Skip to main content
Methodology

The Chayan Score, fully transparent.

Two numbers, kept honest and separate: a quality score anchored to the verified customer rating, and a confidence level from how many people actually reviewed it. When we don't have enough evidence to score a product, we say so — we never invent a number. No black box. No trade secrets.

Last calibrated: Jul 15, 2026 Version 3.0
The Formula

Two axes, kept separate

A great rating from 8 people is not the same as a great rating from 8,000. We measure both, and we never let one masquerade as the other.

Axis 1 Chayan Score quality — from the verified customer rating, calibrated for India
Axis 2 Chayan Confidence evidence — how many verified reviews stand behind that rating

Axis 1 — the Chayan Score (quality)

The Chayan Score answers one question: is this product actually good? It is built from the single signal that reflects real ownership — the verified customer rating — recalibrated for how Indians rate on marketplaces.

Why not just show the star rating? Because 91% of rated products on Indian marketplaces sit at 4.0★ or higher. The raw star is so compressed it can barely separate good from great. Our curve stretches the top end so a genuinely excellent 4.8★ pulls clearly ahead of an ordinary 4.1★:

Verified rating Chayan Score Band
5.0★92Top pick
4.5★75Top pick
4.0★67Recommended
3.5★60Recommended
3.0★50Below average
Below 3.0★<50Avoid

Ratings between the rows are interpolated. The algorithmic score is capped at 94 — the top six points, 95 to 100, are never computed. They are reserved for the handful of products our editors designate an Editor's Pick.

Axis 2 — Chayan Confidence (evidence)

Confidence answers a completely different question: how sure are we? It is set by how many verified reviews stand behind the rating — nothing else.

Confidence Reviews What you see
High100+Full Chayan Score, shown with high confidence.
Medium10–99Full Chayan Score, labelled “Medium confidence.”
Lowunder 10No number — “Rating not established.”

Crucially, review count is a confidence dial, not a quality booster. More reviews make us surer of the rating; they don't make the product better. (We checked our own catalogue: a product's rating and its review count are statistically independent.) So a 4.5★ product with 12 reviews scores just below a 4.5★ product with 5,000 reviews — but only slightly.

The four bands

Band Rating What it means
Top pick4.5★+The best the data can say.
Recommended3.5–4.5★A solid, safe buy.
Below average3.0–3.5★Proceed with caution.
Avoidunder 3.0★We'd skip it.

When we show “—” instead of a number

If a product has fewer than 10 verified reviews, we show “Rating not established” and no score. We could estimate one from the brand or the spec sheet — we choose not to. A guess dressed up as a score is the fastest way to lose your trust. Instead we show what we do know: the brand, how many stores carry it, and whether it is a current model. About two-thirds of our catalogue sits here today, and we would rather be honestly silent than confidently wrong.

The rejection line — “Avoid”

A product lands on the rejection list when its Chayan Score falls below 50 — which, on our India-calibrated curve, means a verified rating under 3.0★. By Indian marketplace standards that is genuinely poor, not merely average. We only judge products with at least 10 reviews, so the list stays small and every entry is one we can stand behind rejecting.

Editorial judgment — kept out of the number

Earlier versions of the Chayan Score added quiet per-category “boosts” — a few points here and there for categories we felt the formula under-rewarded. We have retired that. A number you can't fully reconstruct from its inputs isn't transparent, however well-intentioned the nudge.

Editorial conviction now lives where you can see it: the Editor's Pick designation. The algorithm caps itself at 94; the top of the scale (95–100) and the Editor's Pick ribbon are the only places human judgment touches a product — always visible on the card, never buried inside the score. The number itself is now purely the verified rating and the weight of evidence behind it.

Where the data comes from

Amazon.in
Pricing · Reviews · BSR
Flipkart
Pricing · Reviews
Croma
Pricing · Reviews
Reliance Digital
Pricing · Reviews
Nykaa
Pricing · Reviews
Nykaa Fashion
Pricing · Reviews
Cuelinks
Affiliate attribution layer
Public review threads
Reliability signal
Category schemas
Feature checklists (ours)

How we don't cheat

Calibration schedule + changelog

The scoring model is reviewed every quarter. When we change it, we publish it here — so if we drift, you can see exactly when and why.

Jul 15, 2026 v3.0 — Two-axis rebuild. Retired the weighted 5-factor formula. The Chayan Score is now anchored directly to the verified customer rating (calibrated for India's rating inflation), with review volume acting as a bounded confidence multiplier — never a quality booster. Chayan Confidence became a separate evidence axis. Products with fewer than 10 verified reviews now show “Rating not established” instead of an estimated score. Per-category editorial boosts were retired; editorial judgment moved to the visible Editor's Pick tier.
Jun 13, 2026 (later) v2.6 — Lowered rejection threshold from 65 → 50. The previous threshold caught ~50% of any real-world catalog, devaluing the "rejected" signal. New three-tier system: 65+ Recommended · 50-64 Listed · <50 Rejected. The rejection list now contains only products that fail 2+ sub-scores — when we say we wouldn't buy something, we mean it.
Jun 13, 2026 v2.5 — Added editorial boost system. Introduced category-level score boosts for cases where the 5-factor formula systematically under-weights value. First boost: +9 for spiritual / contemplative books. Every boost is shown as a labelled pill on the product card with a hover-rationale; pre-boost score is also exposed.
May 28, 2026 v2.4 — Reduced popularity_trend from 18% → 15%. We saw too many promotion-spiked products climbing rankings briefly. Redistributed 3% to value_for_money.
Feb 14, 2026 v2.3 — Added Croma + Reliance to user_reviews_signal calculation. Previously Amazon-only. Now weighted average across 4 marketplaces.
Nov 1, 2025 v2.2 — Added review velocity sub-metric (25% of user_reviews_signal). Catches products gaining momentum that raw rating averages miss.
Aug 12, 2025 v2.1 — Initial public release. Five-factor model with current weights and category schemas published.

FAQ

Why don't you physically test products?

Physical testing requires buying every product (capex ~₹3 lakh/quarter) + storage + a lab + writers who can compare. We chose a different bet: cross-marketplace verified data + category schemas + radical transparency. Wirecutter physically tests; we methodologically score. Different model, equally defensible — as long as we're honest about it. We are.

How do you handle paid affiliate relationships?

Every product link is an affiliate link. We earn a commission when you buy through it. The commission rate does NOT affect the Chayan Score. If we recommended only the highest-paying products, you'd notice — and so would Google. Our reputation is the asset; the affiliate commission is the rev-share. Full disclosure on the Disclosure page.

What's the difference between Chayan Score and Amazon's star ratings?

Amazon stars are raw and uncalibrated — and on Indian marketplaces almost everything sits at 4.0★+, so they barely separate good from great. The Chayan Score starts from the verified rating but recalibrates it for that inflation (a 4.0★ is only average here), then pairs it with a Chayan Confidence level so you know how much evidence stands behind it. A 4.5★ product with 30 reviews and a 4.5★ product with 30,000 reviews look identical on Amazon; on ChayanKart they don't.

Can I challenge a score?

Yes. Email the author named on the article. If your challenge identifies a data error or factor we missed, we update the score (and credit you in the changelog if you want). We have already adjusted 14 scores from reader feedback in the past year.

How often are scores updated?

Prices update weekly. Reviews update weekly. BSR updates daily. Full score recomputed weekly (every Monday). Major calibration quarterly.

Can a product score well and still carry a warning?

Yes. A high Chayan Score means the people who own it rate it highly — that's necessary, but not the whole story. Where public review threads surface a reliability or after-sales problem, we add a caution on the product even if its score is strong. The score measures buyer satisfaction; it doesn't erase a documented service or safety issue.

Disagree with a score?

Every article names its author. Email them. If you're right, we update. We've already changed 14 scores from reader feedback.

Contact the editors →
CK
Editor in chief, ChayanKart
Methodology owner · Last reviewed Jun 1, 2026