Nobody is average on everything. Almost every system assumes somebody is.

Every data system carries an assumption about who it is for. Usually it is an average: an average customer, an average applicant, an average patient. I have spent fifteen years watching that assumption fail quietly, in production, on real people, and never showing up in any dashboard.

The work has been architecture, strategy and leadership. I have built data functions from scratch, fixed ones that were not working, and helped organisations find out what they actually needed before spending money on the wrong thing. I have been inside companies and outside them. I have been a founder.

The technical part is the easy part. What is harder, and more useful, is the gap between the system and the decision. That is where the expensive problems live, and it is where almost nobody is measuring.

So I started measuring. Two papers so far, both open, both with the data and the code and the experiments that failed. A book in 2027.