Data inspired vs. data driven

“I knew it!”

As human beings, we are in a constant state of confirmation bias. We validate our thoughts rather than creating actions from the data our brains' process.

Confused? Or did you know this already? 🤷‍♂️

“What are you on about this week?”

So, confirmation bias is the tendency to interpret new evidence as confirmation of one's existing beliefs or theories.

And we’re always at it.

Imagine you are thinking about someone and they call, you always say “How weird, I was just thinking about you!”. But it’s not weird at all. How many times did you think of that person and they didn’t call? How many times did they call when you weren’t thinking about them? Our brains are tricking us! 😱

The best example I read came from Cassie Kozyrkov, Head of Decision Intelligence at Google. She uses consumer reviews to explain that we don’t use them to decide which products to buy, but instead use them to validate the purchase we know we already want to make. In other words, whether a 4.2 out of 5 product review is good or bad depends on the decision we have already made as to whether or not we want to buy it.

“Go on then, what’s it got to do with fuel pricing?”

Kozyrkov goes on to explain the difference between data inspired and data-driven.

For example, cost prices are rising and we need to take some margin. We already know the sites you want to change, so we go in and check how they are performing and where our competitors are. The data ‘confirms’ our thoughts and our brains shouts “I knew it” and BANG we make our pricing decision.

This is what Kozrykov refers to as data-inspired. Our minds has been made up and really we need something to tell us not to make that decision in order to change our minds.

“So how do we become data-driven?”

In order to be data-driven, we need to set parameters before we look at the data and simply say “If I see this then I will do this”. Not easy to do when volume, margin and competitors are fluctuating throughout the day.

Using the example above where cost prices are rising, before we start looking at any data we should make at least one data-driven decision e.g. “If my margin isn’t at 10ppl then I’m not budging, no matter where the competition or the volumes are”. We should be deciding on acceptable parameters beforehand.

And we’re not suggesting that we remove all objectivity from the process, as there is a lot of art in fuel pricing as well as science. Plus it’s nearly impossible to shake our bias and be completely data-driven.

But, as with any science, having a hypothesis will help you make more sense of the data we are looking at.

If you want to read the article cited in this week’s wrap, you can find it here