Methodology
How We Test
Every verdict on this site comes from a published protocol: identical inputs, recorded measurements, and the assumption list printed next to the results, so you can rerun our thinking with your own numbers.
The house rules
We buy everything we test, at retail, through the same listings you would use. No samples, no loaners, no preview access, and no manufacturer sees a verdict before publication.
Identical inputs across machines. A comparison is only as honest as its constants, so the same food, the same preparation, and the same loading go into every machine in a category, on the same days when weather allows.
Measurements over adjectives. Wet and dry weights, cycle times, seal integrity at fixed intervals, oil clarity on a cadence: every claim on this site traces to a number in a log, and the assumptions behind every number are printed with it.
Honest gaps are published as gaps. When a listing does not state a spec, we say so rather than estimating. When a listing contradicts itself, we record the contradiction. When we cannot re-verify a figure on update day, we say that too.
How scores work
Our 0 to 100 score is weighted toward what affects food: batch completeness and seal integrity carry the most weight, then cycle times and capacity reality, then build, documentation, and parts availability, then owner consensus from verified ratings. The exact weighting is stated in each guide's methodology note, because a score without its weighting is just a decoration.
A score is a snapshot of a machine at a moment, tied to the verification date printed beside it. Machines get rescored when the field changes, when we retest, or when long term ownership data changes the picture, and the updated date on every page tells you which snapshot you are reading.
The protocols, by category
Freeze dryers
- 1Three standard foods run through every machine: fresh strawberries, blanched sweet corn, and cooked chicken thigh, prepared identically and loaded to 80 percent of tray capacity.
- 2Cycle time logged from loaded to finished, with finished defined as no visible ice in cut samples from tray centers, the last place to dry.
- 3Every batch weighed wet and dry on a calibrated 0.1 g scale; dry mass ratio is our cross machine comparison metric.
- 4Cold trap pull down verified with a probe where listings publish a temperature, and scored indirectly through pump oil clarity where they do not.
- 5Pump oil inspected weekly and rated for clarity, because oil condition is where shallow cold traps and incomplete batches show up first.
- 6Every manual read, every support channel contacted with a standard question, and every listing contradiction recorded as itself.
Vacuum sealers
- 1Three hundred bags per machine across three series: dry oats, moist marinated vegetables, and wet brined pork, in each machine's own recommended bags.
- 2Seal integrity checked at 24 hours and again at one week, by feel and by submersion bubble test, at room and freezer temperatures.
- 3Consecutive cycle endurance runs with manufacturer mandated rest, recording thermal behavior and cycle consistency.
- 4Bag economics priced per seal at verified prices and amortized at realistic weekly volumes, because consumables are half the ownership cost.
- 5Ratings and counts extracted live from each product's own review widget, never from search snippets or carousels.
Grain mills and future categories
- 1Every new category ships with a published protocol before its first guide, so the methodology precedes the verdicts rather than excusing them.
- 2Mills are evaluated on texture range against real recipes, effort per cup for hand units, thermal behavior for electrics, and parts availability for the decade test.
- 3New categories inherit the house rules: identical inputs, recorded measurements, published assumptions, and retest dates.
Retesting and updates
Guides are re-verified on a rolling schedule and every page shows its last verified date. Prices, ratings, and availability are checked at each update and the date is printed with them, because a price without a date is a rumor.
When we get something wrong, the correction is published on the page with a note, not applied silently. If you find an error, write to us and you will see your fix credited on the record.