Capability — Can the Team Do the Work?

The question

Can this team do the specific basket of work?

Note the three load-bearing words: specific, basket, work. Capability is not a general property of people — not how smart they are, not how senior, not how impressive on paper. It's a relationship between these people and this work. Which means you cannot assess Capability until you've defined the work.

The precondition: a current, accurate job spec

This is the part everyone skips, and it's load-bearing. Before "are they capable?" comes "capable of what, exactly?" — and the answer is a job analysis: in Gael's definition, "an efficient method of assessing what tasks are involved in specific jobs and how workers perform these tasks."¹ You enumerate the tasks, write them as statements, weight them by the time and value they carry, and you treat the inventory's reliability and validity as seriously as any other measurement.¹ A task, in this tradition, is "a mixture of decisions, perceptions, and/or physical activities … directed toward a specific purpose"¹ — a unit of work, not a personality trait.

The behavioral and expertise traditions arrive at the same discipline from other directions. Daniels calls it pinpointing — name the valued accomplishment, then the behavior that produces it, and stop describing people with adjectives ("'lazy' is not a behavior").² Ericsson's expertise research defines skill as "reliably superior performance on representative tasks" — tasks that actually stand in for the job.³ All three say: define the work as tasks and accomplishments first. That defined spec is the precondition for everything downstream — measurement, development, selection. It is also, historically, the expensive part, which is why most organizations don't have one.

Capability is not tenure (or résumé, or seniority)

The most common Capability signal in use is the worst one: years of experience. The expertise research is unusually blunt here. Length of experience is largely unrelated to improvement in professional performance; in one well-known case, radiologists' diagnostic accuracy plateaus — well short of perfect — and reading more scans after the plateau doesn't move it.³ Experience produces familiarity, which feels like skill and often isn't. What actually builds capability is deliberate practice with feedback, not accumulated time-in-seat.³

So Performix measures Capability the way the science says is valid, not the way org charts make convenient:

• Skill and knowledge demonstrated on work samples / representative tasks, not inferred from tenure.³
• Learning velocity — the rate at which a person acquires new capability under practice — which predicts future performance better than the stock of past experience.
• Judgment and role fit against the defined task inventory.

(A note on a boundary the model enforces: the feeling of competence — the sense of being effective — belongs under Motivation, where Self-Determination Theory puts it. Capability is about actual skill against actual tasks. Don't let a confident team read as a capable one, or an anxious capable team read as incapable.⁴)

Measuring it well — and why that used to be impossible for most

Measuring Capability properly is a psychometric problem, and the methodology is mature: criterion and construct validity (does the signal predict performance?), item response theory, and adaptive testing that estimates ability efficiently with stated uncertainty rather than a single hard number.⁵ This is exactly the machinery that large organizations buy I/O psychologists to run.

That's the whole point of where this is going. The reason most organizations don't assess Capability well isn't that the science is unsettled — it's that the science has been inaccessible and unaffordable, available only to those who could staff a measurement function. The unlock is using AI as a consumer of that substrate to make the same machinery affordable for everyone else.⁶ Performix doesn't reinvent the engine — the adaptive psychometric engine and the universal job schema live upstream, and Performix surfaces them — but it puts I/O-grade Capability measurement within reach of organizations that have never had it.

From individual to team

One honest open edge: the assessment canon is built to predict an individual's job performance, and CAMS asks a team question. Team Capability is not the average of résumés. The more defensible reading is coverage: given the task inventory that defines the work, does the team — in aggregate, with its mix of skills and learning velocity — cover the basket? A team can be individually strong and collectively miss a critical task class, or individually middling and fully cover the work. That's the construct we're building toward.⁴

When Capability is the binding constraint

If the diagnosis lands on Capability, the lever is develop-in-place (deliberate practice with feedback against the missing tasks — Part VI) or, at the boundary, select-in (bring in the missing capability). What it is not is more motivation or clearer goals — a team that genuinely cannot do the work will not be exhorted or aligned into doing it. That is the entire value of getting the diagnosis right: it tells you this is a skill problem, so you spend on skill — and not on the three things that would have been the answer if the constraint had been Alignment, Motivation, or Support instead.

---

Footnotes

1. Gael, Job Analysis: A Guide to Assessing Work Activities — job analysis defined; the job-inventory (WPSS) approach; task = "a mixture of decisions, perceptions, and/or physical activities … directed toward a specific purpose." ↩ ↩² ↩³

2. Daniels, Performance Management (OBM, 5e) — pinpointing; "'lazy' is not a behavior." ↩

3. Ericsson et al., The Cambridge Handbook of Expertise and Expert Performance — "reliably superior performance on representative tasks"; experience length unrelated to improvement; deliberate practice. ↩ ↩² ↩³ ↩⁴

4. The CAMS construct model — Capability sub-constructs; the competence-feeling (Motivation) vs actual-skill (Capability) boundary; the team as focal unit. ↩ ↩²

5. van der Linden, Handbook of IRT Vol 1–3; Embretson & Reise, IRT for Psychologists; the computerized-adaptive-testing canon; DeVellis, Scale Development; Schmitt (ed.), Oxford Handbook of Personnel Assessment & Selection. ↩

6. The assessment-affordability thesis: the psychometric machinery is mature but has been affordable only to organizations that could staff a measurement function. The engine lives upstream of Performix; the product surfaces it. ↩