Technical assessment is the most situated of the five types. It measures specific role-relevant competencies against the levels the role requires. Unlike behavioural or cognitive — which are general — a technical assessment is built around a particular competency and a particular format.
What a technical assessment is
Each technical assessment in the library covers exactly one competency and is exactly one format — either MCQ (multiple-choice questions) or Case Study (written response). Assessments are created upfront in the library and assigned to people from there. The bundled multi-format approach (a single assessment containing MCQ, Case Study, and Interview sections assembled at assignment time) is superseded by this model.
If you need to assess two formats for the same competency, create two separate assessments — one MCQ and one Case Study — and assign whichever is appropriate. If you need an interview, schedule a standalone interview at assignment time (see the Assigning an assessment page).
What it measures
For each competency in the assessment:
- The employee's self-rating on the four-level proficiency scale (Beginner, Developing, Proficient, Expert). Each level has anchored descriptions (Beginner — works under supervision, Developing — handles mid-complexity with light guidance, Proficient — owns end-to-end delivery independently, Expert — sets standards and mentors others).
- A short evidence note — the employee writes 1–3 sentences explaining what they have actually done at the level they claim.
- The line manager's rating on the same scale, with the manager's own evidence note.
The platform blends the two ratings into a final per-competency score. The blend is weighted toward the manager (0.6) over self (0.3), with the remaining 0.1 reserved for any other documented evidence in the system (recent assessment results, interview ratings).
How it gets created and assigned
Create first: /{slug}/app/develop/library → + Create custom assessment → pick Technical → pick a competency → pick a level range (L1–L4) → pick MCQ and/or Case Study, and optionally tick Interview preferred → review and confirm.
Then assign: /{slug}/app/develop/assign → choose the employee → pick the saved assessment from the library → set a deadline → assign.
Interview preferred does not generate any content. It is a recommendation flag: when ticked, the saved assessment is marked as best paired with a human interview. If you pick any level above L1, the wizard shows a soft, non-blocking reminder suggesting you enable it. When you later assign a flagged assessment, a note reminds you it is best paired with an interview — schedule one from the assign screen.
The employee receives an ASSESSMENT_ASSIGNED notification. They take their part of the assessment first; the line manager's part unlocks once the employee submits.
What the runner looks like
For the employee:
- Each competency is shown one at a time.
- The four-level scale appears with anchored descriptions visible — the employee picks the level that matches their honest self-assessment.
- A free-text field below for the evidence note.
For the manager (after the employee submits):
- The same competencies, with the employee's self-ratings visible.
- The manager picks their own rating per competency, and writes their own evidence note.
- The manager can disagree with the employee — this is expected, not a problem to avoid.
Time to complete
Around 5 minutes per competency, for both the employee and the manager. A typical six-competency assessment takes 30 minutes from each.
Why the dual-rating model
Single-source ratings are noisy. Self-ratings are systematically optimistic at junior levels and systematically pessimistic at senior levels (the Dunning-Kruger effect, replicated in workplace data). Manager-only ratings have selection effects — a manager who likes the person rates them up; a manager who finds them difficult rates them down — that the employee's self-rating partially counterweights.
The blend is not a compromise; it is a diagnostic. Where employee and manager strongly agree, the rating is reliable. Where they disagree by two or more levels, the platform flags the competency for conversation rather than averaging through the disagreement silently.
How the result reaches the competency profile
Technical results contribute to the competencies they explicitly rate. The fidelity weight is 0.8 when the manager has rated; 0.4 when only the self-rating is in. This is why a technical assessment that the manager skips contributes much less to the profile than one they complete.
When to use it
- After hiring, to validate the new employee's self-described capabilities against the role's required competencies.
- During annual development cycles, to track competency movement year-over-year.
- Before considering an internal mobility move, to confirm the candidate has the required levels for the target role.
- In team formation, to see at a glance which competencies the team has covered and which it does not.
When not to use it
- For one-off competency snapshots where there is no manager available. A self-only technical assessment is weak signal; do something else if no manager can rate.
- For psychological or behavioural attributes — those are what the behavioural assessment is for.
Next
Structured interview — the fourth type — is covered next.