This page is the most important one on the site. It tells you exactly what our verdicts are built from — and, just as importantly, what they are not.
We do not test products ourselves
We want to be blunt about this, because a great many sites in this niche imply otherwise. SolarGenLab does not own a bench, a load tester, or a pyranometer. We have not personally discharged a power station or measured a solar panel’s output in the field. Any site telling you it “tested” thirty power stations this year, complete with identical stock photography, is very probably telling you a story.
What we do is harder to fake and, we would argue, more useful.
What our verdicts are actually built from
1. Independent bench measurements
We collect the published, instrumented results of testers who genuinely do own the equipment — outlets like OutdoorGearLab, The Solar Lab, StorageReview, TechRadar and Trusted Reviews. We record the measurement, the number, and who took it. When we tell you the EcoFlow DELTA 3 Plus surged to 2,600W against a claimed 3,600W, we will tell you who measured it.
2. Manufacturer specifications — treated as claims, not facts
Every figure a brand publishes goes into our database labelled as a claim until somebody independent has verified it. Cycle life is the clearest example: no one has ever cycled a portable power station 4,000 times and published the result. Every cycle-life number you have ever read, on any site, is a datasheet claim. We will not present it as anything else.
3. Verified owner reviews, at volume
We read hundreds of verified purchaser reviews per product and look for the complaint that repeats. One person with a loud fan is noise. Seventy-eight people in one thread describing the same solar input dropping in and out every five seconds is a defect. We only report a criticism when the pattern is real, and we tell you how many owners it came from.
4. Documented contradictions
We read the manufacturer’s own pages closely, and we report it when they contradict themselves — a weight given two different ways in the same spec sheet, an expansion capability the marketing promises and the FAQ denies, four different charging times for one product. These are not gotchas. They are a signal about how carefully a company treats you.
What we will never do
- Invent a number. If a spec is not published, the field reads UNKNOWN. We do not infer it from a similar model.
- Claim first-hand testing. Not in a headline, not in a caption, not by implication.
- Publish a decibel figure we cannot source to a named human who measured it. Most of the noise ratings circulating online trace back to AI-generated affiliate pages. Manufacturer dB claims almost never state the load or the distance, which makes them meaningless.
- Write a puff piece. Every product we recommend gets its documented criticism printed alongside the praise. If a product has no recorded flaw, that usually means nobody has tested it yet — and we will say that instead.
- Recommend a discontinued product because it still pays a commission.
Scores and evidence counts
Where we give a score, it always appears with the evidence behind it — for example, 8.4, from 412 verified owner reviews and two independent bench tests. A score without its evidence count is just an opinion with a decimal point. You will not find one here.
Corrections
Prices move, firmware changes, and products get quietly discontinued. When we find we were wrong, we correct the article and note it. If you spot an error before we do, please tell us — we will fix it and credit you if you would like.