A model that produces a correct conclusion nobody acts on has changed nothing. The path from analysis to behaviour runs through people, and that path is where most football analytics fails.
The output is an input to someone else's decision
An analyst produces a finding, a coach decides whether it fits the plan, and players decide how to apply it under pressure. Each step can stop the finding entirely.
A recommendation that survives all three has been reformulated at every stage into something actionable in the specific situation.
Measuring a model's value by its accuracy therefore misses most of what determines whether it mattered.
Instructions must be executable at speed
A player has fractions of a second to decide, which means a recommendation has to reduce to a rule simple enough to apply without deliberation.
A finding about optimal positioning becomes useful only when expressed as a reference point on the pitch or a cue to watch for.
Translating a statistical conclusion into a cue is skilled coaching work, and it is rarely done by the person who produced the analysis.
Credibility is earned on specifics
Coaching staff assess analytical work against what they have seen with their own eyes, and a finding that contradicts obvious observation is discarded regardless of method.
Analysts who establish themselves usually do so by being right about something small and checkable before proposing anything larger.
This is why analytics departments are frequently built around one person who has earned standing rather than around a platform.
Explanation determines adoption more than accuracy
A recommendation accompanied by a mechanism a coach can evaluate will be adopted ahead of a more accurate one presented without reasoning.
That preference is often criticised as resistance and is a reasonable response to being asked to act on something unexaminable.
Methods that expose their reasoning consequently outperform opaque ones in practice, even where they perform worse in testing.
Timing decides whether anything can change
Analysis delivered after a squad has trained a plan all week arrives too late to alter it, however sound it is.
Departments that influence matches are integrated into the weekly cycle and deliver conclusions when decisions are still open.
The organisational question of when analysis is produced therefore matters as much as the technical question of how, which is not how these departments are usually evaluated.