How the Shadow Score Works
Every deal on ShadowMerchant has a Deal Ranking Score from 0–100. Here is exactly how it's calculated — and why commission rates have nothing to do with it.
🛡️ Our Conflict of Interest & Evidence Limits Disclosed Upfront
ShadowMerchant earns a small affiliate commission when you buy through our links. This is how we keep the service free.
Commission rates are not an input to our scoring formula. The score is calculated from observed price history, discount depth, and product ratings — data that exists before any commission relationship is considered.
⚠️ Evidence Limitation: A high ranking score indicates a favorable relative price drop and product rating based on observed scraper data; it does not constitute legal proof of original manufacturer MRP or merchant pricing intent.
The Formula
The percentage off the listed MRP, normalized against a 70% ceiling. A 70%+ discount scores the maximum on this component. This is the single largest factor — deep discounts are rewarded most.
The raw rupee amount you save, normalized against ₹3,000. A deal saving ₹3,000 or more scores the maximum on this component. This rewards large-ticket deals where the absolute saving matters, even when the percentage is moderate.
The number of customer reviews, normalized against 10,000. A product with 10,000+ reviews scores the maximum. Products with no review count available (common on Myntra, Nykaa, Meesho which do not expose this data) receive a neutral 0.3 score rather than being unfairly penalised.
The star rating out of 5.0. A perfect 5-star product scores the maximum. Products with no rating data available receive a neutral 0.3 score rather than being penalised for missing platform data.
Deals decay linearly over 7 days (168 hours). A deal scraped right now scores 1.0 on this component; a 7-day-old deal scores 0. This prevents stale inventory from dominating the feed.
What the Score Means
🏆 Great Value
80–100
👍 Good Deal
60–79
🆗 Fair Deal
40–59
😐 Low Score
0–39
📊 Scoring Adjustments
The five components are combined in a weighted sum, then passed through a sigmoid normalization. This means the displayed 0–100 score does not change linearly with the underlying weighted sum — it becomes progressively harder to reach very high scores, making 95+ rare even for strong deals.
Two additional penalties can reduce a score before normalization:
- Price above 7-day average: If the current price is higher than the 7-day observed average, up to −20 points is deducted proportionally.
- Suspicious discount: Discounts ≥80% or prices below 1/5th of the original trigger a trust penalty (heavy deduction), because such extreme discounts are typically signs of inflated MRP.
🌐 Currently Active Platforms
Products from Amazon, Myntra, and Nykaa are scored using the same formula. No platform receives preferential treatment in scoring. Platforms that do not expose review counts or ratings receive neutral scores on those components rather than being penalised for missing data.