Methodology
How we score gear
Five rubrics, every weight published, every specification linked to its source, and a plain account of how each page lands on its picks. If a score on this site disagrees with this page, this page is right and the score is a bug.
By Mike M. · Updated September 2026
What stands behind every score
The method is reproducible, and that is the point: we read the maker's product page and manual, quote its exact wording for the seal, the material, the ratings and the warranty, link every quote, do the arithmetic the maker didn't, cite the USDA and the FDA for every food-safety number, compare the models in a category side by side, and score the result on a rubric you can read below. Every figure on every product card can be checked by clicking it. At launch that is 96 products researched and compared and 616 sourced specifications.
The evidence rules
- Maker pages first. Every specification links to the manufacturer's product or support page. Where the maker's site blocked our automated reader, the maker's URL is still cited (it loads for a person) and the label says the wording reached us through the search index. Where a maker publishes nothing for a figure, we say "not stated" rather than fill it in.
- Retailer listings only when labeled. A retailer listing is cited only when the maker's page doesn't exist or doesn't carry the figure, and the source label says so.
- Food safety from the agencies. The 2-hour rule, 40°F, 3–4 days, 165°F, the two-ice-pack minimum, the freezer windows — quoted from USDA FSIS and FDA pages, linked on every page that uses them.
- Live prices only. The button on every pick checks today's price at Amazon. See below.
- Real products, sourced figures. Every product is a real, currently sold model chosen from published research, and every figure beside it is quoted from the document it came from and linked.
What the published score means
Every score on this site is published on a 8.4 to 9.8 scale, and that needs explaining, because a scale that starts at 8.4 is worth being suspicious of.
It starts at 8.4 because of what has already happened before a product gets a score at all. Everything listed here is a real, currently sold model that survived the research for its category: the makers that publish their seal wording, their thermal ratings and their warranties, against the ones that do not. Products that cannot clear that bar are not rescaled down the page — they are not on it. So the published band is a recommendation band, and the number inside it tells you how strongly, not whether.
- 9.5 – 9.8
- Category-defining
- The strongest published case in its category. Buy it unless a specific constraint on its card rules it out.
- 9.0 – 9.4
- Outstanding
- A clear recommendation with one trade-off worth knowing — usually size, format or price band.
- 8.4 – 8.9
- Excellent, strongly recommended
- Does its one job well. Often the right buy for a narrower need than the leaders serve.
Below 8.4 we do not recommend a product, which is why no score on this site sits there. If a pick stops earning its band — a model is discontinued, a seal claim is withdrawn, a listing goes dead — it is replaced with one that does, and the change is logged on the page.
Underneath, the arithmetic is the rubric below and nothing else. Each product is assessed 0–10 against every criterion in its category's rubric; those marks are combined using the published weights; and that single weighted result is expressed on the 8.4–9.8 band above by one site-wide mapping applied in the build. The mapping is strictly increasing, so it changes no ranking anywhere, and it is the same mapping for every product, so a score means the same thing on every page it appears on. Open "the rubric behind the score" on any product card to see the criterion marks on their own 0–10 scale.
Scores are editorial judgments built from documented research. The visible score on a page is content only — it never appears in structured data as a review rating, because it isn't one, and this site publishes no retailer star ratings or review counts.
The five rubrics
One per gear category, with every weight published. The overall is the weighted mean of the criterion marks, computed by the site's build and expressed on the published band, so no page can print a score that disagrees with the rubric behind it.
Containers
| Metric | Weight | How it is scored |
|---|---|---|
| Seal & leak design | 25 | The lid mechanism as the maker describes it (snap-lock tabs, silicone gasket, vent), and whether the maker itself calls it leak-proof or airtight. A gasketed four-latch lid scores highest; a press-on lid with no gasket scores lowest. |
| Material & thermal range | 25 | What the maker states it is safe for: microwave, oven (and to what temperature), freezer, and any thermal-shock caveat. Tempered or borosilicate glass with an oven rating scores highest; plastic with a microwave-only rating scores lower; an unstated rating scores lowest. |
| Lid longevity & replacement availability | 15 | Whether the maker sells replacement lids or gaskets, states a warranty on the lid, or says nothing. A published replacement-lid page or a stated lid warranty scores highest. |
| Stack / nest footprint | 15 | Whether the maker states the containers nest and the lids stack, and the published dimensions. Same-size sets that nest and stack flat score highest; mixed-size sets that do not nest score lowest. |
| Dishwasher & stain resistance | 10 | Dishwasher-safe rating (top rack or any rack), and the material's stated stain and odor resistance. Glass and Tritan score highest; polypropylene scores lower. |
| Value | 10 | What ships in the set for the price band, judged as a method, not a number: pieces, lid count, and whether the set covers a week of meals. Prices come only from the live layer. |
Food scales
| Metric | Weight | How it is scored |
|---|---|---|
| Resolution & capacity | 25 | The stated graduation (1 g or 0.1 g) and maximum load. A 0.1 g graduation with an 11 lb capacity scores highest; 1 g with 5 kg scores mid; unstated scores lowest. |
| Nutrition database / app sync | 25 | Whether the scale pairs to an app that tracks macros, the size of the food database the maker states, and whether it exports to MyFitnessPal or Apple Health. A basic scale with no app is scored on its own merits here (a 4 or 5, not a 0) because a phone app does the same job. |
| Tare & unit handling | 15 | One-touch tare, unit switching (g / oz / lb / ml / fl oz), and whether the scale can weigh into a bowl already on the platform. |
| Platform size & readability | 15 | Published platform dimensions, backlit display, and whether the display sits under a bowl's shadow. A large platform with a bright display scores highest. |
| Battery / charging | 10 | Rechargeable via USB, or replaceable AAA / coin cells, and any stated battery life or auto-off. |
| Value | 10 | Feature set against the price band, judged as a method. Prices come only from the live layer. |
Shaker bottles
| Metric | Weight | How it is scored |
|---|---|---|
| Mixing mechanism | 25 | How the maker says it mixes: a wire whisk ball, a plastic agitator, a molded vortex, or a motor. Judged on what the design does for a thick casein or a mass gainer, not just whey. |
| Leak / seal | 25 | Whether the maker calls the lid leak-proof, describes a gasket or a locking loop, and how the flip cap secures. A stated leak-proof claim with a locking cap scores highest. |
| Cleanability | 20 | Parts count, dishwasher rating (top rack or all parts), and whether the shape has ridges or a narrow neck. Fewer parts, all dishwasher safe, scores highest. |
| Capacity vs scoop math | 15 | Stated capacity against a two-scoop shake plus 12 oz of liquid with headroom to shake: 28 oz clears it, 20 oz does not. Measurement markings count. |
| Material / insulation | 15 | Tritan or polypropylene (BPA-free stated) versus 18/8 stainless steel, and whether the maker states an insulation rating in hours. |
Lunch bags & meal-management bags
| Metric | Weight | How it is scored |
|---|---|---|
| Container fit (count × size) | 30 | How many meal-size containers the maker states it holds, or the interior dimensions we can do the arithmetic on. A bag that states its container count scores highest; a bag that states only can-count scores from the math. |
| Insulation & ice-pack slots | 25 | The insulation the maker describes, any stated cold-hold time, and whether ice packs are included or have a dedicated slot. |
| Leak-resistant liner | 15 | Whether the maker states a leak-resistant or heat-welded liner, and whether it is wipe-clean. |
| Carry / durability | 15 | Fabric denier if stated, shoulder strap or backpack straps, and any stated warranty. |
| Cleanability | 15 | Removable liner, wipe-clean interior, or a machine-washable claim. |
Bulk-cook & freezer gear
| Metric | Weight | How it is scored |
|---|---|---|
| Batch capacity | 30 | The stated capacity in the terms a meal-prepper needs: cooked cups of rice, quarts of chili, pounds of chicken, or bag width for a sealer. Bigger scores higher, up to what fits in a home kitchen. |
| Hands-off cooking | 20 | Timer, delay start, keep-warm, and preset programs as the maker lists them. A keep-warm that lets you sleep on it scores highest. |
| Cleanup | 20 | Removable inner pot or liner, nonstick or enamel surface, and dishwasher rating on the removable parts. |
| Footprint | 15 | Published dimensions and weight against a counter or a cabinet. Compact scores higher unless the capacity metric justifies the size. |
| Value | 15 | What the unit does for its price band, as a method. Prices come only from the live layer. |
How a page's picks are chosen
Roundups are ordered by the author's pick for the page's stated buyer, first to last, with each product's rubric score shown beside it. The two can disagree: the rubric scores a product against its whole category, while the order answers the page's narrower question, and a product that scores higher on the rubric can be the wrong buy for that page (a five-container set with a lifetime lid warranty scores above a nine-container set, and still isn't the one that covers a week). Where they disagree, the per-page "best for" label and the pick note say why, and the score stays exactly what the rubric produced. Comparisons carry every product compared. Guides carry two to four picks placed after the section that explains what to look for, so the recommendation is earned rather than leading. Every commercial page discloses its affiliate links above the first one. No pick is ever chosen or moved for a commission; Amazon pays the same rate whichever product you buy in a category, which removes the incentive, and we'd say so if it didn't.
How prices work
Every buy button links to Amazon through this site's tagged link and reads "Check price on Amazon" until a live price exists. When a live price layer is enabled, the number renders inside the button with the date it was checked, and it is removed automatically once the data is older than 48 hours — a stale price never sits on a page. Right now: live prices are on, last checked November 5, 2026. Every price you see on this site comes from the retailer on the date shown, never from copy written months ago.
How pages get re-evaluated
Roundups and comparisons: every pick's availability and spec page re-checked quarterly, discontinued models swapped, the page's modified date bumped only when a pick, score or spec actually changes. Guides: sources re-verified twice a year. Corrections: on report, within seven days, noted at the foot of the page. A price refresh never changes a page's modified date. The rules for independence and corrections are on the editorial policy; the funding statement is on the affiliate disclosure.