The Three Levers

From diagnosis to action

CAMS tells you which condition is binding. The levers tell you what to do about it — and there are only three, because there are only three ways to change a team's performance: change who's on it, change what they can do, or manage who leaves. Every intervention is one of these.

• Select-in — who enters the system.
• Develop-in-place — diagnose the binding constraint and intervene on it.
• Manage exit-risk — who leaves, and why.

The diagnosis routes to a lever. A genuine Capability gap with no internal path points at select-in or develop; a starved environment points at develop (fix conditions); a flight-risk among your strong people points at exit-risk. The levers aren't independent programs — they're the action end of the same diagnostic.

Lever 1 — Select-in

The cheapest performance gain is often the one you make before anyone starts: don't let a misfit into the system. This is the domain of the personnel-selection canon — validity, structured assessment, work samples — and the discipline is the same one Part II insists on: select on demonstrated capability against the defined work, not on tenure or résumé, because experience length is largely unrelated to performance.¹²

Select-in is also where the affordability thesis bites hardest. Rigorous selection has been the province of organizations that could staff an I/O function; most can't, and so they select on gut and proxy. Making valid, job-analyzed selection affordable is a large part of owning the whole of performance — and it's the lever most organizations run worst, which makes it an opening, not a liability.²

Lever 2 — Develop-in-place

This is the lever the diagnosis points to most often — the learning arm that turns average into great. Development only works if it's aimed at the actual binding constraint, which is the entire reason the diagnosis comes first: training a team that's misaligned, or coaching a team whose environment is broken, burns goodwill and changes nothing.

When the constraint really is Capability, the science of how to develop is unambiguous: deliberate practice with feedback, against the specific missing tasks — not time-in-seat, not generic training.¹ When the constraint is in the consequence environment, Daniels's tools apply: fix what's reinforced so the right behavior is positively, immediately, and certainly rewarded.³ And the delivery vehicle is ongoing, on-the-job coaching and feedback, not the annual review — with the sharp caveat from Part VII that feedback is not automatically benign (it reduced performance in over a third of studied cases), so it must be diagnostic and well-aimed, not reflexive.⁴ This is why Performix embeds feedback in the work rather than bolting on a separate review: development is continuous, in-context, and tied to the diagnosed constraint.

Lever 3 — Manage exit-risk

The turnover science makes the lever precise, and corrects a common error: the goal is not to minimize all attrition. Mobley's foundational point is that turnover is a predictable process running through measurable antecedents — satisfaction, perceived alternatives, intention, commitment — diagnosable from the employee's perspective rather than guessed at after the fact.⁵ And not all of it is bad: functional turnover (the misfit or disruptive leaver) can improve the team that remains.⁵

So exit-risk, done right, is the mirror image of select-in: it's about who, not how many. Performix scores regretted vs. non-regretted exit-risk — flagging the valued people at risk of leaving (where retention pays) separately from the exits that are net-positive. The modern theory sharpens this further: people leave less from slow dissatisfaction than from shocks (the unfolding model) and weak embeddedness — both of which are measurable, and both of which fit the adaptive, probabilistic frame.⁵ This is the exit-risk analogue of the Attraction instrument: two protected-feedback indices, one on who you pull in, one on who you're about to lose.⁶

The levers and the diagnosis are one system

The point of three levers — and only three — is that a CAMS diagnosis is actionable by construction: every binding constraint has a defined response, and the response is chosen by the diagnosis, not by fashion or by whichever program a vendor is selling. Refuse the list, find the constraint, measure it honestly, move it — and "move it" always resolves to select-in, develop-in-place, or manage exit-risk.

---

Footnotes

1. Ericsson et al., The Cambridge Handbook of Expertise and Expert Performance — deliberate practice with feedback; experience length unrelated to improvement. ↩ ↩²

2. The personnel-selection canon — Schmitt (ed.), Oxford Handbook of Personnel Assessment & Selection; Cook; Gael's job analysis — plus the assessment-affordability thesis. ↩ ↩²

3. Daniels, Performance Management (OBM, 5e) — consequence engineering; PIC/NIC. ↩

4. Carter & McMahon, Improving Employee Performance Through Workplace Coaching — ongoing on-the-job feedback; Kluger & DeNisi (feedback reduced performance in more than a third of 131 studied systems). ↩

5. The turnover canon — Mobley (intermediate-linkage model; functional vs dysfunctional turnover); Hom & Griffeth; Lee & Mitchell (unfolding model / job embeddedness); Park & Shaw 2013 (turnover→organizational performance). ↩ ↩² ↩³

6. Attraction and Attrition as paired protected-feedback indices — one on who you pull in, one on who you're about to lose. ↩