Editor’s note: Welcome to Throwback Thursdays! Every third Thursday of the month, we feature a classic post from the earlier days of our company, gently updated as appropriate. We still find them helpful, and we think you will, too! The original version of this post can be found here.
What does it mean to be data-driven? Everything we do is driven by data. Even our intuition is really an accumulation of data points through experience. However, experience isn’t easily won or shared, and often we settle for repeating previous actions, or defer to the resident Hippo!
When we talk about being data-driven, what we actually mean is that we would like to make decisions based on the best data, made available to the most people. What does that mean for business, and how do you start?
By far the most difficult thing in being data-driven is getting the right data in the first place. CIOs frequently inherit a legacy of software applications that have been created with functionality in mind, not reuse of data. Data almost seems relegated to be a side effect of the operation of the machine, and hard to liberate. That’s changing, but it’s still remarkably easy to forget to think about exploiting data when you’re creating new systems. If we maintain a focus on the life cycle of data, we can avoid designing ourselves into silos.
The way in which data is made available is also key. There’s a difference between data for exploration, and data for operation. You wouldn’t expect anybody to pick a driving route just by watching the car’s speedometer. You can’t do without it, but neither is it enough. A dashboard is just table stakes: great for monitoring, but not for understanding patterns and correlations. Productive use of data in business involves learning the art of asking questions and finding that fingertip feel for the problem. That’s one reason that data lakes have people excited, as they hold the promise of making it quicker and cheaper to ask questions.
As with all business buzzwords, “data-driven” does us a disservice at the same time as highlighting a really important concept. We actually want to be “data-informed,” because humans are still the best decision makers we have in business. The aim is to force-multiply our most valuable asset: people. We should automate away all the messy bits of making sense of raw data, and present the human in the loop with the most useful information possible.
The point of being data-driven is to be able to take action. It’s an upgrade to the traditional reporting functions of analytics: we’re moving the use of data into the everyday operation of our business. It’s not a coincidence that companies who are lauded for being data-driven have also updated their infrastructure and working practices in order to be agile and move fast. Having the best data is of limited use if you’re unable to exploit it in the marketplace.
Creating a data-driven culture is a big undertaking, but you can start small and expand out. Start with a manageable problem and a willing team, where you can make an impact in three months or less. Work backwards from your business goals, and figure out what data or tools people need to make a difference. Do you need new data, faster data, more complete data? Does the team have the freedom to explore and ask questions?
Lastly, always hold the business goal in mind. In your words and plans, use the language of creating value, and fostering, rather than smothering, innovation. Being data-driven is about making the data work for you, helping you go further, faster.