EN
Wooden mechanism with curved rails, metal pieces, and a blue rod on a white background.

How to choose the first track for a skills pilot

The first track in a skills pilot should generate useful learning without sacrificing relevance. Choosing it requires separating indispensable conditions from criteria used to compare viable candidates.

PRYSMAP6 min read

This distinction changes the decision. If every criterion is immediately converted into points, an important track that cannot be executed may rank first. The most orderly track may also win even when its result has little value for the organization.

Selection works better in two stages. First, determine whether each candidate meets minimum conditions. Then compare only those that could sustain a useful pilot.

Before choosing the track, define what the pilot must resolve

A pilot should not begin with a list of available areas. It should begin with a decision that needs better answers.

That decision may concern development, mobility, succession, hiring, or clarity of expectations. Its formulation determines which track is relevant and what the organization learns beyond the exercise.

The OECD notes that organizations can begin skills-based approaches through pilots focused on a small number of roles. It also recommends identifying priority roles and involving relevant stakeholders early. The same source acknowledges that available implementation evidence remains limited and largely anecdotal. This reinforces a cautious idea: starting with a narrow scope enables learning, but it does not guarantee that every small scope will produce useful learning.

The initial question must be concrete:

Which organizational decision should we be able to make better after this pilot?

If the answer is vague, there is not yet a sufficient basis for choosing the track. Testing the methodology is not enough either. The pilot needs to clarify which assumption, risk, or decision it will test.

Four conditions every candidate must meet

Entry conditions are pass/fail, not weighted. If one is missing, the track needs further preparation before it can be compared with the others.

1. A real decision exists

The result must have an intended use—for example, reviewing expectations, prioritizing development, or understanding proximity between roles. Without a defined decision, the pilot may produce interesting information and no verifiable consequence.

2. Sponsorship and operational accountability exist

Someone must be able to resolve priorities, reach participants, and remove obstacles. Someone must also be responsible for coordinating the day-to-day work. Symbolic sponsorship contributes little when operational decisions have no owner.

3. Experts and accessible evidence exist

The track needs people able to describe the work, distinguish performance levels, and review the observable manifestations of each skill. Relevant documentation, practices, results, or other evidence must also be accessible.

Familiarity with the roles does not replace this condition. Knowing the people who work in the track may facilitate access, but it does not ensure consistent definitions or sufficient evidence.

4. The track remains stable during the trial

The track need not remain static. It does need a reasonable period in which roles, expectations, and accountable people do not change so much that the observations become impossible to interpret.

An intense transformation can make a track highly relevant and unsuitable as the starting point at the same time. In that scenario, the initial priority may be to stabilize definitions or narrow the track.

Compare the tracks that remain viable

After the filter, five criteria make it possible to compare candidates. They have no universal weights. Their importance depends on the question defined for the pilot.

Criterion Comparison question Favorable signal
Relevance Which decision will improve if we understand this track? An identifiable organizational consequence exists.
Stability Will we be able to interpret results during the pilot period? Roles, expectations, and accountable people maintain sufficient continuity.
Urgency Does the need justify acting now without distorting the trial? There is real pressure, but still room to learn and adjust.
Access Can we work with experts, participants, and relevant evidence? The necessary sources are available and stakeholders can participate.
Reusable learning What can we later apply to other tracks? The candidate allows assumptions that will recur at larger scale to be tested.

The comparison should preserve the reasons, not only the ratings. Two teams may rate reusable learning as high for entirely different reasons. Those differences will matter when the pilot is interpreted.

Three small cardboard models on a worktable, including one held by a blue clamp.

Dependencies among criteria should also be reviewed. Extreme urgency may reduce the time available to define expectations. Very easy access may make a low-relevance track attractive. High strategic relevance may not be enough if no expert is available to review the architecture.

Break ties by asking which insight matters most

If two candidates are close, adding points rarely resolves the disagreement. The tie forces the decision-making team to decide what it needs to learn first.

The UK guidance on testing and piloting services recommends setting clear objectives, validating assumptions, and recognizing unforeseen challenges before scaling a solution. The UK government’s Service Manual proposes a similar logic for the alpha phase: focus the minimum necessary effort on the riskiest assumptions.

Applied to a skills pilot, a tie can be resolved with three questions:

  1. Which assumption could most strongly invalidate later expansion?
  2. In which track can we observe that assumption with less ambiguity?
  3. Which outcome, even if negative, would yield the most useful insight?

These questions shift the choice from convenience to the value of the information it can produce. A pilot is also useful when it reveals unstable definitions, insufficient evidence, or an intended use that needs adjustment. To learn from that outcome, the conditions and reasons for the selection must be recorded from the beginning.

End with a documented selection

Feasibility The team can access the evidence and has time to complete the pilot.

Relevance The output can support an upcoming decision.

The process can be completed in five steps:

  1. Formulate the decision the pilot must support.
  2. Identify two to four candidate tracks.
  3. Apply the four entry conditions and remove candidates that are not yet viable.
  4. Compare the remainder by relevance, stability, urgency, access, and reusable learning.
  5. Choose a track and record the main uncertainty it must resolve.

The final record need not be extensive. It should include the selected track, the affected decision, rejected candidates and the reasons for rejecting them, verified conditions, determining criteria, and the priority assumption.

The selection will still depend on judgment. The matrix makes that judgment visible, open to discussion, and revisable when new information appears.

The right first track will be important enough to justify the effort, stable enough to interpret the results, and accessible enough to enable learning. That combination makes it possible to begin with a concrete decision without confusing a small start with an irrelevant one.

  5. Four mechanical modules are aligned; the two central ones are connected by a blue looped circuit.

    When to split, merge, or retire a skill

    Splitting, merging, and retiring a skill correct different failures. Even if the catalog looks tidier after the change, the choice depends on which distinction in the work must be preserved and which decisions will change.