Product Specifications Explained: Which Numbers Actually Matter

A specification table looks like evidence. It has units, decimal places and a grid, and it invites you to assume that the product with the bigger numbers is the better product. Occasionally that is true. More often the table is a mixture of measurements taken under conditions you cannot see, figures that describe consumption rather than performance, and numbers that mean nothing at all outside a marketing meeting.

This is the glossary the rest of our category guides refer back to. It explains what the common numbers describe, where each one breaks down, and how to compare two spec sheets without being led by whichever manufacturer rounded up more enthusiastically.

Three kinds of number on a spec sheet

Physical facts. Weight, dimensions, capacity in litres, cable length, material composition. Hard to fake, easy to check against a tape measure, consistently the most useful numbers on the page — and consistently the ones buyers skip.

Measured performance. Energy consumption, noise in decibels, runtime, brightness, suction. Real measurements, but each depends on a test procedure. Where the procedure is set by regulation the figures are comparable; where the manufacturer chose it, they are comparable only with themselves.

Marketing constructs. Peak power, "up to" durations, invented indices and star systems with no defined scale. These exist to be larger than a competitor's. The clue is the absence of a stated test, a stated unit, or both.

Sorting a table into those three groups takes a minute and does most of the work. After that, as in how to judge product quality online, the question is whether the construction supports what the numbers promise.

Power: watts, "peak" figures, input vs output

A watt is a unit of power, and on most appliances it describes what goes in, not what comes out. A 2,000 W vacuum cleaner draws two kilowatts from the socket; how much of that becomes airflow depends on the motor, the seals and the airway design. Two machines with identical wattage can differ enormously in useful output, which is why efficiency rules across several categories have pushed input power down without reducing performance.

Where output is quoted, check whether it is continuous or peak. Continuous figures describe what a device sustains; peak figures describe an instantaneous maximum under ideal conditions. In audio, "peak music power output" and similar constructs can be several times the continuous rating and tell you very little — the distinctions that matter are set out in headphone specifications explained.

Rules of thumb: compare watts only within one product type; prefer a stated continuous output to any peak; and where a category has efficiency labelling, treat the label class as the more meaningful number.

Capacity: mAh, Wh, litres and usable volume

Milliampere-hours are the most misread figure in consumer electronics, because mAh is only half a measurement. Energy capacity is watt-hours: multiply mAh by the voltage and divide by 1,000. A power bank rated 10,000 mAh at a 3.7 V cell voltage holds about 37 Wh. Delivering that at 5 V leaves about three-quarters of the headline mAh before the converter has done anything, and the conversion itself loses more as heat — so expect roughly 60 to 70 per cent of the printed figure to reach your phone. Comparing two devices by mAh alone is valid only when their voltages match.

The same gap between nominal and usable runs through physical capacity. Fridge listings often quote gross volume, while the label figure is net — the space left once shelves, ducting and the door line are accounted for. Vacuum bin capacity is measured to a fill line you will never reach with fine dust. Kettle capacity is quoted to the maximum mark, above the usable one.

Runtime figures behave the same way, quoted at the lowest power setting or the lowest volume. What that means in practice is covered in battery life claims explained.

Protection: IP ratings and drop ratings

An IP code has two digits. The first describes protection against solid objects on a scale of 0 to 6, ending at dust-tight. The second describes water on a scale of 0 to 9, running from dripping water through spray and jets to immersion. IPX4 means water resistance was tested but dust protection was not: the X is "not rated", not "none".

Two limits matter more than the digits. The tests use clean fresh water at defined pressures and durations, on new samples — not sea water, chlorinated pool water, hot showers or a garden hose. And seals and adhesives age, so a rating describes the product as it left the factory. Manufacturers say as much in their warranty terms, which is usually where you will find liquid damage excluded regardless of the rating. The full breakdown is in IP ratings explained.

Drop ratings work the same way: a stated height, onto a stated surface, for a stated number of impacts, in a laboratory. A claim of testing to a military standard references a test framework, not a warranty against your kitchen floor.

Textiles: thread count, GSM, denier, fill power

Thread count counts threads per square inch of woven fabric. It is a real measurement that is routinely inflated by counting the individual plies within a twisted yarn, so a "1,200 thread count" sheet may be a 300-count fabric counted four times. Above roughly 400 in a standard weave, fibre quality and yarn fineness predict feel and durability far better.

GSM — grams per square metre — is a weight measure, and harder to game than thread count. It describes towels, jersey and heavier cottons usefully: a higher-GSM towel is denser and more absorbent, and also slower to dry.

Denier describes yarn fineness by mass — the grams in 9,000 metres of fibre. Higher denier means a thicker, tougher yarn, which is why it appears on luggage and outerwear rather than bedding.

Fill power measures the loft of down in cubic inches per ounce. Higher fill power gives more warmth for less weight, which is a different claim from more warmth overall — that depends on the total quantity of down, quoted separately in grams. Duvet tog and the rest of the bedding vocabulary are in bedding thread count, GSM and tog explained; what the fibre percentages mean is in fabric composition explained.

Test conditions decide the number

Almost every performance figure is a sentence with the conditions removed. Appliance energy figures come from a standardised programme, generally an eco cycle run at low temperature over several hours — not the quick wash most households actually use, which is why real bills exceed label estimates. Battery figures assume moderate brightness, one task and no background activity. Screen brightness is often quoted as a peak on a small portion of the panel rather than a full-screen sustained figure. Noise ratings are taken at a set distance in an anechoic room.

Two consequences follow. Regulated figures are comparable with each other because everyone runs the same test; unregulated figures are only comparable within a brand. And a specification without its conditions is not a measurement, it is a claim.

Reading the gaps: what an omitted spec implies

Specifications are voluntary unless a regulation compels them, so manufacturers publish what flatters them. The absences are therefore informative. A laptop with no weight listed is heavy. Bedding with no fibre length or origin stated is ordinary cotton. A device with no stated software support period has not committed to one. An appliance with no noise figure in a category where competitors publish theirs is likely to be loud.

Keep a short list of the specification that matters most in each category and check whether it is present at all before you compare anything. Missing is an answer.

Comparing two spec sheets fairly

1. Normalise the units. Convert mAh to watt-hours, gross volume to net, peak to continuous. Half of all specification comparisons fail here.
2. Check the test regime. Regulated against regulated, manufacturer against the same manufacturer. Never across.
3. Pick three deciding numbers. Choose them before you look at the products, based on the job you need done.
4. Score the gaps, not the values. A 5% difference is noise; a 40% difference is a design decision.
5. Verify one physical fact. Measure the space, weigh the equivalent object, check the plug type. This catches more bad purchases than any performance figure.
6. Write down what would change your mind. If nothing on the sheet could, you had already decided and the table is decoration.