Decision scoring
How recommendations get scored.
Coodra scores candidate decisions by urgency, confidence, expected impact, data quality, and review history before they reach the queue.
Start with candidate decisions
A candidate decision is something Coodra might recommend: reorder this product, reduce that overstock, investigate a margin change, clean a catalog field, or check a supplier risk. Most candidates should never reach the top of the queue. Scoring exists to sort the useful few from the noisy many.
Score factors
Factor
Question Coodra asks
Effect on ranking
Urgency
Will waiting make the problem worse?
Raises items that are close to stockout, waste, or margin loss.
Confidence
Is the supporting data strong enough?
Demotes weak recommendations or marks them for investigation.
Impact
Would action materially help the retailer?
Keeps low-value noise below decisions that matter.
Review history
How did the team handle similar decisions?
Adjusts ranking based on approvals, skips, edits, and corrections.
Data quality
Are key fields missing or stale?
Routes questionable items toward cleanup or investigation.
Ranking is not a single magic number
Good ranking is a set of tradeoffs. A high-margin product with weak stock data may need investigation. A low-margin item moving quickly may still matter if it prevents a bad customer experience. Coodra should surface the tradeoff instead of hiding it behind a score with three decimal places.