Each of the pieces below is a retrospective, written after the work it describes was done, and so each carries two dates: the period it covers and the date it was written. I list the period first because, if you are here doing your homework on me, the chronology is likely what you are after, and I list the writing date so it is clear the piece was written later.

I name public programs and past employers where the work itself was public. My current employer appears as a 3,500-person global manufacturer, and the figures I use there are the ones I’ve already put on the record elsewhere. Where I haven’t put a figure on the record, I use ratios and orders of magnitude.

I’d encourage you to read them in the order below - oldest period first. The work timeline runs the other way, newest role first.

  • 2025–2026 · a 3,500-person global manufacturer
    Causal business models, or what to build once the whole business is in one data lake
    A metric tree that breaks the FP&A plan down into the drivers teams can actually move, kept as a living dataset rather than a dashboard and refreshed from the data lake often enough to compare what projects promised against what the business actually did - and how the natural experiments hiding in the data made the estimates better.