How the Savvy Score works
A 100-point offer score built from observable shopping evidence—not commission rates or sponsored placement.
One score, seven visible factors
V1 scores are deterministic. AI may explain the result, but it cannot secretly change the number.
30-day median
How far the current effective price sits below the product's observed 30-day median.
30-day low
How close today's effective price is to the lowest trustworthy 30-day observation.
Market comparison
How the offer compares with matching variants and conditions at other retailers.
Product value
Quality, performance, review signals, and lifecycle fit.
Retailer protection
Retailer trust, warranty, returns, and seller quality.
Shipping
Availability, delivery cost, and delivery confidence.
Data confidence
Freshness, observed-day coverage, source quality, and identity certainty.
Recommendation bands
90–100: Exceptional / Buy · 80–89: Strong / Buy · 65–79: Fair / Consider · 45–64: Weak / Wait · 0–44: Poor / Pass.
Low confidence, lifecycle risk, or a mismatch with the user's needs can soften the recommendation. Missing data is shown rather than guessed.
Dupe Scores answer a different question
Dupe Scores measure feature similarity, performance, build quality, use-case fit, savings, retailer quality, and confidence. A cheaper product is not automatically a good substitute.