What March Madness Taught Me About Business Analytics

Think analytics is overhyped? I’m about to prove you wrong.

March Madness is arguably the hardest prediction challenge in sports. With 63 games and 9.2 quintillion possible bracket combinations, the odds of a perfect bracket are essentially zero. Every year, millions of people fill one out based on gut feel, school loyalty, mascot strength, and vibes. And every year, most of them are wrong by the second weekend.

This year, I built a machine learning model to do it differently.

I trained it on years of college basketball data — offensive and defensive ratings, pace, strength of schedule, turnover rates, rebounding — and let it generate a bracket based purely on the numbers. No manual adjustments, no gut feel, no “I just have a feeling about this team.”

Here’s what happened:

• The model’s raw predicted bracket finished in the top 0.3% of 27 million national brackets.
• My final bracket — built using the model’s predictions as a foundation, with a small number of intentional adjustments I made based on additional context — finished even higher.
• The chalk bracket, the one you’d get if you simply picked the higher seed to win every single game? Not even close to that.

That gap is the point. That gap is what analytics does.

The Business Translation

I’m not writing this to talk about basketball (even though I do love to talk ball). I’m writing this because that gap shows up in business every day, and most companies are on the wrong side of it.

Here’s what I see repeatedly:

• “We’ve always done it this way.”
• “My gut says this is the right call.”
• “I don’t trust the numbers. They don’t capture the full picture.”

These aren’t unreasonable instincts. Experience and judgment matter. But when gut feel is the primary decision-making mechanism, you’re essentially filling out your bracket by picking mascots. Sometimes it works. Most of the time, you’re out by the second round.

The businesses that consistently outperform their competition aren’t necessarily smarter or more experienced. They’re more systematic. They’ve replaced “I think” with “the data shows.” They’ve built feedback loops so they know when they’re wrong and can correct quickly. They treat decisions as hypotheses to be tested, not conclusions to be defended.

That’s analytics. And it works.

What “Informed” Actually Means

Here’s the nuance I want to be careful about: I’m not advocating for blind trust in models.

My final bracket outperformed the raw model output because I layered informed judgment on top of the data. I knew things the model didn’t: a late injury, a matchup dynamic the historical data couldn’t capture. Analytics gave me a rigorous foundation. Human judgment added context the model couldn’t see.

That’s the real lesson. Analytics doesn’t replace decision-making. It enhances it.

The best business decisions I’ve seen come from teams that start with the data and then apply expertise. Not teams that use data to justify decisions they’ve already made, and not teams that ignore data entirely because “the model can’t capture everything.”

The sweet spot is in the middle: data-informed. Not data-blind, but not data-free.

The Uncomfortable Question

If a machine learning model — built by one person, using publicly available data, for fun — can outperform 99.7% of human judgment on one of the most complex prediction challenges in sports, what does that say about the decisions your business is making without similar rigor?

I’m not asking rhetorically. I’m asking because most organizations have far more data available to them than I had for this project, and far less analytical infrastructure in place to use it.

The competitive advantage is sitting there. Most companies aren’t picking it up.

Did I Convince You?

Analytics isn’t a silver bullet. It requires good data, the right questions, and people who know how to interpret the output. But the alternative of relying on instinct and precedent while your competitors build systematic edges is a losing bracket strategy.

If you’re curious what that looks like applied to your business, that’s exactly what I do at Integ Analytics.

marchmethods.org — the full model, prediction suite, and results tracker are all live if you want to dig in. I'm anxiously awaiting the next year...