STRATEGYWeeks to result90% confidence

Standardized Drill Benchmarking

Replace subjective scouting with smartphone-recorded drills scored against age- and gender-matched benchmarks.

Problem it solves

Evaluator scarcity and geographic bias mean most qualified candidates are never seen, and the few who are get judged inconsistently by one person on one day.

Best for

Organizations evaluating large pools of candidates where human reviewers can only cover a tiny fraction and bias or geography distorts the results.

Not ideal for

Highly contextual or creative qualities that resist measurement, or tiny candidate pools where one-on-one assessment is already feasible and unbiased.

Overview

Why this framework exists

Standardized Drill Benchmarking replaces opinion-led, geography-bound talent evaluation with a repeatable measurement system that any candidate can run themselves. The framework starts from an uncomfortable truth: when a small number of expert evaluators each see only a tiny fraction of the eligible population, talent outside their physical and social reach is functionally invisible. The fix is to decompose the evaluator's intuition into a set of predefined, recordable tasks, capture the raw performance data via a low-friction tool the candidate already owns, and score each result against benchmarks built from peers of comparable age, stage, and context. The framework has four parts. First, sit with the experts and reverse-engineer their intuition into the specific drills they actually use to form a verdict — sprints, jumps, dribbling, passing in the sport case; equivalents in other domains. Second, define what 'comparable, benchmarkable, reliable' data looks like for the people who will act on it, so the output maps cleanly onto an existing decision. Third, collect raw measurements at scale and build cohort-matched benchmarks (you cannot compare a 13-year-old to a 22-year-old). Fourth, position the system as augmenting, not replacing, the expert: top-ranked candidates still go to a human gate. The result is wider reach, faster screening, lower geography- and access-bias, and a transparent record that both evaluator and candidate can interrogate.

Core principles

5 total
  1. Reach beats taste — evaluator coverage is the binding constraint, not evaluator quality.
  2. Decompose intuition into observable drills before you try to score anyone.
  3. Benchmark within cohorts — compare like to like or the data lies.
  4. Augment the expert, do not replace them — keep a human gate on the final call.
  5. No black boxes — both evaluator and candidate must see where each score came from.

Checklist

Saved in your browser

Origin story

How this framework came to be

Richard Felton-Thomas, a biomechanist, was approached by Darren Peries — whose son had been hurt by the unfairness of youth scouting — with a question: could the lab's motion-capture protocols be compressed into smartphone drills any kid anywhere could run? They built ai.io's aiScout with Burnley and Chelsea, validated it when an unscouted 17-year-old living minutes from Chelsea's training ground topped their drills and went on to a Premier League contract.

Source

Traced to primary
Source · PODCAST
How AI Is Discovering Athletes That Human Scouts Miss
Richard Felton-Thomas
Open source →

Related frameworks

Browse all Strategy →