Experiment like a data pizzaiolo!

The 1 big idea: Experimentation is the engine that powers innovation. In data (and pizza), the biggest achievements (and best failures) come about when someone tries something new.

Why it matters: It’s easy to follow proven recipes. No one ever created anything new by following what someone else already did. Want to innovate? Experiment.

The backstory: I hosted a team of colleagues for pizza last week. I love pizza – especially woodfired pizza. (I may even be a bigger pizza nerd than data nerd.) I love experimenting with new dough techniques and unconventional toppings.

I made a pickle pizza. It was a first. It was also a huge hit. I’m adding it to my rotation (alongside my brie, fig, pistachio, and honey pizza and Capricciosa pizza with capers).

I also experimented with a new sour dough Detroit-style deep dish BLT pizza. It was a flop. The dough didn’t rise. Imagine bacon, lettuce, and tomato served over pound of cheese melted on pan of Play-doh.

But I learned from it. The sour dough had great flavor. I used too much sour dough starter. I’m not giving up on it. Thanks to Jonah Hohner for affirming it's a start!

Experimentation is the process of trying methods to discover what effect they have (The Britannica Dictionary). Without it, we’d never progress. We’d never learn. We’d never innovate.

Experimentation is also one of the most undervalued processes in analytics. Not because we don’t do it, but because it's easier to remember the figurative sour dough flops more than the pickle pizza hits. You can’t have one without the other.

7 tips to experiment with your data

Back to data: Here’s a few ideas from an aspiring pizzaiolo to help your experimentation:

1. Don’t fear failure. I’ve had a lot of them (with data and pizza). Every single one taught me something, refined my method, and was a catalyst for innovation.
2. Document everything. You can’t learn from a failure if you don’t understand it. I documented my failed dough recipe. Experimentation without documentation is just playing with data.
3. Balance your portfolio. I served 11 pizzas Wednesday night -- 2 were unproven and 9 were safe. Invest in a few high-risk, high-reward experiments along with a few safe bets.
4. Take inspiration from others. I wish I could claim credit to the pickle pizza. I saw it online once and put my own spin on it. Build on ideas from others. It's not cheating.
5. Solicit honest feedback. Did the experiment work? Why (or why not)? Was it just a cool idea or did it create value? Measure the results to improve next time.
6. Assess the right outcome. I’ve seen too many technologies written-off after the first use case was a failure. The BLT was a failure because of the dough -- not the toppings.
7. Try. Try again. The value of experimentation is deeper understanding. Don't give up after a failed data experiment else the deeper understanding can never provide its value.

Call to action (for leaders): Enable your team with the time and tools to experiment with data. Encourage their creativity. Celebrate successes, forgive (and learn from) failures.

Call to action (for data nerds): Obsess on the problem, not the solution. Too often we search harder for new uses of technology rather than new problems we haven’t solved.

The Runner vs. The Pizzaiolo

Final thought: This is a running blog -- here's the connection. I've learned there’s an inverse correlation between how much pizza I make and how much I run.

• My 2024 goal: 20 miles per week. I’m 100%.
• I made 11 pizzas Wednesday and 6 more on Friday.
• Prepping dough takes time. I ran 7.74 miles through Friday.
• That’s the most weekly pizza and fewest weekly miles in 2024.
• But... I hit a 15.4 mile run Saturday (longest ever) for my weekly goal!

This article is part of my blog, Running Thoughts on Data. My first post, The Story My Data Cannot Tell, shares the genesis of my blog. The views and postings on this site are my own and do not necessarily represent those of Plante Moran.