Data Driven Design v. Intuition: When to Use Each
Data v. Intuition is a debate that continues to intensify in every industry that relies on each for answers.
Until recently, intuition was widely accepted as a great source of decision-making, but the arrival of big data has changed the way we look at our world. The old tropes of “experience” and “creativity” are no longer sufficient for making smart business decisions. Intuition will never be eliminated (nor should it), but knowing when to use intuition over data is the key to making the right decisions.
When should intuition be used in place of data-based decisions? Never. Intuition should only be used to ask the questions that lead to data analysis or to guide the data process of planning, collecting, and analyzing.
For example, you have two designs for a landing page. You like one much better than the other and would like to move forward with your preferred choice. The problem is your preference, and there is no way to know which design the majority of users will prefer without testing. In this case, it is very possible that the data validates your intuition, but you don’t know for sure without testing.
Further, data analysis should not be viewed as a waste of time even when it matches up with your intuition. Instances like this don’t mean much of anything beyond being lucky. Every situation is different, and experience can make your intuition more focused, but it will never be a sure thing. At most, intuition should be used as a hypothesis to test.
Unfortunately, intuition often plays a larger role than simply leading to the hypothesis. Problems often arise when the data returns answers that differ from one’s intuition. It is far more common for someone to validate or discredit data based on how it lines up with his intuition rather than change his previous thoughts based on the data. This is much like when we get a diagnosis from a doctor. We are much more inclined to seek a second opinion when we receive a troublesome diagnosis. This makes sense too. It’s hard to trust something we don’t want to believe. Conversely, why look to data or a second opinion when we can have a quick answer with which we already agree?
Intuition should not be eliminated altogether just because it isn't a reliable source for answers. Intuition or experience can help guide the experiment. Intuition should be present throughout the whole data process, but it should never be a source of answers. Data analyses don’t always present tangible or confident answers, and intuition can be a source of how to change analyses that don’t come up with correct numbers. Predictions often fail and equations need to evolve over time to hone in the accuracy of predictions they are making. Intuition serves as the first step in correcting an incorrect analysis. If something about the analysis “feels off”, that should be the first place to check, but don’t get caught up on hunches if they turn out incorrect.
As I have stated before, there are limitations to using data, but these limitations only make data that much more useful. We can acknowledge the shortcomings of the data in every instance and say how much we know. The world is not black and white. The data may not necessarily be wrong if there is no clear answer, and this doesn’t mean go with whatever feels right. In fact, finding out there is no correlation, nor anything significant to look at further, is an answer.
In the end, it comes down to accountability. Intuition contains no accountability. You can change your opinion or make excuses why your intuition was wrong. There is no hypothesis or benchmark. On the other hand, accountability is inherently a part of data analysis; there are clear results, clear answers, and clear numbers. Data provides answers even when the analysis is wrong, and predictions fail. You can evolve the analysis and keep working towards an end goal. Basically, you know if you are right or wrong and where to go from there. If the goal is to make well informed decisions, you can’t do much better than using data analysis.
Banner image courtesy of Krissyho