You've achieved a great deal at this point and are well on your way to realizing the full power of the dark side your data.
In this post, I wanted to share some additional avenues for you to explore as you travel down this path with your data and master the basics.
I want to first talk about a powerful method of using data to steer your company and grow your businesses. This is known as Split Testing or also some call it A/B Testing.
There has been a lot said about this and there are some awesome tools online to help with this however I wanted to put out some basics if you're new to the concept.
Split Testing is at the foundation of a Data-First approach to business. This is an idea that really resonates with me as a data geek and hopefully you as well by this point 🙂
The idea is that you put out multiple versions of something and measure users behavior to understand the impact this variation had on your user's behavior. This stems from clinical trials in medicine where some participants are given the magic weight loss pill and the others are given a placebo.
Isolating the variation between user behaviors to just one variable helps you improve results by learning what your users want. That is to say, it's a quantitative approach to business strategy and product design. I love this approach, because for as much as I love people, I don't believe most of what people tell me.
I do advocate for surveying your users to understand their psychology and sentiment however people rarely do what they say. Don't believe me? Here's some proof…
Christian Rudder wrote an excellent book, Dataclysm, on this topic. Christian is a data geek, like you, and was a co-founder of the online dating site OkCupid. As a data geek, he shared a lot of his great insights on OkTrends for years until he published this book.
In it, he tells many stories about how we don't actually do what we say we do online. The most poignant example I found in his book is this
It's a little blurry but what it shows is the age of men on OkCupid on the left, the shaded area of their stated preference for women, and then the red number of the age range they message the most.
Long story short, we're not who we say we are online. So don't put all your trust in qualitative feedback or sentiment analysis from your users. Try split testing to see how they actually behave when presented with varying versions of your content.
There is a great business mind from the fifties who revolutionized the Japanese auto industry after World War II named W. Edwards Deming. He was what we today would call a Data Scientist.
He had many great quotes, but the one above is particularly applicable to this post. What he's saying here is that just busting your ass and hustling won't cut it. You need to know where to focus and what to bust your ass doing in order to be successful.
As a prescription for this problem, he developed a system for figuring out what to do. Perhaps I'll do a future series or longer post on this process but for now, here is the summarized version.
PDSA = Plan Do Study Act
Design an experiment you would like to try to improve your business.
Run the experiment
Review the results and dig deep into the causality and if your experiment gave you the expected outcome.
Decide to actually revise your experiment or make a whole scale change to your system (business).
Deming was the godfather of systems thinking about business. That means when you're looking at your business, through data, you can't change one thing w/o another being affected. Remember our data map from the first post in this series? That right there will show you how everything is connected and how changing one variable can have an impact on the rest of your business.
This sucks…because this means there are no easy answers. We can't merely address the little variations in our data one at a time to solve them. We need to design experiments, PDSAs, and run them to understand how the system reacts and if our experiments gave us the outcome we were looking for.
Curbing Your Ego
The last thing I want to cover is probably the most difficult of them all. Curbing Your Ego.
This is hard because as humans, we tend to believe our own bullshit and believe we're right all the time. When we're trying to build an business however we need to think of ourselves as experimenters that use data to guide us rather than our gut.
It's not to say our gut isn't right from time to time! But we definitely need to at the very least understand the impact our actions have on our business to gain perspective about how to make better decisions in the future.
Ready to become Datapreneur?