01
The score is a summary, not a verdict
A 94% match should mean that the available evidence aligns strongly across the model’s stated dimensions. It should not mean a 94% chance of an interview, offer, strong performance, or happiness. Those claims would require outcomes data and controls that a recommendation score does not provide.
02
Read dimensions before the total
Two roles can share the same total score for very different reasons. One may be an excellent capability match with weak compensation alignment. Another may fit preferences perfectly but require evidence the professional has not yet demonstrated.
- Strong qualification + weak direction may be safe but stagnant
- Strong direction + evidence gaps may be a credible stretch
- Strong compensation + poor constraints may be impractical
- A missing salary range should remain unknown, not silently neutral
03
Distinguish concern from unknown
A concern is supported negative evidence—for example, a required certification that the professional does not hold. An unknown is missing information—for example, an undisclosed travel requirement. Treating an unknown as a failure is unfair; treating it as a pass is misleading.
04
Use feedback carefully
Saving, dismissing, or expressing interest can improve relevance, but one action should not rewrite a professional’s entire career intent. Feedback is most useful when the reason is explicit: compensation, location, scope, industry, timing, or lack of interest in the work itself.