You cannot choose a water filter sensibly before you know what you are filtering. Certifications are issued for named contaminants, not for “clean water” in general, so the decision runs in one direction only: find out what is in your supply, then buy the certification that covers it.
Doing it the other way round — picking a filter and hoping — is how people end up with a carbon pitcher certified for chlorine taste, believing it handles lead.
Step one: read your water quality report
If you are on a public water system, your utility is required to publish an annual water quality report, usually called a Consumer Confidence Report. It is free, it arrives or is published each year, and it lists the contaminants detected in your supply along with the levels found and the regulatory limits.
It is the single most useful document in this entire subject, and almost nobody reads it. What to look for:
- Anything detected above, or close to, its limit. These are your filtering priorities.
- Disinfectant type. Chlorine and chloramine behave differently in a filter — chloramine is more stubborn and needs a longer carbon contact time, so a filter certified for chlorine is not automatically adequate for chloramine.
- Hardness. Decides whether you have a scale problem, which is a softener question rather than a filter question.
- Any PFAS results. Increasingly reported, and the reason behind a growing share of filter purchases.
- Lead. With an important caveat, below.
The lead caveat that the report cannot cover
A utility tests water as it leaves the plant and at sample points in the distribution system. Lead usually does not come from there. It comes from the service line into your building, and from solder and fixtures inside it.
So a report showing no lead does not tell you your tap water has no lead — it tells you the supply does not. If your building predates the mid-1980s, or you do not know what your service line is made of, the only way to know is to test at your own tap. Many utilities keep service line material records and will tell you; some run free or subsidized testing programs.
This matters because lead is one of the contaminants where filtration is genuinely the answer, and where the certification is specific: a NSF/ANSI 53 claim naming lead.
If you are on a private well
Nobody tests it for you, and the report does not exist. A certified laboratory test is the starting point, and the panel should cover bacteria, nitrate, arsenic, and whatever is locally common — your county or state health department can usually tell you what that is. Treatment selection for a well is a different exercise from municipal filtration, because the risk profile includes microbiological contamination that treated supplies have already addressed.
Step two: match the contaminant to the standard
| What you found | What to look for |
|---|---|
| Chlorine taste or smell | NSF/ANSI 42, chlorine or taste-and-odor claim |
| Chloramine | NSF/ANSI 42 with a chloramine claim specifically |
| Lead | NSF/ANSI 53, lead claim named |
| Cysts (Cryptosporidium, Giardia) | NSF/ANSI 53, cyst reduction claim |
| VOCs | NSF/ANSI 53, VOC claim |
| PFOA / PFOS / Total PFAS | NSF/ANSI 53 (carbon or anion exchange) or 58 (RO), with the PFAS claim named |
| Nitrate, arsenic, fluoride, high TDS | NSF/ANSI 58 — reverse osmosis |
| Bacteria or viruses (untreated source) | NSF/ANSI 55 Class A ultraviolet, or NSF P231 |
| Hardness and scale | NSF/ANSI 44 — a softener, not a filter |
| Sediment, rust, visible particles | NSF/ANSI 42 particulate claim, by particle class |
Notice the pattern. Carbon handles chlorine, taste, VOCs, many organics and — with the right medium — lead and PFAS. Reverse osmosis handles dissolved inorganics that carbon cannot touch: nitrate, arsenic, fluoride, total dissolved solids. Ultraviolet handles organisms and nothing chemical. No single technology does all three, which is why combination systems exist.
Step three: choose the form factor by where the water is used
Capability and convenience pull against each other here, and the right answer depends on how much water needs treating.
| Form | Suits | Limitation |
|---|---|---|
| Pitcher | One or two people, drinking water only, renting | Small cartridge, slow, limited claim lists, frequent replacement |
| Faucet-mounted | Drinking and cooking at one tap, no plumbing work | Low flow, bulk at the tap, not every faucet fits |
| Under-sink carbon | Households wanting good flow and longer cartridge life | Needs a cabinet and a connection; usually a separate tap |
| Under-sink reverse osmosis | Dissolved inorganics — nitrate, arsenic, fluoride, TDS | Produces wastewater, slow, needs a tank and a drain connection |
| Whole-house (point of entry) | Sediment, chlorine, hardness for the entire building | Large, plumbed, not the right place for fine drinking-water claims |
| Refrigerator inline | Convenience at an existing dispenser | Claim lists vary enormously; check the specific cartridge |
A common and sensible arrangement is a whole-house sediment and chlorine filter for the building, plus a certified under-sink unit at the kitchen tap for drinking water. The coarse work happens once, and the expensive certified medium is only treating the water you actually drink.
Step four: check flow rate and capacity, not just the claim
Two specifications decide whether a filter stays effective in use.
Rated capacity, in gallons, is how much water the cartridge was certified to treat while still meeting its claims. Past that point the claim no longer holds, even if water still flows. Divide your household’s drinking and cooking consumption into the rated capacity to get a real replacement interval, then compare it with the manufacturer’s stated interval — whichever is shorter wins.
Rated flow rate matters because contaminant reduction depends on contact time with the medium. Running a filter faster than its rated flow reduces its effectiveness, which is a real risk with under-sink systems plumbed to a full-pressure tap.
And bear in mind the trade-off: a tighter filter achieving more claims runs slower. Filling a kettle from a high-capability under-sink filter takes longer than from the cold tap, and that is the mechanism working rather than a fault.
What filters do not fix
Three honest limits.
A filter past its capacity can be worse than none. Spent carbon can harbor bacterial growth, and an exhausted cartridge may release accumulated contaminants. Replacement intervals are a safety matter, not a performance suggestion.
Hot water taps are not filtered by most systems, and carbon filters are generally specified for cold water. Fill from the cold tap and heat it afterward.
A filter cannot fix a plumbing problem. If the lead is coming from your service line, a certified point-of-use filter protects the tap it is fitted to — not the bathroom tap, not the shower, not the ice maker. Replacing the line is the actual fix, and filtration is the interim measure.
The order to decide in
- Read your water quality report. Free, and it answers most of the question.
- Test at your own tap for lead if the building is older or the service line is unknown.
- List the contaminants you actually need to reduce. Usually one to three, not twenty.
- Find the matching standard and named claim, and check it on the certifier’s own listing.
- Choose the form factor by where the water is used and how much of it.
- Check rated capacity and flow, and work out your real replacement interval.
- Price the cartridges over three years. This is where the cost of a filter lives.
Related guides
- NSF/ANSI 42 vs 53 vs 401: which certification covers what
- Pitcher, faucet, under-sink or RO: choosing by contaminant
- Reverse osmosis: what NSF/ANSI 58 covers and what it costs you
Sources: NSF, NSF/ANSI 42, 53 and 401: Filtration Systems Standards; NSF, Standards for Water Treatment Systems, for the standard-to-technology mapping including NSF/ANSI 44, 55, 58 and 62; NSF, Forever Chemicals and the Advancement of Filtration Standards, for which standards carry PFAS claims and which technologies are used. The annual water quality report is a requirement on public water systems under the Safe Drinking Water Act; contact your utility for your own. We do not test filters, and the form-factor table reflects the design characteristics of each type rather than measured results.