Why Team Performance Varies

The question almost nobody asks correctly

Walk into most organizations and ask why a team isn't performing, and you'll get a list. Skills gaps. Engagement. Accountability. Better goals, better feedback, better incentives, better culture. The standard textbook on the subject — Aguinis's Performance Management, the field's most widely taught — opens its very first chapter with a table of sixteen things a good performance system is supposed to deliver, all at once: motivation, self-esteem, clarified roles, fairer pay decisions, retention, engagement, and a dozen more.¹

That list is the problem.

A list says everything matters, which is another way of saying you don't know what matters. It's the opposite of a diagnosis. A doctor who responded to "I'm tired" with "here are sixteen things healthy people do" would not be practicing medicine. Yet that is more or less how performance is managed: a standing program of best practices applied to everyone, all the time, whether or not any given one is the thing actually holding a particular team back.

Performix begins from a different premise. At any moment, for any team, one condition is binding. Find it, fix that, and performance moves. Spend your effort anywhere else and it doesn't — no matter how good the practice is in the abstract.

See the magnitude for yourself: Part I·b — The Impossible Checklist assembles the field's mainstream advice into a single daily checklist, shows where it openly contradicts itself, and names why managing without a diagnosis isn't just hard — it's worse than arbitrary. It is the problem this guide exists to answer.

Capability is not performance

Start with the finding that the whole field quietly knows and rarely acts on: capable people do not automatically produce performance.

Pfeffer and Sutton named it the knowing-doing gap — the hard part of management isn't knowing what to do, it's getting it done; the companies that win are the ones that turn knowledge into action.² Their point generalizes past knowledge to capability of every kind. A team can have the skills and still not deliver, because skill only becomes output when three other things are present: the team knows where to point it, has a reason to, and works in an environment that lets it.

This is the core of the CAMS model:

• Capability — can this team do the specific basket of work?
• Alignment — does it know what matters and where effort should go?
• Motivation — does it have the energy, commitment, and reason to act?
• Support — does its environment enable the work?

Capability is necessary. Alignment, Motivation, and Support are what convert it into realized performance. When performance is missing, the diagnostic question is not "are these people good?" — it's "which of the four is starving?"

This isn't a new intuition; it's a well-aged one. Thomas Gilbert put it plainly in 1978: when the plumber comes to fix your pipes, "we expect the expert to identify the precise cause of our problem" — "we would be angry indeed if this expert inspected our pipes with a list of true-false questions… and then gave us an achievement score."³ That is exactly what the performance-review apparatus does — a checklist and a rating where a diagnosis belongs. Gilbert's Behavior Engineering Model is the diagnostic engine inside the Human Performance Technology tradition: six cells split between the environment (information, resources, incentives) and the individual (knowledge, capacity, motives), with a rule to find the deficient cell before you intervene.⁴ Gilbert went further than most are comfortable with — he held that low motivation is usually a symptom, "a red flag to look for deficiencies in information, resources, or incentives," because the work environment, not exhortation, is what actually moves people.⁴ Whether or not you take it that far, the move is the right one: locate the constraint, then act.

What "performance" even is

Before you can diagnose it, you have to define it — and "performance" is two things people routinely conflate. Aguinis's own framework draws the line: there are results (outcomes) and there are behaviors (what people do that produces the outcomes).¹ Daniels, from the behavioral tradition, sharpens it into a discipline he calls pinpointing: name the valued accomplishment first, then the specific behavior that yields it — and note that "'lazy' is not a behavior."⁵ You cannot manage, or measure, a vague noun.

A second confusion worth killing early: experience is not capability. The research on expert performance is blunt about this. Length of experience is largely unrelated to improvement; radiologists' diagnostic accuracy plateaus — well short of perfect — and simply reading more scans afterward doesn't help.⁶ Expertise comes from deliberate practice with feedback, not from tenure. So a team full of veterans is not, on that basis, a capable team. (We measure Capability by skill, work samples, and learning velocity for exactly this reason — never by years served.)

Where the orthodoxy goes wrong — in its own words

You don't have to take the diagnostic case on faith. The performance-review apparatus that sits at the center of conventional practice is failing on the evidence of the people who run it.

• Employees don't believe in it. By the field textbook's own citation, only 3 in 10 employees think their company's review system actually helped them improve.¹
• It doesn't measure what it claims to. The largest studies of people rating other people find that about 60% of the variance in ratings is about the rater, not the person being rated — the "idiosyncratic rater effect" — and more elaborate rating scales make it worse, not better.⁷
• Even Harvard Business Review — the establishment — now documents the retreat: more than a third of U.S. companies have dropped annual reviews; Deloitte calculated its process consumed "1.8 million hours" that "didn't fit our business needs"; a Washington Post writer called it a "rite of corporate kabuki."⁸ At its peak, 60% of the Fortune 500 ran forced ranking.⁸
• And feedback itself — the supposed remedy — is not reliably benign. A landmark analysis of 131 feedback interventions found they decreased performance in more than a third of cases.⁹

Two conclusions follow, and they point in opposite directions from where most reform goes.

First, the answer to bad measurement is not no measurement. When HBR's reformers throw out ratings, they mostly land on the manager's unaided judgment — which is precisely the idiosyncratic rater the data just indicted. Drucker drew the right line sixty years ago: measurement exists to make self-control possible for the performer; "to use [measures] to control people from outside and above … is to abuse measurements."¹⁰ The fix is better measurement aimed at the right target — not a return to gut feel.

Second, the answer to the laundry list is diagnosis. Measure the four conditions, find the binding one, act only there, and re-measure. That is what the rest of this guide is about.

The shape of the argument

The remaining parts take the four conditions in turn — Capability (II), Alignment (III), Motivation (IV), Support (V) — defining each, its sub-constructs, and the science under it. Part VI covers the three levers a diagnosis points to: who you let in (select-in), how you develop the people already there (develop-in-place), and how you manage who leaves (exit-risk). Part VII is about measurement itself — how to measure these conditions well enough to bet on, without drowning in metrics or being fooled by them.

The throughline is one discipline, repeated: refuse the list, find the constraint, measure it honestly, move it.

---

Footnotes

1. Aguinis, Performance Management (3e, 2013), Ch.1 — Table 1 (16 "contributions"); results-vs-behaviors distinction (Ch.4); the 3-in-10 Watson Wyatt statistic (p.9). ↩ ↩² ↩³

2. Pfeffer & Sutton, The Knowing-Doing Gap (2000). ↩

3. Gilbert, Human Competence: Engineering Worthy Performance (1978) — the plumber analogy: diagnose the precise cause; don't run a true-false checklist plus an achievement score. ↩

4. Gilbert's Behavior Engineering Model — six cells; "evidence of low motivation is a red flag to look for deficiencies in information, resources, or incentives." Primary: Gilbert, Human Competence; operationalized in Van Tiem et al., Fundamentals of Performance Technology. ↩ ↩²

5. Daniels, Performance Management: Changing Behavior That Drives Organizational Effectiveness (5e) — pinpointing; "'lazy' is not a behavior." ↩

6. Ericsson et al. (eds.), The Cambridge Handbook of Expertise and Expert Performance — "length of experience unrelated to improvements"; the radiologist plateau. ↩

7. Buckingham & Goodall, Nine Lies About Work, Lie #6 — "about 60 percent of the variability in ratings" is the rater; complex scales worsen it. ↩

8. Cappelli & Tavis, "The Performance Management Revolution," in HBR's 10 Must Reads on Performance Management — Deloitte's 1.8M hours; "rite of corporate kabuki"; 60% of the Fortune 500 on forced ranking. ↩ ↩²

9. Kluger & DeNisi (1996), meta-analysis of 131 feedback interventions — performance decreased in more than a third of cases (cited via Carter & McMahon, Improving Employee Performance Through Workplace Coaching). ↩

10. Drucker, People and Performance — "to use [measures] to control people from outside and above … is to abuse measurements." ↩