Reviews You Can Trust

“Where do you go for product reviews?”

That is one of the more common questions in my inbox. Thank you to those who have asked. This piece is for you.

In ‘On Independence’, I touched on the reasons why it has become nearly impossible to find trustworthy reviews of consumer products online.

Today, I want to share the sources that I have bookmarked and return to whenever I make a purchase.

It is by no means a complete list. It is just an account of the resources I have personally come to value over the years. My hope is that we might build a more complete list together in The Community.

Before the list, I'll outline the three mechanisms that corrupt almost every review you read online, then the three questions I ask of any reviewer. Together, they'll tell you in about five minutes whether a source is worth your time.

The degradation of trust in online product reviews is caused by three primary mechanisms.

1. Seeding - Get free products in return for reviews.

Brands and marketing agencies send product to reviewers on a schedule, and the units keep coming as long as the coverage stays positive. ToolGuyd exposed the terms of this sort of arrangement:

"We require you to post on social media before the 30th," "one pre-determined item from each brand every other month," and "In return...you keep the product."

Very little is disclosed. Often, all we see is a “Paid Partnership” tag, with no mention that product was free, or that there was a deadline, or that future freebies depend on the content of the review.

The FTC's own guidance counts a free product as a material connection and says a platform's built-in tag is "not necessarily" adequate disclosure.

But it’s not surprising that with this level of vagueness in the rules, many creators choose the minimum-viable disclosure.

And what happens when brands don’t get the ratings they want from influencers?

In December 2020 NVIDIA told Hardware Unboxed it would stop sending review cards because its reviews "continued to focus singularly on rasterization performance," and would revisit the decision "should your editorial direction change." The email was published, the backlash was immediate, and NVIDIA reversed within two days.

Motorcycle review channel FortNine's Ryan Kluftinger described his experience: "we get a press bike, we make a video, and they don't like some negative things we've said, so they never give us anything else."

2. Sponsorship - Get paid in return for videos, posts or sales.

Furniture maker Anne Briggs explains the lexicon well in an interview on Christopher Schwarz’s substack:

"Ambassador" means free product, "affiliate" means paid per sale, and "sponsorship" means paid for the affiliation.

Agencies offer "a three- to six-video deal," and the contracts "typically require you to give the brand way too much control over what you say about their products."

Here, the sponsor relationship is at least disclosed in the ad read. However, what isn’t disclosed is that the same brand’s products get reviewed elsewhere on the channel or blog, or that the discount code in the description pays the reviewer.

The FTC has settled with companies over exactly this. Warner Bros. paid influencers "from hundreds to tens of thousands of dollars" to cover a video game, "required the influencers to promote the game in a positive way and not to disclose any bugs or glitches they found," and had them put the disclosure where it was "visible only if consumers clicked on the 'Show More' button." The videos were viewed 5.5 million times.

Global influencer spend is now about $33 billion a year.

3. The costume of testing - Adopt the language of objectivity, without the methodology.

When you search “best air purifier”, Google’s algorithm decides which page you see first. Ranking really matters, because the first results get all the clicks. Today, nearly every page linked to a query carries shopping links that pay the publisher a commission when you buy.

So a publisher’s ranking is its revenue. A page that sits at the top of the “best air purifier” query is worth a lot of money every month.

In 2022 Google announced it would move pages up if they showed evidence of actual testing. Its guidance still asks for "quantitative measurements," and "the number of products that were tested, what the test results were, and how the tests were conducted."

Testing is expensive. Claiming you test rigorously is free.

Real Simple's best air purifiers page has a section called "Our Testing Process." It describes unboxing, app setup, filter changes and appearance. On performance, it says this:

"Finally, we looked at the performance, considering any substantial difference we felt in the overall air quality."

No test instrument is named. No robust methodology. No measurement appears anywhere on the page. The only numbers they provide are the manufacturer’s own specs.

Its sister title Better Homes & Gardens says "we tested 144 options in our lab and at home" with "an air quality monitor that measures PM2.5s and TVOCs" and "a decibel meter." Some numbers do appear: a noise range on each product, and on a few, a line like "decreased PM2.5 numbers by 90% and TVOCs by 80% when tested in an enclosed environment."

None of it is checkable by the reader. There is no standardized methodology, no data table, no charts, and no way to tell which figures came from their meter and which came from the manufacturer. There is no consistency in how they report numbers at all.

Both titles belong to Dotdash Meredith, which reported $1.1 billion in digital revenue last year, with affiliate commerce up 21 percent.

HouseFresh is an independent outfit that actually tests air purifiers, publishes their methodologies and provides the raw test data. They documented in 2024 that BH&G's testing claim first appeared in the twenty days before Google's update took effect.

To see the difference, let’s look at the same product reviewed by both outlets. BH&G's page names the PuroAir 240 "Best for Large Rooms". HouseFresh bought the same unit and gave it one star.

Do yourself a favor and actually look at both reviews, then judge them for yourself.

BH&G is just one offender. Every industry has dozens of publishers playing this game.

A real test produces numbers, and numbers produce losers. A costume lab produces ranked lists of only winners in spurious categories with buy buttons on every row.

Sadly, the pantomime wins. In a study of 186,000 Google searches where a content site ranked first, 86 percent of those first places belonged to a large media brand. HouseFresh's own traffic from Google fell from 4,000 visitors a day to 200.

There are three simple checks that make identifying trustworthy sources a whole lot easier.

1. Who bought the unit? Retail purchase is the single strongest signal of honesty there is. A manufacturer that supplies units can provide hand-picked ones. It can also withhold the next one if a review goes badly.

Consumer Reports states their standard plainly: “CR buys all the products and services it rates at retail, just as consumers do”.

Solereview literally posts an image of each receipt for every running shoe it reviews.

That is trustworthy.

2. Who pays for the verdict? Entirely reader funded is best, but that is a tough standard to uphold.

Consumer Reports spends about $30 million a year of its membership revenue buying and testing products. Almost nobody else can achieve that kind of scale. Readers rarely subscribe based on a review they saw. They search for one, read it once, and leave. So most reviewers get paid one of two ways: a commission when you buy, or money from the brand.

Those are wildly different incentives. A commission pays the same whichever brand a buyer picks. Brand money pays for a specific verdict. That’s why affiliate revenue can be ok, assuming the reviewer is objective and discloses the relationship. Sponsorships are never ok, because they necessarily undermine editorial independence.

3. Show me a negative verdict. A real test produces losers. If a tester’s results page only names the 10 best in various spurious categories, that is not a good sign. Thankfully, it is an easy-to-spot format.

So look for the loser. Does the reviewer tell you what not to buy? Vacuum Wars keeps a list of robot vacuums to never buy.

Independence is expensive. RTINGS spent $714,000 buying 618 products last year. I already mentioned Consumer Reports’ $30 million a year commitment. Torque Test Channel is a one-man operation. It sells a $15 rankings spreadsheet that has moved 13,800 copies. That spreadsheet is easily the best money I’ve spent on tools in years. Providing true value like that is how publishers can reject brand money and become incorruptible.

That is no small feat. Every source below has found a way to get paid that doesn’t depend on the brand liking the verdict.

The sources below come in two kinds. Both are useful for different reasons.

Testers aim to provide an objective number. They conduct the same test, across every competing product, side by side, with a method anyone could repeat. This enables them to produce objective rankings of the products they are testing, in the tasks they were designed to do.

Practitioners provide something more: judgment. They did the work for a living. A cobbler has seen a thousand soles come unglued and knows which construction fails and why. A master technician has had the same part in his hands hundreds of times and knows which model years to avoid. That knowledge is demonstrated in how they handle the object and in what they notice that you wouldn't. When you watch them work, you borrow a competence you never had time to build.