AI for Retail

CHAPTER 12Hierarchy of Data Maturity

“If we have data, let's look at data.

If all we have are opinions, let's go with mine.”

—Jim Barksdale

As we stated in the beginning of the book, to leverage AI, you need real‐time, accurate data. If you do not have real‐time, accurate data, then you just have your intuition or worse, misleading or stale data. You can never pull off something like the OODA loop. Garbage in, garbage out.

Back to the self‐driving car example, if you accidentally got some mud on the front camera, and you could not observe the car coming at you very clearly, then you wouldn't orient yourself properly and detect the oncoming car, you wouldn't make the decision to stop, and you would smash into the car and be liable for damages. To show another failure mode, if you had some system lag where the OODA loop slowed down and lagged for three seconds between spins around the OODA loop, assuming you were going 70 mph, that would be about 300 feet during those three seconds, during that time you just traveled the length of a football field, which is more than enough time to smash through a few cars.

This is why accurate and real‐time data are so important to make autonomous systems work. “If you can't measure it (and measure it accurately), you can't improve it.”

How would you grade the accuracy of your data today? How would you grade the freshness of your data today?

Hierarchy of Data Accuracy

Back to our question that we asked in Chapter 3: Is Coke available for ...