Healthy Habits to Enable Automation
- Frictionless Decisions brings you counterintuitive, original, jargon-free ideas for connecting data to decisions. Every week.
In 25 years leading high-performing corporate data teams, somehow I’d never heard of this: my client asked me to explain how they could implement “self-healing” data.
Imposter syndrome set in; I stalled for time.
Then I called the smartest data engineer I know, wondering if he’d heard of data healing itself. He asked me if I just made it up. That’s when I realized it wasn’t me - it was them.
Why did something so obviously unrealistic make perfect sense to a room full of executives? For someone who believes in AI, everything is possible. Maybe self-healing data appeals to you just as much as it did to those executives.
The True Believers
Before we dismiss self-healing data as pure imagination, think about the reason people compare good data to a healthy life.
Good habits improve your chances for good health, but they don’t guarantee it. Like recovering from an illness or an injury, you follow your doctor’s orders, but it’s not your doctor who makes you well. We’d like more control, if such a thing exists, so we look to technology to solve our data health problems.
Nowadays, you can lose weight without changing your diet or exercise habits, but you’ll miss out on the intangible benefits that those good habits produce. If you immediately jump to an AI solution for better data quality, might you miss out on the benefits of a good business process improvement?
Learning to say no to your appetite benefits you in many ways beyond your appearance. My client, however, was looking for an instant fix.
Automation vs. Autonomy
Oracle (the world’s largest database provider) began marketing self-healing database features back in 2004; today, they call this “autonomous” database management. Using all available statistics, the most innovative databases available today can fully automate your database’s physical health.
“[W]e have created a way of essentially figuring out what indexes you need, based on the workload characteristics…All that is done under the covers without the DBA doing anything,” said Jenny Tsai-Smith, VP of Oracle’s Database Product Management group.
As amazing as these technologies are, the obvious question remains:
Will delivering data to people faster really improve their decisions?
Healthy Habits
James Clear, author of Atomic Habits, says, “Your weight is a lagging measure of your eating habits. Your knowledge is a lagging measure of your learning habits. Your clutter is a lagging measure of your cleaning habits. You get what you repeat.”
Here are some effective data habits you can start repeating today:
Write business rules instead of code. You can’t identify bad data unless you first define what good data looks like. A business rule is a specific definition of how things should work.
Wouldn’t you rather have your AI solution use your rules rather than make up its own?
- Organize data instead of writing reports. It takes a lot more effort to prepare data that’s useful for everyone in your company, not just the one report someone needs this week.
Do you want your AI solution to “guess the most statistically probable meaning” of your data, as Jessica Talisman asks, or to “reference the established definition for this context”?
- Create roadmaps instead of closing tickets. Deciding where to invest in a data solution that benefits everyone requires significant effort to get them all on the same page.
Do you really want your AI solution to follow a road to nowhere? You can start this work with my federal model for data governance.
Process vs. Product
Finally, I offered my client just one recommendation: Exception Reports. I proposed a process rather than a product, but they preferred to jump right into any AI solution anyone offered. Maybe these executives missed the irony of good data: AI can improve data quality, but only as part of a business process.
A process, not a product, makes data look like it heals itself.
To help remind you of these concepts, I add a hit 80s song to the Frictionless Data Spotify playlist each week. This week, enjoy Doctor Doctor by The Thompson Twins.
For the full story about making data flow faster and better, check out Frictionless Data on Amazon.
