1. We buy everything we test
We do not accept review units, sponsored placements, or "send us one and write a post" deals. Every kit, light, meter and pump that appears on this site was purchased anonymously through the same retailers our readers use — usually Amazon UK, Amazon US, or the manufacturer's own web store. We keep receipts and link to them in our internal test log.
2. The standardised test bench
Each new system runs through an identical eight-week cycle in our test space, a 4 m² spare room in north-east London held at 20–23 °C and 45–55% relative humidity. Everything is plugged into a metered smart plug so we can record real watt-hours rather than spec-sheet claims.
- Seeds: Genovese basil, buttercrunch lettuce, curly parsley — the same batch for every system in a comparison.
- Water: Local tap (typically 280–320 ppm) unless the manufacturer specifies otherwise.
- Nutrients: Whatever the system ships with for the first cycle; General Hydroponics MaxiGro for the second cycle to normalise.
- Light: The unit's built-in lights only — no supplemental lighting unless we are reviewing a light.
3. What we actually measure
- Setup time from sealed box to seeds in water, timed with a phone stopwatch.
- Germination rate per pod at day 7 and day 14.
- Total harvest weight at week 8, weighed on an Etekcity 0.1 g scale.
- Noise at 30 cm with a Reed R8050 sound-level meter, day and night.
- Energy use with a TP-Link Tapo P110 over a full 14-day cycle.
- PAR / PPFD at canopy height with an Apogee MQ-500 quantum sensor for lighting reviews.
- pH and EC drift between top-ups with an Apera PH20 and Bluelab Truncheon.
4. How we score
Every system gets a 100-point internal score split across five categories: ease of setup (20), yield (25), reliability over the eight-week cycle (20), running cost (15), and design / build quality (20). We publish the rankings, not the raw scores, because we found readers care more about "should I buy this" than about a number to two decimal places.
5. The retest rule
Any system we recommend gets a fresh second cycle six months later. If it fails differently — a pump dies, the LEDs dim noticeably, the app stops talking to it — we update the article and add a dated "Long-term update" note rather than quietly editing the original copy.
6. How we choose what to test
We prioritise the systems readers actually ask us about in email and in the comments, plus anything new that crosses 1,000 verified Amazon reviews in its first six months. We do not test every product in a category — we test the ones a typical apartment grower is most likely to find when they search.
7. How we make money
Most of our recommended products link out through affiliate programmes, primarily Amazon Associates and a handful of manufacturer partners. We earn a small commission only if you choose to buy — and only after our review verdict is locked. Read our full affiliate disclosure for the legal version.
Spotted something we got wrong?
We are a small independent publication and we do make mistakes. If a number looks off or a product has changed since we tested it, please tell us — we will rerun the test and update the article.