Validation at scale

Assessing a handful of people honestly is a discipline. Assessing fifty — while levels feed reviews, promotions, and eventually pay — is a system design problem. The pressure is predictable: once levels touch money, levels inflate. Not because people are dishonest, but because incentives lean on the scale — on self-ratings and manager ratings alike — person by person, a little at a time.

The counterweight is making every level a claim others confirm:

Peer validation. A level isn't declared by one voice — whether the person's own or their manager's; the people who work alongside them corroborate it. This isn't a tribunal — most of the time it's a colleague nodding "yes, she delivers that on her own." The point is that levels become shared observations, not private declarations. Disagreement is a conversation about one skill at one level — small, concrete, and usually resolved in minutes.

Skill champions. For each critical skill, name a champion — someone at expert level who can answer questions, mentor others, and validate assessments in that skill. Champions give every skill a reference point: an agreed example of what expert actually looks like here. When someone wonders whether they're confident yet, the champion has seen enough of the skill to say. (Champions carry more weight in the team-development story too — see the team chapter.)

Teaching-back as the standing bar. The confident→expert test scales beautifully, because teaching is public. A claimed expert who has walked two colleagues to confident needs no further audit.

Handled this way, validation barely feels like control. It feels like what it is: a team keeping its shared map accurate, because everyone's plans are drawn on it.