Using data

There’s a belief close to dogma in UX design and product management: that examining data will reveal something that will improve the UX or the product in some way. Some even refuse to do any design until they have done “research” or have access to web analytics, customer feedback, or some such.

There is nothing wrong with having data in the design process. But you would not fit a speedometer in a car and expect that car to go faster or run better as a result of the data it gave you. You need a reason for having more data about the car’s speed. So if you suspect your poor judgement of speed is responsible for you getting speeding fines, then getting a speedometer might help that problem.

So it is with data. What would you do better (or differently) if you knew something like click-through rates were high or low, or what keywords drive referrals? This data on its own won’t tell you what to do, but matching the data with an hypothesis can. And an important implication of the fact that data on its own can’t tell you why something is happening, is that the hypothesis can only come first.

Yet I see designers often reluctant to state hypotheses (which I would say is not the same as opinion) due to their belief that they need to have as much data as possible before they can do any design.

Breaking out of that vicious cycle is the key to progress.