How to Convert Units: The Method That Replaces Memorised Factors

1. Why Memorising Conversion Factors Fails

Ask an engineer how many millimetres are in an inch and you will get an instant, correct answer: 25.4. Ask how many square feet are in a square metre and the answers get slower — and wrong more often than you would expect. The number of unit pairs a working engineer meets is quadratic in the number of units. Eighteen categories, a few dozen units each, and you are well past the point where memory is a reliable storage medium.

The failure is structural, not personal. Memory has no error signal. If you misremember 25.4 as 25.0, you produce a number that looks entirely reasonable — a part 1.6 mm undersized is not visually wrong, it is scrap. If you remember "a metre is about 3.28 feet" but then apply 3.28 to a square metre, you get an area 7.5 times too small, and nothing about the arithmetic will tell you so.

There is a better way. It is older than the metric system, it is taught in introductory chemistry as dimensional analysis and in engineering as the factor-label or unity-bracket method, and it has one property that memory lacks: when you do it wrong, the units fail to cancel and you can see it before you ever compute a number.

2. The Only Rule: Multiply by One

Any statement of equivalence can be written as a fraction that equals one. If 1 inch is 2.54 centimetres, then

Multiplying a quantity by one does not change the quantity — it only changes the label. That is the entire method. Write the conversion as a fraction whose numerator and denominator are the same physical quantity expressed in different units, multiply, and cancel.

Take 180 centimetres and ask for feet. The site's own centimetres to feet converter returns 5.90551181 ft, but you can derive it from two anchor definitions — 1 in = 2.54 cm and 1 ft = 12 in:

Read the unit chain left to right. Centimetres appear in the numerator of the first term and the denominator of the first bracket, so they cancel. Inches appear in the numerator of the first bracket and the denominator of the second, so they cancel. Feet remain. The chain is the proof: if you had written the first bracket upside down — (2.54 cm / 1 in) — the centimetres would stack up in the numerator and never cancel, and you would see the error immediately without computing anything.

2.1 The units tell you which way up the fraction goes

This is the part people who memorise factors never learn. You do not have to remember whether to multiply or divide. Put the unit you are leaving in the denominator and the unit you are going to in the numerator, and the cancellation does the rest. Ask yourself: "what unit do I want to be left holding?" — then build a fraction that removes everything else.

2.2 You memorise two or three definitions, not two hundred factors

Every metric-to-imperial length conversion on this site descends from a single sentence in the 1959 International Yard and Pound Agreement: 1 yard = 0.9144 metre exactly, which fixes 1 inch = 25.4 mm exactly. From that one definition you can build twelve others on paper:

Three anchor definitions — 2.54 cm, 0.45359237 kg, 3.785411784 L — cover most of what an English-speaking engineer needs. Everything else is a chain of unity fractions. This is why the length guide and the weight and mass guide both open with the definition rather than the table of factors: the definition is what you actually need to carry.

3. Three Worked Examples, Increasingly Awkward

The method scales from a single ratio to compound units and powers without any change in technique. What changes is the number of brackets.

3.1 Single ratio: 150 lb to kilograms

The pound is defined exactly as 0.45359237 kg by the same 1959 agreement, so the conversion is one bracket:

The bracket is a definition, not a measurement. It has no significant-figure limit of its own — the only precision limit is the 150 you started with. This is the case where people ask "how many decimal places should I keep?" and the answer is: as many as your input justifies, and not one more. See pounds to kilograms for the reciprocal direction.

3.2 Compound units: 60 mph to metres per second

Speed is length over time, so two brackets are needed — one for the distance, one for the time. The trap is the second bracket: hours appear in the denominator of "mph", so the time fraction must be written with hours on top to cancel them.

Check the chain: miles cancel (numerator against denominator), hours cancel (numerator against denominator), and what remains is metres over seconds. Had the time bracket been written as (3600 s / 1 h), the hours would have stacked and the answer would have been off by 3600² — a number so large it would have been obvious. That is the point: with factor-label, the large errors become visible before arithmetic, and the small ones become visible in the check step of Section 5. For the reverse direction, see km/h to m/s and mph to km/h.

3.3 Powers: 1 m³ to cubic centimetres

This is where the memorised-factor habit produces its most expensive errors, because the factor is not the one the memory offers.

The bracket is cubed because the unit is cubed. The factor is 10⁶, not 100. The same rule gives 1 m² = 10,000 cm² (factor squared), and it is the reason 1 m² = 10.7639 ft² rather than 3.28084 ft². Whenever the unit carries an exponent, the whole bracket carries it too — which is why the area guide and the volume guide both warn that early rounding error is squared (or cubed) along with the unit.

4. The Five Cases Where Multiplication Is the Wrong Tool

Everything above assumes a conversion is a ratio — that zero in one scale is zero in the other, and that the relationship is a straight proportional line through the origin. For most engineering units that is true, which is exactly why the habit is dangerous: five common families of units break the assumption, and in all five the multiplication still produces a plausible number.

4.1 Scales with a zero offset: °C and °F

The Celsius and Fahrenheit scales do not share a zero point. Water freezes at 0 °C and at 32 °F; it boils at 100 °C and 212 °F. The relationship is affine, not proportional:

Applying a pure factor gives 0 °C → 0 °F, which is wrong by 32 degrees — a comfortable autumn day reported as a hard freeze. The two scales cross at exactly one point, −40°, which is the one temperature you can convert from memory without the offset. Kelvin is different again: K and °C share a step size but differ only by an offset (K = °C + 273.15), so that pair needs addition and no multiplication at all. Use Celsius to Fahrenheit, Fahrenheit to Celsius, or Kelvin to Celsius rather than a factor.

4.2 Logarithmic scales: dBm and dB

The decibel is not a unit of power; it is a unit of ratio, read on a logarithmic scale. dBm anchors that ratio to one milliwatt:

There is no conversion factor between dBm and watts, because the mapping is not linear. 0 dBm = 1 mW, 30 dBm = 1 W, −30 dBm = 1 µW. Adding 3 dB multiplies power by 1.995 (call it 2 in the field); adding 10 dB multiplies it by 10. And because the scale is logarithmic, cascading gains add where powers would multiply: two amplifiers of +20 dB each produce +40 dB, which is 10 W, not 20 W. Any engineer who "multiplies by the dBm factor" has misunderstood what the unit is. See dBm to watts for the exact values, and the electric conversion guide for the two laws underneath.

4.3 Reverse-ordered scales: AWG wire gauge

American Wire Gauge runs the wrong way round: a higher number is a thinner wire. AWG 10 is 2.588 mm across; AWG 40 is 0.0799 mm. The gauge is defined geometrically, as a geometric progression:

Because the definition is exponential, AWG to square millimetres is not a multiplication — it is an exponential curve, and wire area is what actually carries current. A designer who assumes "AWG 12 to AWG 10 is a fixed ratio" and scales it across the range will be badly wrong at the extremes. Use AWG to mm² or mm² to AWG, which apply the definition above.

4.4 Conversions that need a third quantity: kVA and kW

Some conversions cannot be answered from the two units alone, because the answer depends on something you have not been told. Real power in kilowatts relates to apparent power in kilovolt-amperes through the power factor of the load:

A 100 kVA transformer delivers 100 kW to a purely resistive load (PF = 1) and 80 kW to a load with PF = 0.8. The question "how many kW is 1 kVA?" has no single answer — which is exactly why uninterruptible supplies and standby generators carry both figures on the nameplate, and why sizing a generator from the kVA figure alone is a classic commissioning failure. See kVA to kW and the power factor calculator, and the energy and power guide for the rest of the family.

4.5 Same name, different definitions

Here the multiplication is valid but the factor is not unique, because the unit name is shared by quantities that differ substantially. Choosing the wrong definition gives a plausible number and a real error:

The last row bites hardest. A tyre inflated to 32 psig is at 46.7 psia, and any gas-law calculation that uses 32 will be wrong by nearly a third. The number did not lie; the label was abbreviated. This is why the site writes units out in full wherever a variant exists — litres to gallons and gallons to litres both state US gallons explicitly, and the pressure guide treats psia, psig and bar as three distinct starting points rather than one.

5. Catching Your Own Mistake in Ten Seconds

The factor-label method prevents the errors you can see. These three questions catch the ones you cannot, and they take less time than reading the answer a second time.

5.1 Did the units cancel?

Treat the units as algebraic symbols and cancel them across the whole expression. A conversion that is set up correctly ends with exactly one unit — the one you asked for — and no leftovers. A stray unit anywhere in the middle means a bracket is upside down. This check costs nothing because you are looking at the units you already wrote, not at new work.

5.2 Is the order of magnitude right?

Ask whether the answer should be bigger or smaller than the input, and by roughly how much. A kilometre is shorter than a mile, so km-to-miles must come out smaller. A metre converted to feet must come out larger — 1 m is 3.28 ft, not 0.328 ft. This is the check that catches a misplaced decimal point, and a decimal point is the single most common shape of a unit-conversion disaster: see the Isaac Peral submarine case, where a displaced decimal sent a hull design on a scale nobody intended.

5.3 Can you get back?

Convert the result back to the original unit. It should return the number you started with, to within rounding. If 150 lb becomes 68.04 kg, then 68.04 kg must become 150 lb. A round-trip that does not close means one of the two directions used a different definition — and this check is also how you find out that the two conversion factors of a pair are reciprocals, not independently rounded numbers. Multiplying by 0.45359 and then by 2.20462 does not return to the start unless those two values are exact inverses; the site's kg to lbs and lbs to kg converters share one definition so that they do.

5.4 Round last, and only once

Never round an intermediate result. Convert 100 ft to metres (30.48 m), keep the full value, convert again to centimetres (3,048 cm) — the round trip closes. Round 30.48 to 30 first and the round trip returns 98.4 ft, a 1.6% loss created out of nothing but impatience. The same discipline applies to significant figures: the answer to a conversion inherits the precision of the input, not the precision of the definition. If you measured 150 lb, the kilogram answer has three significant figures, whatever the calculator displays.

6. What Happens When Nobody Checks

Unit conversion errors are not usually arithmetic errors. They are check errors — the arithmetic was executed correctly on the wrong quantity, and no one asked Section 5's three questions before the work went out of the door. The case studies on this site are worth reading as a catalogue of skipped checks.

The pattern across all of them is the same: a correct procedure applied without a check, on a quantity whose unit was assumed rather than stated. The main mistakes hub collects the full set, and the cost of retooling essay explains why the opposite error — converting everything, everywhere, all at once — is also expensive.

7. Choosing the Right Converter

You now have the method. Use it to set up the problem and to check the answer — and use a converter for the arithmetic, so that the definition is applied exactly once and at full precision rather than through a rounded constant. Every converter on this site applies the definition, not an approximation.

By category

The worked examples from this page

The cases where multiplication fails

Frequently Asked Questions