two ways to compensate work travel, there are only two possible interfaces—a receipts-driven and a per diem interface. Cultures are interesting because food, values, personalities and lifestyles are different. But from the narrow view of work practice, the variation that affects practice the most tends to be legislated by corporations or government. Other differences between cultures are best understood as variations in emphasis; some cultures emphasize one strategy or cultural value more than another, but most cultures have people in them that represent those values and strategies anyway. So it is possible to sample across cultures and check at the extent of difference before deciding how much global data is needed to characterize the market.

Contextual Data Uncovers Common Work Themes, Not Market Trends

Contextual data does not show trends. Contextual data cannot be counted, like a survey, because we are observing people in the field and letting each person’s activity direct what we see. We simply do not get the exact same data from each person, but we get a myriad of detail that can’t be found in surveys and focus groups. Under these conditions, frequency measures and trends have no meaning. Contextual data is trying to model or show the structure of the practice. Contextual Design gets its power by designing from an understanding of structure that doesn’t lose individual variations.

Start Small; Grow as You Go

If we had infinite time and resources we could satisfy our worry about complete coverage by extensive customer sampling. But we don’t. Luckily, because a design is targeted at a work or life practice, a small amount of data will let us model the important aspects of the practice and use it to drive design. So you can start small, grow the data and reuse it.