In short
A useful scorecard is short and specific: four to six dimensions, each with a written description of what a strong answer looks like, each assigned to the stage and the person best placed to judge it. Agree it before the first interview and every requirement gets tested exactly once, instead of general impressions being scored three times.
Key facts
- How many dimensions
- Four to six is usually enough
- Agreed when
- Before the first interview, not after the first debrief
- Each dimension needs
- A stage, an owner and a description of 'strong'
- Biggest single fix
- Score evidence, not impressions
The problem a scorecard solves
In small teams, hiring decisions are often made from a conversation in a corridor: three people who each spoke to the candidate about roughly the same things, comparing impressions that were never written down.
A scorecard does not make the decision for you. It makes the inputs comparable, exposes where nobody actually tested the hardest requirement, and gives you something to look back at when a hire works well or does not.
Building one in under an hour
Start from the day-one requirements
The scorecard exists to prove those. If a dimension does not map to a requirement, ask why it is there.
Choose four to six dimensions
Depth, breadth, delivery, ambiguity, systems thinking, collaboration, communication, ownership. More than six and people score the same impression repeatedly.
Describe what strong looks like
One or two sentences per dimension, written in terms of what the candidate says or does — not how they made you feel.
Assign each dimension to a stage and a person
Everything gets tested once, by whoever is best placed to judge it. This also shortens the process.
Agree the non-negotiables
Decide before you meet anyone which dimensions you would not compromise on, so a strong performance elsewhere cannot quietly override them.
An example mapping
Illustrative only. The value is in doing this for your own role, not in copying the grid.
| Dimension | Stage | Evidence you are listening for |
|---|---|---|
| Technical depth | Deep-dive conversation | Can explain a decision they made and the alternatives they rejected |
| Systems and trade-offs | Design discussion | Names constraints and what they would give up, unprompted |
| Working under ambiguity | Deep-dive conversation | Describes how they decided what to do when the requirement was unclear |
| Delivery | Screen | Specific about what shipped, what did not, and their part in it |
| Communication | Founder conversation | A non-specialist could follow the explanation and make a decision from it |
Rules that keep it honest
- Score after each stage, before the debrief, so nobody anchors on the loudest voice.
- Write one piece of evidence next to every score.
- A missing score is information: it means nobody tested that dimension.
- Do not score 'culture' as similarity. Define observable behaviour, or leave it off.
- Keep the candidate's total time commitment to something you would spend yourself.
Common questions
- Is this too heavy for a five-person company?
- It is a one-page document written once per role. The alternative — re-arguing the requirements in every debrief — usually costs more time than writing it.
- Should candidates see the scorecard?
- Not usually the scores, but telling candidates what each stage assesses costs nothing and produces better, more relevant answers.
- How does this fit with a take-home exercise?
- Treat it as one stage that proves specific dimensions. If it is longer than the work it simulates, it filters for availability rather than ability.
Talk it through
A short, direct conversation about the role, the market and whether the SUS model fits what you are trying to do.
Arrange a conversationAny percentages or figures shown in this article are illustrative examples used to explain the model. They are not quoted rates, market benchmarks or salary data.