Failure modes
Every assessment source has a bias signature. Knowing them is most of the defence — and it's why the method prefers levels that several sources can confirm.
Inflation. A level claimed (or granted) above the behaviour. It feels harmless, even kind — until the map assigns work that assumes foundations that aren't there. The cost arrives as frustration: tasks that feel impossibly hard, slipping deadlines, and a person quietly concluding they're not good at this — when the truth is only that a level got inflated and the missing foundation was never scheduled. Inflation doesn't skip the learning; it moves it to the worst possible moment.
Modesty. The same error in reverse, with a gentler face. Rate someone beginner where they're confident — whether from their own humility or an assessor playing it safe — and the map routes them around opportunities they were ready for: the stretch project, the mentoring role, the promotion case. The map can only open doors it believes someone can walk through.
Source-specific tilts. Self-assessments drift toward whichever of the two errors matches the person's temperament. Manager assessments lean on what the manager sees — skills exercised out of view get under-rated, recent wins outshine steady delivery. Credentials go stale: the certificate is forever, the skill isn't. None of this disqualifies a source; it just means no single source should be the whole story on a skill that matters.
Three habits keep placement honest, whoever is assessing:
- Anchor on the behavioural tells. Help along the way, or only at the edges, or almost never? That's a fact about last month's work, not a judgment of worth — and it's the same question for every source.
- Be specific. Broad skills invite vague ratings in any direction; "social-media posting: confident, SEO basics: beginner" leaves little room for ego or halo.
- Correct cheaply, early. The map has no memory of pride. Fixing a level costs nothing; living with a wrong one costs every plan built on it.
