Every organization, to be competitive and grow, is now facing the tough task of how to scale the talents of their people and processes through technology. It has become a necessary skill for leaders to learn how to effectively extract the knowledge needed for them to make decisions that will shape the future of their organizations. I asked Richard Hammer, a Senior Consultant with Cortex, whose expertise spans data analysis to enterprise architecture for scaling organizations and the technology that powers them the question: What is the difference between making data-driven decisions versus data-informed ones? The following was his answer.
Leaders are trying to make sound decisions driven by the data they have available to them. If you’re data-driven, though, you may run the risk of being on auto-pilot. Whatever scripts, macros, or code were written to analyze the data you are using to make decisions with is likely no longer reviewed and you have systems and people making decisions solely on the data, usually without questioning it. This is the single best way to give your competitors the advantage … especially if they figure out your patterns and differentiate themselves.
The trap of being data-driven is you tend to tune out your instinct. You tend to focus on a specific set of data, a specific result, and ultimately a specific outcome. You create your own data bubble and start eliminating data that may give you different insights. You zoom in, and in, and in, and in.
A better approach is to be data-informed. Sure, there are some practices where being data-driven is valuable, but I always question why, analyze the algorithm, and make small adjustments to the process to see if it becomes more efficient. If it does, we go forward. If it does not, we back the change out, analyze the data and results again, and make another adjustment. Systems can be built that do this automatically. This is the root of predictive analytics and artificial intelligence. You know you are on the right track here when the system recognizes a pattern you did not specifically tell it to recognize.
What you want to be doing is to take in all the data and keep an eye on the outliers and exceptions for later tuning.
If you are new to being informed by data, our suggestion is to get to know it. Ask questions about it. Discover the boundaries of the data you are mining. Do tests and confirm the results (make predictions, take action, and measure the response based on what you thought would happen).
An example of data-driven decisions is with A/B tests of a marketing campaign. You have 100,000 customers you want to reach. You do not know which message will most appeal: “BUY MY STUFF!!!” or “I think I found what you are looking for”. In an A/B campaign, you would send 1,000 random customers each message. When the results came in (who clicked more may be the indicator you are measuring), you send the remaining 98,000 customers whichever message is better performing.
To be data-informed, take a small step further. When targeting 1,000 customers each, make sure the demographics of the target customers are diverse. Be sure the purchase histories of the target customers are diverse. Look at the best practices for messages (all caps with exclamation points will likely not perform well) and identify a better A/B candidate. Better yet, focus your target list based on demographic, history, and something new. Use your instinct, test your gut, stretch and differentiate. Don’t be driven by the data. Let the data inform you in a way that aligns with your “gut”.
Some of the traps on the data-informed side can be decision constipation and not thinking in a way that allows you to scale the process. It is a fine line. Engineers want you to be data-driven because it is easy, consistent, and testable. Once the machine is in place, it runs. Engineers like that.
If you are looking to disrupt a marketplace, though, you need to sometimes question the machine itself.
Eventually, if you do not adapt, you become the machine and someone comes along to disrupt you.