Extracting Meaning From Research
When designing a product, it’s not enough to simply gather information and observations before plunging into design: there is a crucial point in the process where the designer must leave behind what they’ve learned and make a dizzying leap to imagining what the solution should be. Since that solution is rarely, if ever, arrived at deductively, it can be a source of anxiety for designers and clients both. But designers can devise research to identify real, pressing problems that need solving before moving on to designing a product. Clear problem definition is a key part of successful design. In this article, we outline some techniques and tips about conducting research to help identify which problems are the right ones to solve. Diligent research informs an authentic, but not literal, description of whom the product is for and what their needs are. It helps the designer answer the questions: Who is this really for? What motivates them? What need can this product address?
MODELING THE PROBLEM
Drawing conclusions from research requires modeling the domain and problem space. Being able to visualize the context of the product, its users, and the dynamics at play in the domain will make it possible to prod and experiment with complex information and extract meaning from it. Think of these models as similar to the game Tetris: collected information and observations are plugged into the appropriate space in the model. Sometimes, a piece doesn’t fit, but don’t despair. Pieces with an odd fit often provide important information about points of stress or pain that the product can address.
Recently, a client came to us with a problem. In order to support a specific business practice they wanted to know:
• How is this business practice defined?
• How many businesses really engage in it?
• Which industries should be targeted?
• What problems do those companies need to solve in order to engage in this practice?
We decided to tackle this problem by gathering data using three different methods: telephone surveys, face-to-face interviews, and a literature review. Telephone surveys included both closed-ended and open-ended questions, asked in the same order each time. This gave us quantitative information about issues of scale and frequency, as well as qualitative information about individual variances. Surveys are a great way to learn how a group of people operates, but don’t mistake learning about a group of people with unearthing what motivates an individual or how he or she can be served. Ultimately, understanding an individual’s motivations and designing with them in mind will make a product more successful.
Face-to-face interviews were open-ended and designed to be more free-form than the surveys. We interviewed both vision-keepers within the client organization as well as those who work in industries related to the business practice.
Interviews can be conducted either in groups or one-on-one. Interviews help determine:
• Who could the product help? What are their needs?
• Who else is involved while using the product? Just before and after?
• How does the end-customer think about the objects, data, or stuff that the product involves?
Group interviews are more helpful when designing to support an existing business or other process. Look for areas where friction exists, or where communication seems to break down. These are problems to target in your design.
Whichever you choose, keep the line of questioning open and informal. Brainstorm a list of what you want to know with some sample questions in front of you, but feel free to follow the interviewee’s line of thinking rather than strictly following a script. And always try to conduct interviews in the person’s day-to-day surroundings, this will give you another level of data to draw on.
Literature review: We were able to tap into a wide array of resources using subscription databases, reviewing relevant web sites, and by reading a couple of books. You should read what your client reads to gain the vocabulary of the domain. Magazines, white papers, and secondary research will round out your background.
LOOKING FOR PATTERNS
Both during and after conducting research, we used two techniques known as coding and highlighting to extract meaning from our data.
Coding, Categorizing, and Bucketing
With the telephone surveys, we went through a formal process of grouping the responses into different “buckets” according to theme. “Bucketing” is an iterative process, where a team of designers group responses to survey questions into categories that make sense. Disagreements about categorization spawn further refinement of the patterns. After the responses were grouped together, the team agreed upon an appropriate category description. These categorized responses were then totaled within and across industries. Through this process we were able to identify industry-specific trends and cross-industry trends. An excellent side product of coding is quantifiable data about qualitative information, which can bolster your argument.
While interviewing customers and stakeholders, we continually checked for common threads, problem areas and other themes in people’s responses. Highlighters and Post-it notes are effective tools to help find these patterns in interview notes. We used colors to label different aspects of individual interviews and then scanned for commonalities and differences. This too, is iterative: it means reviewing responses once to ferret out categories and preliminary patterns, and subsequent reviews to group the responses. When these patterns became predictable and repetitive, we knew we had collected enough information to extract the story from the data. At that point we were ready to use our research to define the business problem clearly and inform our “leap” toward envisioning the solution.
There are some other techniques that you can use to improve your findings from research:
• Be a skeptical innocent: It’s important to watch for contradictions between what people say they do and what they actually do. See if you can determine the cause of the disconnect.
• Don’t discard outliers: Divergences can say as much as similarities. Examine them for clues about future trends and problem areas.
• Clarify assumptions: People use words differently. Ask yourself whether two people really are saying the same thing; don’t just assume so.
• Know your constraints: Often designers are given fairly short timeframes for gathering and analyzing research concerning a domain. Spend time identifying what your research hasn’t told you.
Using these techniques to extract patterns from research helps you concentrate on problem areas and describe them accurately. By organizing people’s varied responses, you can deduce where a product can most help alleviate pain for the end-customer. This also helps when, having made that leap into designing a product, you begin evaluating your solutions to make sure they are effectively meeting the needs of those who will use them.
Originally published in 2001, when Gretchen and I were working at Cooper but still incredibly relevant…am I right? http://www.idsa-sf.org/inca_archive/2001.issue1/inca.2001.issue1.pdf