How To Calculate Revenue Growth Rate

Revenue growth rate measures how far a company's top line moved between two points in time, expressed as a share of the earlier figure. A business that booked 1,200,000 in one year and 1,500,000 three years later added 300,000, and 300,000 divided by 1,200,000 is 0.25, so revenue grew by 25%. That single division is the whole of the calculation; everything else on this page concerns which question the resulting number actually answers.

The phrase hides two different questions, and they have different answers. The first is how much revenue grew across the entire span, which is 25% here. The second is how fast it grew per period, which works out to 7.7217% once the compounding is spread evenly across the three years. Neither figure paraphrases the other, and quoting one where the other was expected is the most common reason a growth number gets challenged in a meeting.

The divisor is always the starting revenue, and that is a requirement rather than a preference. The same gap of 300,000 read against 1,500,000 gives 20%, which is not the growth that happened but the decline that would be needed to travel back. Multiplying back is the quickest check available: 1,200,000 x 1.25 lands exactly on 1,500,000, while any version built on the wrong base fails to return the endpoint.

Dividing the total by the number of periods is the error this page exists to head off. Twenty-five percent over three years is not 8.3333% a year; the true figure is 7.7217%, and the shortcut overstates it by 0.6116 percentage points. The reason is that each period's growth is earned on a base the previous period has already enlarged, so the parts multiply rather than add. The distortion grows with the size of the total: a 400% rise over three periods annualises to 70.9976%, not to the 133.33% that dividing would suggest.

Writing the change as a multiple keeps the arithmetic honest. Ending revenue of 1,500,000 against a start of 1,200,000 is a multiple of 1.25, and total growth is simply that multiple minus one, giving 0.25. Multiples chain in the way percentages do not: two consecutive spans of 1.25 and 1.2 combine to 1.5 rather than to anything resembling a sum of their parts. This is also the form to reach for when the starting figure is very small.

Nothing in the formula needs adjusting when revenue falls. A business sliding from 1,500,000 to 1,200,000 has a multiple of 0.8, a total growth of -20%, and a compounded rate of -7.1682% per period across three years. The negative sign emerges from the arithmetic rather than from any special case, and it is a legitimate measurement rather than an error state. What does distort a decline is quietly switching the denominator, which makes the fall look smaller than it was.

A percentage says nothing about scale, and this is where growth figures mislead most reliably. Revenue going from 10,000 to 50,000 is 400% growth, which sounds extraordinary and represents 40,000 of new business. The same 40,000 added to a base of 1,200,000 would be 3.33% and would barely register. Report the absolute increase next to the rate whenever the two endpoints are small, or when the audience is meant to judge commercial substance rather than momentum.

Both endpoints have to be the same kind of number for the ratio to mean anything. Net revenue after returns, discounts and allowances is not gross bookings; revenue recognised under one policy is not revenue recognised under another; and a series restated for a currency change or a reclassification is not comparable with the year before it. When the two endpoints were produced by different rules, the growth rate describes a change in reporting rather than a change in trade, and it should be labelled as such.

Count periods as the number of growth steps between the endpoints, not the number of calendar labels touched. Four annual reports, covering 2021 through 2024, contain three steps, so n is 3; three fiscal years contain only two. The choice of window is a judgement in its own right, because a span beginning in an unusually weak year flatters the result and one beginning in an unusually strong year suppresses it. A single rate also says nothing about what happened in between, so pair it with the yearly sequence whenever the shape of the path matters.