How We Score Capability, and Why We Won't Score You on a Résumé

Every platform that scores people has to answer one question honestly: what is the score made of?

If the answer is "what you told us about yourself", the score measures how well you write about yourself. That is a real skill, but it is not the one anybody is trying to measure.

What we score

Seven capabilities: technical depth, communication, execution, ownership, collaboration, learning ability, and product thinking. Each one is built from evidence with a weight and a confidence attached — a completed challenge, a submitted deliverable, an expert's evaluation, the way somebody directed an AI tool, how their work changed after feedback.

Each piece of evidence carries a weight reflecting how much it tells us. An expert evaluation, written by a person who read the work, counts heavily. A deliverable counts, but less, because output can be polished by tools in ways that reasoning cannot. Behaviour over time — what changed after feedback — counts more than any single artefact.

Why a résumé can never make a score strong

A résumé is a curated document. It is written to be persuasive, it describes the past rather than demonstrating the present, and it is trivially inflatable. So it contributes, with a ceiling: evidence you've told us about yourself can place a capability at a starting point, and cannot push it into "strong". Only work we've observed can do that.

We found this the direct way. An account that had uploaded a résumé and done nothing else was reading 75 and 77 out of 100 on two capabilities — "strong" on both — because eighteen listed skills scored highly and nothing distinguished claimed from demonstrated. Listing skills is a typing exercise. We capped self-reported evidence, and the cap lifts the moment there's one piece of real work behind the score.

Where we deliberately stop

We don't score personality. We don't infer traits from how someone writes, or treat confidence as competence. We don't compare you to a "successful candidate" profile, because those models mostly learn who got hired before, and that is a description of the past, not a measure of the person.

We don't hide the basis of a score either. Every capability shows what it was built from.

What a score is for

A number should never be the whole story about a person, and a recruiter who hires on our score alone is using it wrong. It's a way to make evidence findable — so that a student who did excellent work in a programme nobody has heard of isn't invisible next to someone with a recognisable logo.

The evidence is the point. The score is an index to it.