User research comes in two main forms, qualitative and quantitative. Qualitative data is highly contextual and describes attitudinal and behavioral information about people (assuming we are studying people and not bears). Quantitative data is measured, counted, and easily shown in classic formats like pie charts, grids, and bar charts. Business people and engineers love the quantitative because it is concrete and quickly obtained, especially if you have someone visualize it effectively for you.
Even though qualitative data is not concrete like quantitative data, it can be collected in such a way that does not feel fuzzy. It is important to not push off qualitative data because it seems hard to use and obtain. It is where all the design gold nuggets present themselves. I have never done this type of research with only validating my gut instincts. I am always surprised by things I learn doing this. This very experience tells me it must done to get all the good stuff, the competitive opportunities companies need to differentiate. If you are going to be valuable and set apart from the rest, you need this as your launching pad.
So how have I made this kind of information less fuzzy?
The answer comes in one word – Triangulation.
There are many methods of research. Understanding what methods are and how to choose the right ones for the goal in mind is key. So you notice that the word triangulation has the “tri” prefix, well that will tell you that I am about to say that you need 3 methods to work together in a complementary way to achieve the validity you need so you can really lean on it during design. It is important to understand the methods but in lieu of that, below is a handy quadrant that helps the lay person see how the different methods can be selected appropriately.
You can see that the various methods give you different kinds of data. It is important to choose the methods that compliment each other and avoid redundancy. The rationale for this triangulation approach is the thought that if three different methods reveal anything that trends (give you 3 points), you have something valid.
I chose an example of 3 methods to help illustrate this concept. Interviews, Email Surveys, and Web Analytics can make a complementary approach to understanding the user base. In this case, an application or website has to exist already. You cannot get Web Analytics without it. Interviews are a good choice if you have access to the user populations that represent the user base. Email Surveys are also helpful when you have a large email list you can send. Web Analytics measure behaviors and have a quantitative source with a large N. Interviews give you a small N but have rich qualitative data that deliver behavioral and attitudinal data. Email Surveys can potentially give you a large N if you get high response rates and can give you qualitative attitudinal data.
Experience will help give you the ability to design these studies effectively. When you know you are using them together, you can design them all in a way to give the you knowledge of whether the overlaps exist. As seen in the venn diagram below, when the overlaps present themselves, you know it is real and definitely NOT fuzzy.
When you know it is valid, it is the key to truth about the people you are serving with your product. That truth is gold.