How SiftScore Works in Practice

Every score on TrustSift is produced by a published formula — not by an editor's opinion. This article walks through the exact calculation with a concrete example so you can verify any score yourself.

The five components

A product starts with five component scores, each clamped between 0 and 100:

1. Value (25%) — price compared to the category median, plus the refund window.
2. Trust & legitimacy (25%) — documented refund policy, minus red flags.
3. Evidence (20%) — published ingredient lists and cited study evidence.
4. Market reception (15%) — marketplace gravity and the latest known rating.
5. Transparency (15%) — price, refund and description clearly published, minus red flags.

The unknown-data rule

The most important rule: unknown data sits at a neutral 50 — it never boosts and never penalizes. A product with zero verified data scores exactly 50.0, and the page says "unknown" instead of guessing.

A worked example

Take a product priced at $49 in a category with a median of $50, offering a 60-day refund, with one published ingredient:

• Value: base 50. Price is within ±30% of the median (+20). Refund of 60 days (+10). Value = 80.
• Trust: base 50. Refund documented (+20). No red flags. Trust = 70.
• Evidence: list published (+20) + one named ingredient (+5) = 25.
• Market: no gravity or rating data → neutral 50.
• Transparency: price (+20), refund (+20), description (+10) = 100.

Total = 0.25×80 + 0.25×70 + 0.20×25 + 0.15×50 + 0.15×100 = 65.0.

How red flags change the math

Each red flag (fake countdowns, auto-ship clauses, hidden subscriptions…) removes 15 points from Trust and 10 from Transparency. One flag turns a 65.0 into a 59.8; two flags flip a legitimacy verdict to avoid.

Try it yourself

Open the methodology page — it renders the same constants the scoring engine uses, with a live worked example computed from real current data.