Data-Driven Decision Making?

Thoughts on how to be sure you're really letting the data lead.

We all talk about data-driven decision-making, but let’s be honest—how often is data actually leading the decision, rather than just justifying one we’ve already made?

Daniel Kahneman’s work on cognitive biases reminds us that our minds are wired to seek confirmation, not contradiction. When we cherry-pick data to support our preferred narrative, we’re not making data-driven decisions—we’re making decisions with data as a prop.

Consider these real-world examples:

• The Advertising Spend Fallacy: Global brands often slash digital ad budgets after short-term test showed little ROI. The problem? These tests don’t account for delayed impact on brand equity and long-term sales—a classic case of availability bias.
• The Hiring Algorithm Dilemma: A major tech company built an AI hiring model trained on past successful candidates. What did they miss? The model reinforced historical biases, excluding top talent that didn’t fit past patterns— this is a case of data reflecting bias, not removing it.
• The Retail Location Mistake: Some retailers use foot traffic data to pick store locations but ignore demographic and purchasing patterns. The results can be disappointing. High-traffic locations with low conversion rates can rise to the top, — support for the idea that data needs to be examined carefully and holistically before it is decision-ready.

How Do We Avoid these Traps?

• Interrogate the Data: Think carefully about what's missing. What assumptions are baked in?
• Challenge Your Own Bias: Ask yourself, "Would you accept the data if it told a different story?"
• Use Data for Discovery, Not Just Justification: Practice patience and having the discipline to let insights emerge where they will rather than forcing a narrative.

Remember, data is powerful—but only when we use it honestly and rigorously.