"Looks good" is the most expensive sentence in marketing operations. Every team using it has a quality bar that lives in one person's head — and a content engine that depends on that person being available, in a good mood, and remembering what they decided last time.
What a rubric replaces
A rubric replaces taste calls with a list of criteria, weights, and pass/fail thresholds. It doesn't remove judgment — it makes judgment auditable, teachable, and consistent across the four people who all said "looks good" but meant four different things.
A starter rubric for content
- Brand voice match — does this sound like our last 30 approved pieces? (0–25)
- Hook strength — does the first line earn the second? (0–20)
- Specificity — does this say something only our team could say? (0–20)
- Proof — at least one concrete claim with a citable source? (0–15)
- CTA clarity — is the next action obvious and one-tap? (0–10)
- Compression — could it lose 20% of words without losing meaning? (0–10)
Memory makes it compound
A rubric without memory is a checklist. A rubric with memory is a learning system. Encode the approved pieces, the rejected pieces, and the reviewer comments — and the next draft starts further along the curve.
You can't scale taste. You can scale a rubric — and a rubric with memory is the closest thing to a content quality system that compounds.