Looks good is a friendly reaction, but does not explain why a piece is ready to publish. A few clear criteria can give both writer and editor a more useful conversation.
Separate non-negotiable errors
Invented experience, an unsupported consequential claim or incorrect attribution cannot be offset by appealing prose. I would not average those problems away with a general score. Some criteria are publication conditions rather than preferences.
Source verification and distinguishing examples from experience belong to this basic review. Unverified passages need an explicit resolution.
Read from five perspectives
Accuracy asks whether claims have a basis and clear limits. Coverage asks whether the title’s question is answered. Contribution examines the useful distinction, application or explanation offered. Structure considers the sequence of sections. Voice tests natural expression and fit with the publication.
Short evidence notes may be enough. Instead of structure is weak, identify which two sections perform the same job. The assessment should enable revision rather than merely deliver a verdict.
Adapt to the form
A question-and-answer article and a personal reflection need not share an outline. A checklist can be concise while a substantial guide requires more explanation. Word count alone is not a quality measure.
Different form does not excuse an unanswered question, however. A short piece still needs a complete thought, and a long one should not grow through repetition.
Turn the review into a decision
Ready, specific revisions needed and reconsider the approach are useful distinct outcomes. Each correction should have an owner and a reviewed version. The final pass checks whether previous notes were actually resolved.
I use a rubric to give different writing styles a shared level of care and reliability, not to make every article resemble the others.