Design first
A new day and a new project. The first thing to do in a human-centered approach to designing a new product or service is to understand the user and business goals. You start setting up interviews with your client, reading about competitors, and begin to observe and interview the types of people who might end up using this new project.
But wait. What do you actually know at this point? What questions do you have? Are they well informed, distinct, and pointed? Or are they wide-ranging and fuzzy? I want to challenge the order in which we structure our approach to research in design, and to suggest how design can be used not only to create solutions but also to define the problem.
The current human-centered design process is essential: research, ideate, solution, test, and repeat. This is the wrong order. How can you research if you don't have a well-defined question?
How many times have you seen projects begin by observing users, talking to business and end-users, competitive analysis ... only to end up with a list of pretty general bullet points about what could be done a little better and some pros and cons. This is not a well-considered research process; this is a research process that started without a well-defined question.
So how can you change this? I would say start designing a solution to the problem from day one. Only by making something will you as a designer start to understand the problem both from a macro and micro perspective. Next, instead of testing your design, take a step back and ask yourself what questions do you have and what assumptions have you made? Now use the answers to these questions to formulate your research questions, and begin your research activities with them in mind. The next step would be to iterate on the design this time focusing on a solution. Finally, you would test this idea with end-users.
A bonus to this approach is that you will actually have things to show end-users, not just hypothetical ideas, and you can get a real reaction. The point of this initial design process is to understand the problem by actually doing the thing that designers are good at--designing!
While many designers have a skill for research, this is often more intuitive than formal. Based on many conversations with sociologists, anthropologists, and other social scientists over the years, I've seen how many designers’ research skills are at best naive, without a grounding in real research techniques. The research that designers do is often incomplete or misleading, especially when it is done on questions such as ‘how can we improve shopping experiences?’ or ‘how do we help improve health outcomes?’ This is not a knock on designers; why should they be good at this work? Most were never trained to conduct formal research methods or research design.
You may say I have described the lean UX approach of “design and test.” Rather, I would argue that that process is focused on finding a solution, rather than defining a question. Testing by its nature is about narrowing to a solution. The initial round of design in the approach I have outlined is not to define a solution, it is to define the problem.
By designing first, designers play to their strengths, creating a preliminary solution. By then doing substantive research into the questions raised by this initial design, they can look at more specific questions and find better outcomes. It allows the research process to be much more focused and useful, getting at core design questions more quickly.
