A spec sheet looks like a list of facts, but it mixes three very different things: measured hardware facts, best-case claims tested by the seller, and labels invented by the marketing team. Learning to tell them apart is the fastest way to stop overpaying for numbers that do not change your daily use.
The Review NYU is not affiliated with New York University.
We do not currently earn a commission from any link on this page. If links ever become affiliate links, we will label them clearly.
Short answer. Trust specs that describe fixed hardware and can be checked (storage, screen size, ports, weight, supported codecs, update policy). Treat anything with "up to", "as much as" or a multiplier ("1.5x more", "120x zoom") as a best case measured by the seller under its own conditions. Ignore labels with no defined test ("pro-grade", "military grade", "AI-powered"). Then confirm the numbers you care about with an independent lab that tests every product the same way. Example: in RTINGS' battery test, the Bose QuietComfort Ultra (2nd Gen) lasted 37 hours against Bose's "up to 30 hours" claim, while the 200MP main camera on the Galaxy S26 Ultra scored lower in DXOMARK's tests than the iPhone 18 Pro's 48MP one.
This is an explainer, not a product test. The principles come from how independent test labs separate measured results from manufacturer claims. Every specific example below comes from a maker's own page or a published lab result that we checked on 9 November 2026, and each is linked in the sources. We use phones, headphones and a games console because those are the categories where we have current, like-for-like lab data; the same logic applies to laptops, TVs, vacuums and almost anything else with a spec sheet.
Before you compare two products, sort each line on the spec sheet into one of three groups.
| Type | What it is | Examples | How much to trust it |
|---|---|---|---|
| Fixed facts | Physical or contractual facts that do not depend on how you use the product | Storage, screen size, weight, ports, supported Bluetooth codecs, warranty length, years of software updates | High. Easy to verify and rarely wrong. |
| Best-case claims | Performance measured by the seller, under conditions the seller chose | "Up to 30 hours" battery, "up to 120x zoom", peak brightness, charging "50% in 15 minutes" | Medium. Usually true in the seller's test, often different in yours. Not comparable across brands. |
| Marketing labels | Words or ratios with no published test behind them | "Pro-grade", "best-in-class", "studio-quality", "AI-powered", "dynamic contrast" | Low. Treat as advertising until someone measures it. |
The trap is that all three groups are printed in the same font, often in the same table. A good habit is to mentally add "according to the seller, in its own test" to every line in the second group.
These specs predict everyday experience reasonably well, because they are either fixed or hard to exaggerate.
More megapixels means more pixels, not better photos. Sensor size, lens quality and processing matter more. The clearest 2026 example: the Galaxy S26 Ultra's main camera has a 200MP sensor and the iPhone 18 Pro's has 48MP, yet DXOMARK scored the iPhone's main camera 181 and the Galaxy's 167. Our camera phone comparison covers the details.
"100x" or "120x" zoom figures describe the furthest digital crop the software allows, not how good the picture is. The optical figure (for example 5x) is the one that reflects real lens reach. PetaPixel's review of the Pixel 11 Pro XL says its 120x Pro Zoom uses generative AI to enhance a very low resolution crop, which means detail at that setting is partly invented. Phrases like "optical quality zoom" usually describe a crop from a high resolution sensor: useful, but not the same as a dedicated lens.
Battery claims are real measurements, but in the seller's chosen conditions: a particular brightness, volume, codec and feature set. They are not comparable across brands, because each company tests differently. Independent labs that run every product through the same test are the fix. Sometimes the claim is conservative, sometimes generous; you only find out by checking.
Multipliers are almost always measured against the company's own previous model. Apple, for example, says the AirPods Max 2 offer "up to 1.5x more Active Noise Cancellation than the previous generation". That tells you the new model improved; it tells you nothing about how it compares with Sony or Bose. For that comparison you need a lab: RTINGS rates all three as top of the category for noise isolation.
Screen brightness figures are usually peak values on a small part of the screen for a short time, measured differently by each brand. "Dynamic contrast" ratios on TVs and monitors are often not measured with a standard method at all. Refresh rate (60Hz, 120Hz) is a fixed fact and does matter for smoothness, but only if the content and the chip can keep up.
Words like "4K gaming" describe what output the device supports, not the frame rate you will get in demanding games. Valve's Steam Machine is a useful lesson: Kotaku reported that Valve changed its store page from "4K gaming at 60 FPS" to "4K gaming" on 25 June 2026 after early reviews found many games falling short. Our Steam Machine analysis covers what reviewers measured.
Here is what happens when you put three manufacturer claims next to one independent test that treats every product the same way.
| Headphones | Maker claim | RTINGS measured | Difference |
|---|---|---|---|
| Bose QuietComfort Ultra (2nd Gen) | Up to 30 hours | 37 hours | About 23% longer than claimed |
| Sony WH-1000XM6 | Up to 30 hours | 31.75 hours | About 6% longer than claimed |
| Apple AirPods Max 2 | Up to 20 hours | 22 hours | About 10% longer than claimed |
Two lessons. First, Bose and Sony print the same claim (30 hours), but in a matched test one lasted more than five hours longer. Comparing the two claims would have told you they were equal. Second, all three exceeded their own claims here, which is a good sign, but it is not guaranteed for every product or category. The full comparison is in our AirPods Max 2 vs Sony WH-1000XM6 vs Bose QC Ultra guide.
Fixed facts that do not depend on how you use the product: storage, screen size, weight, ports, supported codecs, warranty length and the number of years of software updates the maker commits to in writing.
It is a best-case result measured by the seller under conditions it chose, such as a set brightness or volume. It is usually true in that test but often different in daily use, and it cannot be compared fairly with another brand's "up to" figure.
No. Sensor size, lenses and processing matter more. In DXOMARK's tests, the iPhone 18 Pro's 48MP main camera scored 181 while the Galaxy S26 Ultra's 200MP main camera scored 167.
The headline figure is a digital crop, not optical reach. The optical figure, such as 5x, reflects the lens. PetaPixel reports that the Pixel 11 Pro XL's 120x setting uses generative AI to enhance a very low resolution crop.
Use an independent lab that runs every product through the same test. For example, Bose and Sony both claim up to 30 hours for their flagship headphones, but RTINGS measured 37 hours for the Bose QC Ultra (2nd Gen) and 31.75 hours for the Sony WH-1000XM6.
Only that the product improved on the company's own older model. It says nothing about how it compares with competitors; you need an independent test that includes both products for that.