I used to think I was bad at math. Not a little bad. Genuinely convinced it wasn’t for me. Then I took undergraduate statistics and experimental design, and something clicked that has never unclicked since.

It wasn’t the numbers. It was the thinking.

I saw that I could construct evidence. I could test a hypothesis. And… more usefully… I could spot the holes in how other people were constructing theirs. That single realization sent me into a doctoral program in neuroscience, then computational neuropsychology, then public health, and eventually a master’s in statistics. Somewhere along that path, statistics stopped being a subject and became a philosophy. A way of seeing.

I’ve carried that lens through every chapter since… from reading brain scans to redesigning patient journeys and advising on customer experience at scale to teaching senior executives how to make better decisions.

Here’s what that philosophy actually looks like in practice:

Certainty is a posture, not a destination. Every inference we make carries the possibility of two errors… concluding something is true when it isn’t, or missing something true because our evidence wasn’t strong enough. What makes statistical thinking powerful isn’t just that it acknowledges this, it’s that it quantifies it. We don’t simply say “we might be wrong.” We operate with explicit confidence, and we measure exactly how much uncertainty we’re carrying. That discipline keeps thinking honest in a way that felt, to me, almost moral.

There are two major traditions for doing this. One asks: if I ran this experiment a thousand times, how often would I be wrong? The other asks: given what I already believed, how much should this new evidence move me? Different questions, different answers from the same data… and both are rigorous. That isn’t a flaw in the discipline. It’s an honest acknowledgement that how you frame the question shapes what truth looks like.

How someone balances those risks tells you everything about their priorities. The threshold for approving a cancer drug and the threshold for running a marketing campaign are not the same… nor should they be. When you see how a person or organization sets their tolerance for being wrong, you understand what they actually value, not just what they claim to.

Perfect data is a myth. Assumptions are everywhere. We almost never have the full picture… the entire population, every variable, every context. So we make assumptions that allow us to proceed. There are principled ways to do this. But those assumptions can also be quietly adjusted to point the results in a preferred direction. That’s not conspiracy, it’s a structural feature of analysis that every honest practitioner knows.

I can engineer almost any conclusion from real data. I say this not to be cynical but because it’s true, and because knowing it makes me a better reader of evidence. If I can do it, so can others. Math on its own does not give me security. Understanding how the math was applied does.

Almost anything is possible. Only a few things are probable. Base rates matter. Prior probability matters. The dramatic explanation is rarely the right one, even when the numbers can be made to support it.

In CX and patient experience especially, where so much is justified by survey scores and satisfaction metrics, this matters enormously. The number is rarely the whole truth. It’s a signal filtered through a dozen quiet assumptions.

In a world increasingly run by algorithmic black boxes, I find this more useful than ever. Not because I distrust data, but because I’ve spent enough time inside the machinery to know that data always has a hand on it somewhere.

The question is whose. And what they were trying to do.

Qaalfa Dibeehi is a Customer-Led Transformation Strategist, author, and keynote speaker. Founder of Human2outcome and former VP and Principal Analyst at Forrester Research. Former Dean of the Majid Al Futtaim Leadership Institute and Board Director at the Customer Institute. He has worked with organisations across Europe, North America, the Middle East, Africa, and Asia on customer experience, behavioural science, and customer-led transformation.