Early in my career at Capital One, I spent weeks building a model I was genuinely proud of. The validation metrics were solid, the methodology was sound, and I was ready to present. The meeting lasted twenty minutes. The stakeholders weren't interested in how I'd built it — they wanted to know what to do differently on Monday morning.
That moment reshaped how I think about analytics.
The model is the means, not the end
A model that nobody acts on is just an expensive calculation. The value of analytics comes at the point of decision — when a stakeholder understands the trade-offs, trusts the recommendation, and changes course.
At the Post Office, this is especially true. When you're presenting to board-level stakeholders about system improvements that affect Postmasters' daily lives, the framing matters as much as the analysis. You need to show the problem clearly, present credible options, and make a recommendation they can defend.
Three things that actually move decisions
Clarity over complexity. The best analysis I've delivered has often been the simplest — a clear comparison of two or three options with explicit trade-offs, not a 40-slide deck of methodology.
Stakeholder language. Translating model outputs into business outcomes (cost, risk, customer impact) is a skill that takes years to develop. It's also the skill that separates good analysts from influential ones.
Follow-through. Presenting a recommendation isn't the finish line. Working with stakeholders to design and implement improvements — and measuring whether they worked — is where real impact happens.
What this means for how I work
Whether I'm building a monitoring system, re-grounding a strategy model, or leading a team through a complex investigation, I start with the decision in mind. What needs to change? Who needs to change it? What would convince them?
The Python, SQL, and Tableau are essential — but they're in service of something bigger. Analytics is about decisions. Everything else is infrastructure.