Every spec sheet is two documents at once. The engineering table — the full PDF on the manufacturer’s support site — describes how the device was tested. The marketing grid on the product page selects which of those numbers to show, rounds them upward, and adds qualifiers that quietly move the promise. Reading a spec sheet well means reading the second document while reconstructing the first: knowing which words are hedges, which numbers are not comparable across brands, and which important numbers are missing entirely.
Why the sheet is written by marketing, not engineering
The spec sheet is a competitive document first. Vendors choose the rows that win comparisons against same-priced rivals, and omit the rows that lose. That incentive explains most of what follows: the qualifier words that turn claims into non-claims, the cherry-picked units, the missing rows. Nothing here means the numbers are false — the marketing grid is usually drawn from real measurements — but the selection and framing are load-bearing, and the selection is not neutral.
The practical habit that follows from this: never evaluate a spec sheet in isolation. Read it next to a rival’s sheet at the same price, and next to the manufacturer’s own detailed specifications document, because the deltas between the marketing grid and the full document are where the concessions hide.
“Up to,” “typical,” and other hedged numbers
Qualifiers do the heavy lifting in modern spec sheets, and they follow patterns worth memorizing.
“Up to” marks a ceiling. Battery life claims are the classic case: Apple’s own MacBook Pro specifications list up to 24 hours of video streaming but up to 16 hours of wireless web on the same machine — a real 8-hour spread produced by nothing more demanding than web browsing. Any “up to” figure should be mentally relabeled as the best case you will rarely see.
“Typical” versus “minimum” signals that variance exists and the vendor chose the flattering end. Display brightness is the sharpest example, because vendors list two different numbers without always flagging which one you are getting. Apple’s MacBook Pro spec page states 1,000 nits sustained full-screen brightness and 1,600 nits peak — but the peak figure applies only to HDR content, small bright areas for a short time. Sustained full-screen brightness is the number that predicts reading a screen outdoors, and it can be half the headline. NotebookCheck’s reviews consistently measure lower real-world brightness than vendor claims, which is the pattern to assume until a measurement outlet says otherwise.
“Peak” versus “sustained” also governs charging watts and processor clocks: a phone advertising 120-watt charging may hold that rate only briefly before thermal management cuts it, and a laptop’s maximum clock speed describes a burst, not a workload.
Numbers that sound precise but are not comparable
Several spec rows look objective and are anything but, because the underlying definitions differ or the conditions are vendor-chosen.
IP ratings are the most misunderstood row in consumer electronics. The two-digit code comes from IEC 60529 — the first digit grades dust protection from 0 to 6, the second grades water protection from 0 to 9. IP67 is fixed by the standard: 1 meter of fresh water for 30 minutes. IP68 is not: the standard leaves depth and duration to the manufacturer, provided they exceed IP67, so one brand’s IP68 might mean 1.5 meters and another’s 6. The IEC itself notes that the rating describes resistance at the time of testing, does not cover hot, salty or soapy water, and says nothing about how long the seals last. A liquid-damaged phone is also usually a warranty exclusion regardless of the rating.
Battery capacity in milliamp-hours (mAh) is only meaningful within one voltage. Across devices with different cell voltages and chemistries, watt-hours (Wh) is the comparable unit — which is why laptop batteries are specified in Wh and phone batteries in mAh, and why a 5,000 mAh phone battery cannot be compared with a 5,000 mAh tablet battery at face value.
Charging watts describe delivery rate into the battery, not the whole picture: the charging curve tapers as the battery fills, so “50% in 20 minutes” is a more honest claim than a peak watt figure. Mouse DPI, similarly, is a sensor ceiling, not a sensitivity recommendation — every gaming mouse now exceeds the DPI any game uses, making the row useless for comparison.
What is omitted from the sheet
The absent rows are more predictive than the present ones. Software support commitment is the largest omission: Apple publishes no formal update schedule for iPhones, and most Android vendors historically omitted the number entirely — which is precisely why the recent shift matters, with Google committing to seven years of OS and security updates for Pixel 8 and later and Samsung matching seven years from the Galaxy S24 onward. A phone’s useful life is set by that number more than by any camera or chip spec.
Also routinely missing: sustained performance under load (thermal throttling behavior), repairability and parts pricing, real-world battery measurements under normal use, and the distinction between included accessories and sold-separately ones. The omission pattern is consistent across vendors because the pattern of embarrassment is consistent — and the fix is to look up each missing row before purchase rather than after. The warranty terms are a related omission: a device’s protection is defined as much by the exclusion list as by the spec rows, a subject covered in whether extended warranties are worth it.
The full PDF versus the marketing grid
Most manufacturers publish the detailed specifications document on a support subdomain rather than the product page, and the difference between the two is diagnostic. Apple’s support spec pages, for example, carry the qualifiers the product page drops: “typical,” “HDR content only,” “varies by configuration.” USB-IF naming is another case where the full document wins — the standards body’s official names (USB 3.2 Gen 1 at 5 Gbps, Gen 2 at 10 Gbps, Gen 2x2 at 20 Gbps) exist precisely because vendor shorthand like “SuperSpeed Plus” obscures the speed being bought.
The five-minute verification routine: find the full spec PDF from the manufacturer, not a retailer summary; read the footnotes, which contain most of the qualifiers; check one independent measurement outlet — NotebookCheck for laptop displays and battery life, RTINGS for TVs and headphones, GSMArena for phone battery tests — to calibrate how the vendor’s numbers relate to measured ones; and search the manufacturer’s support pages for the warranty and software-support terms that the grid omitted. If a spec appears only in the marketing grid and cannot be traced to the full document or an independent measurement, treat it as decoration. That habit pairs well with timing: sheet-reading is most valuable when applied to the model generations currently on sale, which is the subject of when to buy electronics.
What to actually do with the numbers
Convert the sheet into three decisions. First, sustained values only: sustained brightness, sustained clocks, web-browsing battery life — the numbers that describe ordinary use. Second, comparability checks: convert mAh to Wh when cross-shopping, ask which conditions define any IP68 claim, and ignore DPI entirely. Third, fill in the omissions: support years, battery cycle rating (500 cycles to 80% for iPhone 14 and earlier, 1,000 for iPhone 15 and later — the number that tells you when the battery, the first wearing component, needs service), and the warranty exclusion list. A spec sheet read this way predicts experience; one read as printed predicts the marketing department’s quarter.