Methodology · v1.0 · published 20 Aug 2026
How the Meal Prep Score works
The score exists to answer one question: how good is this service right now? Not in 2022, not on its best day — now. This page is published before any score is, and every change to it will be versioned and dated here.
1 · What goes in
| Input | What we record | Cadence |
|---|---|---|
| ProductReview.com.au | Aggregate star rating and total review count, as displayed on the brand's listing, with the capture date. Shown on-site as "X★ from N reviews on ProductReview.com.au, as at [date]" with a link to the listing. | Weekly snapshot |
| Google reviews | Aggregate rating and review count with capture date. | Weekly snapshot |
| Trustpilot | TrustScore and review count with capture date. | Weekly snapshot |
| YouTube reviews | Recent independent review videos — embedded on the brand page, weighted by recency and viewership, summarised in our own words. | Monthly sweep |
| First-hand testing | We order the meals, weigh them, verify labels against claims, record delivery condition and days-to-use-by on arrival, and photograph everything. Scored against a fixed rubric. | Rotating cycle |
| Pricing & macros | Current per-meal pricing and headline macros from the brand's own site, checked and date-stamped. | Monthly verification |
All third-party figures are facts about those platforms at a point in time — they remain those platforms' data, are always attributed and dated, and we never present them as our own rating. Our structured data (the machine-readable markup search engines use) only ever contains our own editorial assessments, never third-party ratings.
2 · How it's combined
- Recency weighting. Source snapshots decay exponentially with age (half-life: 12 months), so the composite tracks the brand's current performance. Alongside the headline score we show a last-90-days sub-score and the trend direction.
- Volume weighting. A platform's contribution scales with its review count for that brand, so 3,000 reviews outweigh 30.
- Editorial layer. Our first-hand test rubric contributes a fixed 25% of the composite. Rubric results are published on the brand page with photos.
- The exact formula, weights and any changes are published on this page. Score movements appear in each brand page's changelog ("+214 reviews ingested this week; score moved 7.7 → 7.9").
3 · The data threshold
A brand gets no published score until we hold: at least 200 aggregated reviews across sources, at least two consecutive weekly snapshots, and at least one completed first-hand test. Until then the brand is listed as "Tracking". We never publish placeholder or estimated ratings.
4 · What never influences a score
- Commissions. Whether a brand pays us a referral commission — and how much — appears nowhere in the formula. Brands cannot pay to improve, suppress or fast-track a score.
- Our parent company's brands. Foober and Vegan Soul are scored by the identical rubric and pipeline, are badged everywhere they appear, and are permanently excluded from ranked "best of" lists.
- Advertising or featured placements. If we ever run them, they will be labelled, fixed in position, and excluded from rankings.
Corrections
Found an error — a stale price, a wrong count, a misattributed figure? Contact details are on the About page. Corrections are applied to the page with a dated note in its changelog.