How we rank shower filters

Every competitor ranking above us in this category leads with something we do not have. WaterFilterGuru has a certified water specialist, test kits and a colorimeter. Forbes has testers using units for months. Interior Medicine has a physician byline. We have none of that, will not claim to, and will never publish a “we tested N filters” line.

Attempting to imitate it is the one move guaranteed to fail, because it would be false and because it is trivially falsifiable. So we compete on something else: a method built entirely from sources you can check yourself.

What we do

1. Verify certifications against the certifying body. A certification names a standard and the organization that issued it, and it appears in that organization’s own product listing. NSF, IAPMO R&T and WQA all maintain searchable directories. If we cannot find a named body, we record the product as uncertified regardless of what the box says.
2. Compile published specifications, with a source on every figure. Media, flow rate, cartridge life, format and anything published about chloramine, taken from the manufacturer’s own page or listing. Our content model makes this mechanical: a specification is either a value with a source URL, or an explicit “not published”. There is no third state, so an unsourced figure is a compile error rather than a judgment call.
3. Credit third-party lab data by name. Where an independent party has measured something, we report it and link it and say whose it is. Citing a competitor’s lab results by name is a trust signal, not a weakness — it proves we did not invent the number.
4. Do the cost arithmetic in public. Purchase price plus cartridges per year times cartridge price. Every roundup in this category publishes a purchase price and none computes the year. The method.
5. Record what each product does not do. Mandatory on every product in our registry. A pick with no stated limitation has not been thought about.

What we do not do

• We do not test filters. We own none of them and operate no lab. Units we claim to have tested: 0.
• We do not hold any credential in water treatment, and we do not employ anyone who does. No CWS, no CWR, no WQA certification, no clinician, no expert panel.
• We do not accept free products or sponsored placement. Sponsored placements: 0. Free products accepted: 0.
• We do not publish a score. Our data model has no score field, deliberately. A number would imply an evaluation we did not perform, and no page on this site emits a Review or rating in its structured data either.
• We do not print a price we cannot verify live. Prices render from a live retailer feed or not at all. Where there is no current figure the buy button reads “Check price”.

How a ranking is built

In this order. Later criteria only break ties among products that have cleared the earlier ones.

1. Traceable certification. The only independent check this category offers. Of the 15 products in our registry, 4 carry one.
2. Media matched to the reader’s disinfectant. A KDF filter on chloraminated water is the wrong tool however well it is made, so this can outrank certification on a page written for chloramine.
3. Independent measurement, where any exists, credited to whoever produced it.
4. Published flow rate. If a maker will not say how much water comes out, you cannot compare the product or check it against your state cap.
5. Cartridge cadence and capacity, because a filter nobody re-cartridges stops filtering around month eight.
6. Format and fit — inline versus replacement head, clearance, handheld compatibility.

Where our numbers come from

On citing brand-owned sources

Some of the clearest technical writing in this category is published by companies that sell filters. Where we cite one — for contact-time figures or media chemistry — we say in the citation label that it is brand-owned, because a filter maker explaining filter chemistry has an interest and you are entitled to weigh that yourself rather than have us hide it.

How we handle disagreement

When sources conflict, we show both and say so rather than picking the one that suits the recommendation. The clearest example on this site is the Sprite Slim-Line 2: it holds a genuine NSF/ANSI 177 certification from IAPMO, and an independent lab measured 42 percent chlorine reduction on their own supply. We publish both figures and explain why they are not in conflict.

What this method cannot do

Stated plainly, because a method with no limits is a sales pitch.

• It cannot tell you how a filter performs on your water. Nothing in published specification can.
• It cannot verify a manufacturer’s testing claim. Where a maker says it was independently tested and names no body, we report the claim and cannot check it.
• It cannot fill a gap. Several major brands publish no media list, no flow rate and no capacity. We record that and do not substitute a guess, which makes those entries thinner than we would like and honest.
• It cannot assess build quality, spray feel or durability, all of which need a unit in a hand.

Update cadence

• Prices — daily, automated through the price layer.
• Certification status — re-verified quarterly against the certifying bodies’ listings.
• Specifications — the current compile was run on 2026-09-01; re-checked when a product page changes or a reader reports a discrepancy.
• Rankings — every six months, or immediately if a certification lapses or a product is discontinued.
• Evidence pages — annually, or immediately when new published research changes a conclusion.

dateModified changes only when the content actually changes — never on the daily price redeploy. A site that claims every page changed every day trains search engines to stop trusting its dates.

How to tell us we are wrong

Specifications change, listings get edited and certifications lapse. If a figure here no longer matches the manufacturer’s page, or if we have read a source incorrectly, tell us. Corrections are made promptly and logged with the date on the editorial policy page. Written by Scooter M., who is an enthusiast rather than an expert, and says so.