Exponential growth is what happens when a quantity changes by a constant percentage of its current size in every period rather than by a constant amount. That single property produces the curve that fools intuition more reliably than any other in applied mathematics: nearly flat for a long stretch, then vertical, with no change in the underlying rate at any point.
Arb Digital built this page to work in all four directions rather than only forwards. Most calculators project a final amount from a rate and a time. This one also solves for the rate given two amounts and a time, for the time given two amounts and a rate, and for the starting amount given everything else — which is usually what you actually have when you are looking at real data rather than setting up a textbook problem.
What This Exponential Growth Calculator Does
It models the relationship A = P(1 + r)ᵗ in the discrete case and A = Peʳᵗ in the continuous case, and rearranges it to solve for whichever of the four quantities you select. It reports the total change, the overall multiplier, the doubling time — or half-life when the rate is negative — and the equivalent continuous rate, so you can move between the two models without redoing the algebra.
The tool is deliberately unit-free. It covers population, bacterial cultures, radioactive decay, audience growth, depreciation and anything else that changes proportionally. Our compound interest calculator is the right page when the subject is specifically money, because it handles compounding frequency, regular contributions and currency formatting that a general growth model has no business assuming.
How to Use It
- Choose what to solve for. Final amount, growth rate, time or initial amount. The selected field becomes the output and is ignored as an input.
- Pick the model. Discrete steps by period, or continuous change at every instant. The two differ by more than most people expect at high rates.
- Fill in the three known values. Keep the rate and the time in the same unit — a monthly rate needs a count of months.
- Enter a negative rate for decay. Everything works identically in reverse, and the doubling time becomes a half-life.
- Read the multiplier and doubling time. They are usually more intuitive than the raw final figure, and they make two scenarios directly comparable.
The Formula and How It's Calculated
The discrete model is A = P(1 + r)ᵗ. Rearranged, the rate is r = (A/P)^(1/t) − 1 and the time is t = ln(A/P) ÷ ln(1 + r). The continuous model replaces the base with e: A = Peʳᵗ, giving r = ln(A/P) ÷ t and t = ln(A/P) ÷ r. Doubling time is the time at which A/P equals 2, which is ln(2) ÷ ln(1 + r) discretely and ln(2) ÷ r continuously.
Work through the default. Starting at 1,000 and growing 7% per period for 10 periods gives 1,000 × 1.07¹⁰ = 1,967.15, so the quantity has not quite doubled. The overall multiplier is 1.9672 and the total change is +967.15. Doubling time is ln(2) ÷ ln(1.07) = 0.6931 ÷ 0.06766 = 10.24 periods, which is why ten periods falls just short. The equivalent continuous rate is ln(1.07) = 6.766%, slightly below the discrete 7% because continuous compounding does more work with a smaller stated rate. Wolfram MathWorld's exponential growth entry derives the same relationships from the underlying differential equation.
Why the Curve Looks Flat Right Up Until It Doesn't
At a constant 7% per period, a quantity gains 70 units in the first period and 129 in the tenth. In the fiftieth it gains over 1,900 — more in one period than it started with in total. The percentage never changed. What changed is the base it applies to, and that is the whole of the effect.
The practical consequence is that early exponential growth is almost impossible to distinguish from linear growth by eye, and any judgement about whether something is "taking off" made from the first few points is unreliable. The reliable test is not the shape of the line but the ratio between successive values: if each period is a roughly constant multiple of the last, the process is exponential regardless of how flat it currently looks. Plotting on a logarithmic scale turns exponential growth into a straight line, which is why epidemiologists and analysts default to log axes when the question is whether a rate is changing.
Doubling Time, Half-Life and the Rule of 70
Doubling time is often the most useful summary of an exponential process because it is independent of the starting size. A 7% rate doubles a quantity every 10.24 periods whether it starts at ten or ten million.
The Rule of 70 gives it almost instantly: divide 70 by the percentage rate. At 7% that is exactly 10, against a true 10.24. At 2% the rule says 35 and the truth is 35.0. At 20% the rule says 3.5 and the truth is 3.8, so the approximation degrades as rates climb, but it is dependable through the range most real processes occupy. The same arithmetic runs backwards for decay: a −5% rate has a half-life of ln(0.5) ÷ ln(0.95) = 13.5 periods. Radioactive decay is the purest example, since the half-life of an isotope is a fixed physical constant — the US Nuclear Regulatory Commission's definition of half-life and the EPA's page on radioactive decay describe the process this model was originally built for. For isotope-specific work, our half-life calculator handles the decay case directly.
Discrete Versus Continuous: When the Difference Matters
The two models answer slightly different questions. Discrete growth assumes the quantity jumps at the end of each period. Continuous growth assumes it changes at every instant, with each instant's growth immediately contributing to the next. At a stated 7% the gap over one period is small — a multiplier of 1.07 against 1.0725 — but it widens sharply as the rate rises. At a stated 100%, discrete doubling gives 2× while continuous gives e, about 2.718.
Choose by the process, not by preference. A population with an annual breeding season is genuinely discrete. A bacterial culture dividing continuously, or a chemical reaction, is genuinely continuous. Where the model matters most is when converting a quoted rate between the two conventions: the equivalent continuous rate reported in the grid is ln(1 + r), and using a discrete rate in a continuous formula overstates growth every time. Our logarithm calculator handles that conversion on its own if you need it separately.
Fitting a Rate to Real Data, and Its Limits
Solving for the rate from two observed points is the most common real use of this page, and it comes with a caveat worth stating plainly. Two points define exactly one exponential curve, so the fit is always perfect and tells you nothing about whether the process is actually exponential. A quantity that went from 1,000 to 1,967 over ten periods yields 7% whether it grew smoothly, sat flat for nine periods and jumped in the tenth, or oscillated wildly around the path.
The honest procedure is to fit the rate across several sub-intervals and see whether the answers agree. If early periods give 12% and later ones give 3%, the process is decelerating and a single exponential rate will overproject badly. That deceleration is the norm rather than the exception: no real quantity grows exponentially forever, because every real system eventually hits a constraint — food, market size, physical space — and bends into a logistic S-curve instead. If you have more than two points and want the trend line properly estimated, use our linear regression calculator on the logarithms of the values, which is exactly how exponential fits are done in practice.
Reading Percentage Change Correctly
Two errors show up constantly. The first is averaging growth rates arithmetically. A quantity that grows 50% then falls 50% is not back where it started — it is at 75% of the original, because the second percentage applies to a larger base. The correct average of successive growth rates is the geometric mean, which our geometric mean calculator computes directly.
The second is confusing a percentage change with a percentage-point change, especially with negative rates. A decay of 5% per period never reaches zero — after twenty periods a value of 1,000 is still at 358, and after a hundred periods it is still at 5.92. Exponential decay approaches zero asymptotically without ever arriving, which is why depreciation schedules that must reach zero use straight-line or declining-balance rules with a floor rather than a pure exponential. A model that never terminates cannot describe an asset that is eventually written off.
Arb Digital models traffic, audience and pipeline growth against actual data rather than a single optimistic rate carried forward.
Browse All Free Tools Talk To Our TeamCommon Mistakes to Avoid
- Mixing rate and time units — a monthly rate with a count of years overstates growth by a factor of twelve in the exponent.
- Fitting a rate from two points and projecting it far forward — two points always fit perfectly and reveal nothing about whether the growth is genuinely exponential.
- Averaging growth rates arithmetically — successive percentage changes compound, so the geometric mean is the correct average.
- Using a discrete rate in a continuous formula — convert with ln(1 + r) first, or the projection runs high.
- Extrapolating exponential growth indefinitely — real systems hit constraints and bend into an S-curve, usually well before the projection says they should.
Related Free Tools From Arb Digital
Handle money specifically with the compound interest calculator, find how long a balance takes to double with the money doubling calculator, average successive rates with the geometric mean calculator, work with the underlying logs using the logarithm calculator, or model a constant-ratio series term by term with the geometric sequence calculator. The free online tools hub lists every maths calculator we publish.
Frequently Asked Questions
It is growth where a quantity increases by a constant percentage of its current value each period, so the absolute increase gets larger every period even though the rate never changes.
Divide the natural logarithm of two by the natural logarithm of one plus the rate. For a quick estimate, divide 70 by the percentage rate, which is accurate to within a few percent for typical rates.
Discrete growth applies the rate once per period in a single step. Continuous growth applies it at every instant, so it produces slightly more over the same period at the same stated rate.
Enter a negative rate. The same formula applies, the multiplier falls below one, and the doubling time is replaced by a half-life.
Yes. Choose to solve for the rate, enter the initial and final amounts and the number of periods, and the calculator returns the constant rate that connects them.
No. Each period removes a fixed proportion of what remains, so the quantity approaches zero without ever arriving at it.
Usually because a short-run rate was extrapolated over a long horizon. Real processes decelerate as they meet constraints, and an exponential model assumes no constraint exists.
This page explains a mathematical model for educational purposes only. It is not financial, investment, medical, or safety advice, and projections should never be used as the sole basis for a decision.
