How we pick

A published method is the only honest alternative to a test lab we do not have. Here is exactly how the scores on this site are produced.

Start with what we are not

We do not run a test bench. We have not pumped water through these units, measured their flow, or run them to failure. A site claiming hands-on testing across a catalog of 56 products is describing something that did not happen, and we are not going to join in.

What we can do — and what nobody in this category seems to bother with — is read the manufacturers' published data carefully, compare like with like, cite every figure, and be explicit when the data does not exist.

The five criteria

Pump roundups score on five dimensions, weighted roughly equally. Other categories use criteria adapted to what matters there — a check valve is not scored on flow at lift — but the structure is the same and the criteria are printed on the page.

1. Flow at lift. Taken from the manufacturer's published curve at 10 feet where one exists. Not the zero-foot headline figure, which describes a pump lying on a bench.
2. Build. Housing material, impeller type, seal design, and how the pump handles sustained duty.
3. Switch. Weighted heavily, because the switch is what fails. Sealed designs and redundant floats score higher than exposed tethers.
4. Serviceability. Quick-disconnect cords, replaceable switches, parts availability, and whether a repair means replacing the whole unit.
5. Value. The first four assessed against the product's current street price band — not against a number we publish, because we do not publish prices.

What the scores are

Editorial judgments on a ten-point scale, derived from published specifications and from how a product compares with its direct alternatives. They are a way of making our reasoning legible and comparable, not measurements.

That is also why no score on this site appears in structured data. Marking an editorial opinion up as a product review rating would tell search engines we tested something we did not, and we will not do that even where it would help us.

Where the numbers come from

• Manufacturer documentation first — product pages, specification sheets, installation manuals.
• Retail listings second, where the manufacturer does not publish a figure.
• Nothing third. Where no source exists, the figure is omitted and the page says so. Of the 56 products in our registry, 56 carry a cited source and 13 have a published flow figure with the lift it was measured at.

The content model enforces part of this mechanically. A flow rate cannot be stored without its lift. A product cannot carry a price field at all. An invented part number is unrepresentable because the registry has no field for one. Those are deliberate constraints: a rule that only exists in a style guide eventually gets broken.

The “skip this if” rule

Every product entry on this site carries a skip this if line, and the content model will not compile without one. The reason is straightforward: a recommendation that names nobody who should avoid the product is not a recommendation, it is an advertisement.

How money affects this

The site earns affiliate commissions. That creates an obvious incentive to recommend expensive things, and the honest answer is not to deny the incentive but to describe what we do about it: ranking is by fit for the stated job, we say clearly when the cheaper product is the better buy, and no manufacturer has ever been consulted about anything on this site. Full detail on the affiliate disclosure and in the editorial policy.

When we get it wrong

We will. Specifications change, products get discontinued, and manufacturers revise their documentation. Tell us at the contact page and the correction gets made and noted. The update date on every page is the real last-modified date, not a figure refreshed to look current.