How We Decide What to Recommend

The short version

We have not tested these products ourselves and we will never say that we have. Every recommendation on Mesh & Signal is built from three inputs: the manufacturer's published specifications, measurements taken by independent outlets that do operate labs, and structural reasoning about what a particular house or symptom actually requires. Where a number appears, it is attributed and linked. Where we reach a judgment, it is labeled as a judgment.

That is a genuine limitation. It is also, we think, a more honest and more checkable position than the alternative — because the reasoning on these pages can be verified by you, whereas an unverifiable lab claim can only be trusted.

What we do, step by step

1. Establish the problem before the product. Almost every purchase decision in home networking is really a diagnosis. We work out which of four categories a symptom belongs to — coverage, congestion, capacity or reliability — because each needs a different purchase and buying for the wrong one is the most common expensive mistake in this category.
2. Verify the product is currently sold. Every model named on this site was confirmed as a current product at the time of writing, via live research rather than memory. The evidence trail is recorded per product.
3. Read the published specifications. Band count, channel widths, port count and port speed, coverage claims, device limits. We treat manufacturer coverage figures as comparisons between that brand's own models, not as promises about your house — they are measured in conditions nothing like a real building.
4. Read independent measurements, and cite them. Outlets with labs — RTINGS, Tom's Hardware, Tom's Guide, Dong Knows Tech, TechRadar, Engadget and others — do work we cannot. We link their pages rather than reproducing their numbers as bare figures, because a throughput number torn out of its test conditions is worse than no number.
5. Apply the structural reasoning. This is where our value actually is. Backhaul type matters more than headline speed; range has not improved across recent Wi-Fi generations; metal in a wall is worse than thickness; a wireless relay halves throughput. These are mechanisms rather than opinions, and they change which product is right far more than any benchmark does.
6. Score against a visible rubric and publish it next to the pick, so you can disagree with our weighting rather than only with our conclusion.

The scoring rubric

Every roundup on this site uses four metrics chosen for that specific page — a page about basements weighs "beats masonry" heavily and ignores device capacity; a page about apartments does the reverse. Scores are out of ten and the overall figure is the mean of the four.

That last row is deliberate and worth explaining. We show a visible score because it is a useful summary. We do not emit it as machine-readable review or rating markup, because doing so would present an editorial assessment as a review rating of a product we have not tested — which would contradict this page.

Where prices come from

From a live price data layer, or not at all. If there is no verified current price for a product, the button on this site says "check price on Amazon" and opens the retailer's own listing. It never shows a remembered or stale figure.

This site currently runs without live price data, which is why you will not see dollar figures on the product cards. When a price does appear, it carries the date it was verified, and prices older than 48 hours disappear automatically rather than quietly going stale. Where we quote a price in prose — a rental fee, a price bracket — it is cited to a named source and dated in the text.

Where product images come from

Licensed stock photography, and we label it. We have no entitlement to manufacturer or retailer product imagery, so the photographs beside our picks illustrate the category — a mesh node on a shelf, a plug-in adapter in a socket — rather than showing the specific unit. Every such image carries an "illustrative" label in the interface and alt text describing the actual photograph. The alternative, using an unlicensed image or implying a stock photo is the product, is not something we are willing to do.

Forbidden phrasings

These never appear on this site, and if you find one it is an error we want reported:

• "We tested", "in our lab", "we measured", "our benchmarks"
• "After X days of testing" or any duration-of-use claim
• Any specific throughput or range figure presented as our own measurement
• Any first-hand-use anecdote about a product
• Any review, rating or review count we did not receive from a real source
• Any price we have not verified, or any invented product identifier

The phrasings we use instead: "independent testers at [named source] measured…", "the manufacturer rates this at…", and "we have not tested this ourselves; here is what the published data shows".

How we choose which products to include

Every product on this site is in a single registry, and pages may only pick from it. That is a structural safeguard rather than a preference: it makes it impossible for a page to name a product that has not been through the verification above, and impossible to invent a product identifier to fill a list.

We include a product if it is currently sold, genuinely suits at least one situation we write about, and we can say something specific about its trade-offs. We leave products out when the honest answer is that nothing in that category fits the page — a bare list is better than a tenuous recommendation.

When we are wrong

We will correct it and say that we did. Specifications change, products are discontinued, providers change their rental fees and approved-equipment lists, and standards move. If you find something out of date or simply incorrect, tell us — the correction process is on our editorial policy page.