How we work out a refill interval

What we do, and what we do not

We do not test products. There is no lab, no test bench and no shelf of units we bought. We say so on every page, and we would rather publish that plainly than imply otherwise. Four of the nine competitors we looked at when researching this category put a testing claim in the page title; two of those had no author identity at all. We are not going to compete on that axis, because we would have to lie.

What we do is compile every published specification we can find, attribute each figure to whoever published it, and compute from those figures using inputs that are also published. Anyone can check the arithmetic, and anyone can substitute a different assumption and get a different answer. That is the point.

Input 1: the watering-frequency band

Horticultural extension services do not publish milliliters-per-day water-use rates for houseplants. What they publish is (a) the instruction to water thoroughly until the container drains, and (b) frequency bands by plant type.

Virginia Cooperative Extension’s publication Properly Watering Container Houseplants (Criscione and Holliday-Goulart, June 2026) gives succulents and cacti as needing water “once or twice a month” and tropical houseplants as “sometimes once or twice a week.” We use those verbatim, as ranges, and we never collapse a range to a single number.

Input 2: the volume of one thorough watering

This is the number we had to derive rather than find, and it is where our one assumption lives. We are setting it out in full so you can replace it.

Step 1 - soil volume from pot geometry. We model a standard tapered pot as a cylinder with a taper factor:

soil volume = pi x radius squared x height x 0.7, taking height as 0.9 times the diameter and 0.7 as the taper factor for a typical nursery-shape pot.

Step 2 - the assumption. We take one thorough watering as 20 percent of that soil volume. Virginia Cooperative Extension defines watering thoroughly as wetting the potting soil all the way through and letting the container drain until water stops dripping out the bottom. Twenty percent is our estimate of the applied volume that achieves that in a dried-down peat-based mix while leaving a little to drain.

It is a stated assumption, not a measurement, and we will not dress it up as one. Virginia Cooperative Extension’s own leaching-fraction publication notes that even the classic 10 to 20 percent leaching target comes from greenhouse production rather than living rooms, so precision here would be false.

A worked example

Take the Lechuza Classico Color 21, whose maker publishes a 5-liter reservoir, holding a 6-inch tropical root ball.

1. One thorough watering of a 6-inch pot: 350 ml, from the table above.
2. Waterings held: 5,000 divided by 350 = 14.3.
3. Published cadence for a tropical houseplant: once or twice a week, so 3.5 to 7 days per watering.
4. Refill interval: 14.3 x 3.5 = 50 days, to 14.3 x 7 = 100 days.
5. Published as: roughly 7 to 14 weeks.

Now the same calculation for a typical small watering globe, whose contents one publisher measured at about half a cup, roughly 118 ml, in the same pot: 118 divided by 350 = 0.34 waterings, so 1.2 to 2.4 days. That is the arithmetic behind our most-read page, and there is no interpretation in it.

How picks are selected

Stated criteria, applied consistently:

• The product is currently sold and available from a retailer we link.
• It has at least one published specification we can attribute, so there is something real to compare.
• It covers a distinct use case, so a roundup is a set of genuinely different answers rather than five versions of the same one.
• At least one pick in every roundup is something we tell you not to buy, with the reason.

Selection is never driven by commission rate or by which affiliate programs we have joined. Where a product with no published capacity outsells everything else in its category, we list it and explain that the number is missing rather than quietly omitting it - which is why 7 of the 30 products in our registry carry a “not published” row.

How the scores work

Each product carries an editorial score out of 10 built from four or five named metrics, which are visible on the card. The metrics are chosen per category - published capacity, computed interval, honesty of the manufacturer’s claim, value, and so on - and they are our judgment applied to published data.

They are not test results, and they are deliberately not marked up as ratings in the page’s structured data. Emitting a review rating for a product we have not tested would contradict the paragraph above it, and we would rather lose the rich-result eligibility than say two different things on the same page.

How prices work

Prices come from a live data layer refreshed against the retailer, and they carry the date they were captured. If that data is more than 48 hours old the number disappears automatically and the button reads “Check price on Amazon” instead.

We never show a price we cannot currently verify, and we never write a price into the text of a page. A stale number is worse than no number, because it looks like information.

The limitations, stated plainly

Every interval on this site is a planning range, not a guarantee. Virginia Cooperative Extension lists the things that change how much water a container plant needs: potting soil type, container type and size, time of year, plant species, temperature, humidity, light intensity and water type. We control for two of those - pot size and plant type - and not the rest.

Specifically, our numbers will be optimistic for a plant in a bright sunny window in July, and conservative for the same plant in a cool room in January. They assume a standard peat-based potting mix, an indoor environment, and a plastic or glazed container rather than unglazed terracotta, which breathes and dries faster.

The right way to use them is to compute, then verify: run whatever you buy for a full week at home and see what actually happens in your room, with your soil. One week of observation beats any calculation, including ours.

How we handle being wrong

We will be, at some point - a spec will be misread, a manufacturer will change a product, a link will rot. When that happens we correct the page and log it on the editorial policy page with the date.

If you have found something wrong, tell us and we will check the source. Corrections go to the front of the queue ahead of everything else.