Design assessment for hiring: tasks, rubrics, and scoring

What to screen for in a design hire

Hiring a designer without a design background feels like judging a wine competition by the labels. Every portfolio looks polished, partly because polishing portfolios is the one task every designer has practiced. What you actually need to know is how a candidate frames a problem, works inside constraints, and responds when someone says "make it pop."

A structured design assessment doesn't measure taste or tool tricks. It focuses on role-relevant competencies and produces evidence you can score consistently across candidates. If you're hiring this role now, the graphic designer hiring guide covers sourcing and interviews around this screen.

Nine competency domains cover most design roles. Weight the ones your role needs:

• Problem framing: defines the problem, success metrics, constraints, and stakeholders.
• Research and insight: chooses appropriate methods, avoids leading questions, synthesizes into actionable insights.
• Concept exploration: produces multiple viable directions and documents rationale.
• Systems thinking: coherent patterns, reusable components, predictable behaviors.
• Craft and execution: hierarchy, typography, layout, interaction clarity, attention to detail.
• Communication and critique: explains decisions, takes feedback, iterates.
• Collaboration: negotiates trade-offs, manages expectations, integrates cross-functional input.
• Ethics, accessibility, and inclusion: contrast, semantics, cognitive load; anticipates harm.
• Process integrity: works well async, documents process, uses AI with disclosure.

How to build the screen (6 steps)

Treat the assessment like a product: requirements, prototype, test, iterate.

1. Define 3-7 measurable competencies. Example: "Candidate can frame ambiguity, propose solutions, and communicate trade-offs." If you can't explain how a competency shows up in real work, don't assess it.
2. Choose evidence that would actually convince you: artifacts (wireframes, flows), decision logs, critique responses, research synthesis. Evidence must be observable and scorable.
3. Select the task format: a time-boxed work sample has the highest relevance; case analysis suits senior roles; a structured portfolio walkthrough with a rubric tests process integrity.
4. Build an analytic rubric with 4 performance levels, criteria mapped one-to-one to competencies.
5. Standardize the run: time limits, allowed resources, deliverables, submission format, and weighting, identical for every candidate.
6. Review after each hiring round: where did strong people struggle from ambiguity, which criteria had low rater agreement, and was the time-box realistic?

The task bank: 9 scenarios

Pick 2 or 3, time-box them, and use the same set for every candidate.

1. Problem framing under ambiguity

Scenario: redesign onboarding to reduce drop-off. No data provided.

Prompt: a one-page brief with problem statement, assumptions, risks, success metrics, and the first 3 research steps.

Covers: framing, outcomes, risk thinking. Strong candidates name what they don't know before proposing anything.

2. Research plan and bias control

Scenario: 5 days to inform a redesign for diverse users.

Prompt: choose 2 methods and justify them, write 6 interview questions, identify 2 bias risks with mitigations.

Covers: rigor, inclusion, method choice. Leading interview questions here predict leading research later.

3. Synthesis to insights

Scenario: messy notes from 8 interviews (you supply them).

Prompt: produce 5 insights, each with evidence, impact, and a design implication.

Covers: synthesis quality. Weak candidates return themes; strong ones return implications.

4. Interaction design with constraints

Scenario: a subscription tier change flow. Constraints: mobile-first, billing impact visible, screen reader support.

Prompt: key screens with state and error annotations.

Covers: systems, accessibility, clarity. The error states tell you more than the happy path.

5. Visual hierarchy

Scenario: a dense landing page that "needs to pop."

Prompt: a layout system (type scale, grid, spacing) with hierarchy decisions explained.

Covers: craft plus rationale. You're scoring the explanation as much as the layout.

6. Critique and iteration

Scenario: a stakeholder says "This looks boring."

Prompt: clarify goals, propose 2 alternatives, define what you'd test.

Covers: communication, stakeholder handling. This scenario is the day-to-day job in miniature.

7. Ethical decision

Scenario: growth wants a dark pattern.

Prompt: identify risks, propose an ethical alternative, define guardrails.

Covers: ethics, maturity, and whether the candidate can disagree usefully.

8. AI-era integrity

Scenario: AI tools are allowed on the task.

Prompt: a decision log covering AI use, what was validated, and what changed after feedback.

Covers: transparency and judgment. You want candidates who use AI and can tell you exactly where.

9. Cross-functional trade-offs

Scenario: engineering says the design is too complex to build.

Prompt: a phased plan (MVP, v1, v2) with trade-offs and acceptance criteria.

Covers: collaboration, feasibility. Candidates who treat this as a fight, lose the scenario.

Scoring rubric

Rubric levels (1-4)

• 1, foundational: incomplete, unclear, misaligned with the brief.
• 2, developing: partially there, gaps remain.
• 3, proficient: clear, complete, handles the constraints.
• 4, advanced: anticipates risks, strong trade-offs, clear judgment.

Default weights

• Problem framing and outcomes: 15%
• Research and insight: 10%
• Exploration and concepting: 10%
• Systems thinking: 10%
• Craft and execution: 15%
• Communication and critique: 15%
• Collaboration: 10%
• Ethics and accessibility: 10%
• Process integrity: 5%

Total: weighted score out of 4.0.

Hiring bands

• 3.6-4.0: advanced. High autonomy and strong judgment. Can own design decisions without a design manager over them, which is usually your situation.
• 3.0-3.5: proficient. Reliable performer; give clear briefs and they'll deliver.
• 2.3-2.9: developing. Capable but inconsistent. Only hire if someone can review their work.
• 1.0-2.2: foundational. Not ready for a solo seat.

Must-review thresholds

• Accessibility and ethics below 2.5: review regardless of total. This is compounding risk you'll live with in every shipped screen.
• Communication below 2.5 for client-facing roles: same rule.

Calibration (20 minutes, do this before you trust any score)

1. You and a second reviewer score one sample independently.
2. Compare domain scores.
3. Discuss discrepancies.
4. Rewrite anchors where you disagreed.
5. Re-score until you land within about half a point of each other.

Time-boxes that respect candidates

• Portfolio walkthrough with questions: 45-60 minutes.
• Take-home simulation: 2-3 hours maximum.
• Live exercise: 60-90 minutes.

Long unpaid take-homes hurt candidate experience and distort performance. The candidates with options decline them, which quietly filters your pool down to the ones without options.

Fairness, accessibility, and integrity rules

Non-negotiables for a hiring assessment:

• Job relevance: every task maps to work the role actually does.
• No spec work: fictional or clearly non-production briefs only. Never ship candidate work.
• Transparent criteria: tell candidates what you're evaluating.
• Consistent conditions: same task, time, and resources for everyone.
• Accessible formats: no color-only instructions, sufficient contrast, plain language, defined acronyms, multiple response modalities, and a clear accommodation process.
• Don't penalize communication styles that aren't job-critical.

For AI use, state it plainly: "You may use any standard design tools and AI assistants. Disclose where AI was used and what you validated independently. We evaluate reasoning and decision quality." Require decision logs and include a short live discussion of the work. Open-resource tasks that reward judgment beat surveillance every time.

Run this screen automatically with Truffle

The framework above assumes someone distributes the brief, collects submissions, applies the rubric, and keeps scoring consistent from candidate 2 to candidate 40. That someone shouldn't have to be you with a spreadsheet.

Truffle is a candidate screening platform that combines talent assessments with resume screening and one-way video interviews. Run the scenario prompts as an assessment, then use a one-way interview for the critique scenario: candidates walk through a design decision on camera, and you see how they explain trade-offs before booking a single call. Truffle scores every response against your criteria and ranks the pile with the reasoning shown. AI surfaces the evidence. You judge the work.

The 7-day free trial includes 30 credits, no card required. Start free trial