SiftScore Methodology
The exact, public formula behind every score on this site. Nothing is hidden — you can reproduce any score by hand.
The formula
Every product receives a score from 0 to 100 as the weighted sum of five components:
| Component | Weight | What it measures |
|---|---|---|
| Value | 25% | Price compared with the category median and the refund window. |
| Trust & legitimacy | 25% | Documented refund policy minus red flags (auto-ship, fake countdowns, hidden subscriptions…). |
| Evidence | 20% | Published ingredient lists and ingredients with cited study evidence. |
| Market reception | 15% | Network marketplace gravity and the latest known rating. |
| Transparency | 15% | Price, refund and description published clearly, minus red flags. |
Total = 25%×Value + 25%×Trust + 20%×Evidence + 15%×Market + 15%×Transparency
Component rules
- Value — base 50. Price within ±30% of the category median: +20; within ±60%: +10; otherwise 0. Refund: ≥180 days +30, ≥90 +20, ≥60 +10, >0 +5, unknown 0.
- Trust & legitimacy — base 50. Refund policy documented: +20. Each red flag: −15.
- Evidence — unknown data stays at neutral 50. With data: list published +20, each named ingredient +5 (cap +20), each ingredient with cited evidence +20 (cap +60).
- Market reception — base 50. Gravity ≥200 +25, ≥100 +15, ≥50 +10, >0 +5. Rating ≥4.5 +25, ≥4.0 +15, ≥3.0 +5.
- Transparency — base 50. Price documented +20, refund documented +20, description published +10. Each red flag −10.
All components are clamped to 0–100.
Red flags
- Auto-ship / negative-option clause
- Fake countdown timers
- No refund policy
- Undisclosed subscription
- Exaggerated claims
The unknown-data rule
Missing data never boosts a score and never invents a value. Components with no data sit at a neutral 50 and the page shows "unknown" — we would rather show a lower, honest score than guess.
Live worked example
Real calculation for Exipure using current data:
| Component | Weight | Score | Contribution | Breakdown |
|---|---|---|---|---|
| Value | 25% | 100.0 | 25.00 | Price vs category median (+20) · Refund window (+30) · |
| Trust | 25% | 70.0 | 17.50 | Refund policy documented (+20) · |
| Evidence | 20% | 50.0 | 10.00 | No ingredient data — neutral (+0) · |
| Market | 15% | 50.0 | 7.50 | Market gravity (+0) · Rating (latest known) (+0) · |
| Transparency | 15% | 100.0 | 15.00 | Price documented (+20) · Refund documented (+20) · Description published (+10) · |
| Total | 100% | 75.0 / 100 | ||
When scores are recalculated
Scores are recomputed automatically on every data refresh (daily pipeline, manual re-fetch, imports) and cached on the product at write time. The example above is always computed live from the same code that produces stored scores.