CAGR Formula: How to Calculate Compound Annual Growth Rate

Compound annual growth rate (CAGR) compresses a multi-year growth story into one comparable annual rate. The CAGR formula is (Ending Value ÷ Beginning Value)^(1/n) − 1, and it's the standard way finance teams measure revenue growth, investment returns, and market expansion - while getting routinely confused with the simple average of yearly growth rates, a mix-up that can overstate the real trend by 20+ percentage points. This guide covers the formula, a full worked SaaS revenue example, three ways to calculate CAGR in Excel, and exactly where it breaks down.

CAGR compresses an entire multi-year growth path into one number you can compare or extrapolate - but it discards everything about the path itself.

What CAGR Measures (and What It Doesn't)

CAGR answers one specific question: if a metric grew smoothly, at a constant rate, every single year between two points in time, what would that rate have to be? It is not a description of what actually happened in any individual year - real growth is lumpy, seasonal, and often volatile. CAGR is a modelling convenience: a single number that lets you compare two companies, two investments, or two time periods on equal footing, and a single number you can extrapolate forward in a forecast.

That convenience is also its biggest risk. Because CAGR smooths away the actual year-by-year path, two businesses with wildly different risk profiles - one that grew steadily and one that swung between a blowout year and a near-death year - can post the identical CAGR. The number is only ever a starting point for analysis, never the whole story.

The CAGR Formula

CAGR = (Ending Value / Beginning Value) ^ (1 / Number of Years) − 1

Three inputs, all of which must be defined precisely:

• Beginning Value: The metric's value at the start of the period (Year 0). For revenue or ARR, this is typically the first full period you're measuring from.
• Ending Value: The metric's value at the end of the period (Year n).
• Number of Years (n): The number of full compounding periods between beginning and ending value - not the number of data points. Four years of annual data (Year 0 through Year 4) is n = 4, not n = 5.

The formula is a geometric mean in disguise: it finds the constant per-period growth factor that, compounded n times, turns the beginning value into the ending value.

Worked Example: CAGR for SaaS ARR Growth

Assume a SaaS company's Annual Recurring Revenue (ARR) grew as follows over three years:

Notice the path: a breakout year, a down year, then a strong rebound. Applying the CAGR formula to the beginning and ending values only:

CAGR = ($8.0M / $2.0M) ^ (1/3) − 1
CAGR = (4.0) ^ (0.3333) − 1
CAGR = 1.5874 − 1 = 0.5874 → 58.7%

So this company's ARR compounded at an effective 58.7% CAGR over the three-year window - a single number that describes the net effect of a triple, a pullback, and a rebound, with no visibility into any of those individual swings.

Calculating CAGR in Excel (Three Ways)

Using the ARR table above, with Year 0 ARR in cell B2 and Year 3 ARR in cell B5:

1. The direct formula

= (B5/B2)^(1/3) - 1

2. Repurposing the RATE function

Excel's RATE function is built for loan and annuity math, but with no periodic payment it solves the same equation as CAGR. Enter the beginning value as a negative (an outflow) and the ending value as the future value:

= RATE(3, 0, -B2, B5)

Both return 58.7%.

3. Cross-checking with GEOMEAN

CAGR is mathematically a geometric mean of the year-over-year growth factors (not the growth percentages - the factors, i.e. End/Begin for each year). With growth factors in D3:D5 (3.00, 0.8333, 1.60):

= GEOMEAN(D3:D5) - 1

This also returns 58.7%, confirming the direct formula. Use this as a sanity check any time you're computing CAGR across a long or irregular series - it's easy to mis-count the number of periods, and GEOMEAN forces you to lay out every year explicitly.

CAGR vs. Average (Arithmetic) Growth Rate

This is the single most common CAGR mistake in financial models: averaging the yearly growth percentages instead of compounding them. Using the same ARR table, the arithmetic average of the three YoY growth rates is:

Average Growth Rate = (200.0% + (−16.7%) + 60.0%) / 3 = 243.3% / 3 = 81.1%

Compare that to the actual CAGR of 58.7%. The arithmetic average overstates the true compounded growth rate by 22.4 percentage points - a huge gap, driven entirely by the volatility in the middle year.

The reason is structural, not a rounding artifact: percentage gains and losses are not symmetric under compounding. A 50% loss requires a 100% gain just to get back to even, but a simple average of −50% and +100% is +25%, which implies growth that never happened. The geometric mean (CAGR) correctly accounts for this; the arithmetic mean does not. This holds for any series with variation - the arithmetic average is always greater than or equal to the CAGR, and the gap widens as the underlying growth gets more volatile. For a perfectly steady growth path (the same rate every year), the two converge to the same number.

Using CAGR to Project Future Values

Once you have a CAGR, you can run the formula forward to project a future value:

Future Value = Beginning Value × (1 + CAGR) ^ n

Continuing the example, projecting ARR two more years forward (to 2027) at the true 58.7% CAGR:

2027 ARR = $8.0M × (1.5874)^2 = $8.0M × 2.5198 = $20.16M

Now compare what happens if you'd mistakenly used the arithmetic average of 81.1% instead:

2027 ARR (wrong) = $8.0M × (1.811)^2 = $8.0M × 3.2798 = $26.24M

That's a $6.1M overprojection - roughly 30% too high - flowing straight from picking the wrong growth-rate concept. In a real model, that error cascades into headcount plans, burn-rate assumptions, and runway calculations built on top of the ARR forecast. This is exactly the kind of assumption that belongs on a dedicated drivers tab - see our startup financial model guide for how to structure ARR, headcount, and burn assumptions so a change like this is a one-cell fix, not a rebuild. You can see CAGR-driven ARR projections in a full working model in our SaaS MRR/ARR forecast template.

Where CAGR Breaks Down

CAGR is useful precisely because it's simple, and that simplicity is also where it fails:

1. It hides volatility and path risk. Two companies can post the same CAGR while one grew steadily and the other lurched between a blowout year and a near-collapse. CAGR alone can't distinguish them - always look at the year-by-year path, not just the headline rate.
2. It breaks with negative or sign-flipping values. A metric like EBITDA moving from −$1.0M to +$3.0M has no meaningful CAGR - the formula requires a positive beginning value, and a negative base produces a result that isn't economically interpretable. Describe swings like this in dollar terms instead.
3. It's sensitive to the start and end dates you pick. Choosing a start point right after a downturn (an artificially low base) or an end point at a cyclical peak will flatter the CAGR. Always state the exact window and check whether either endpoint is unusually high or low relative to the trend.
4. It assumes a single lump-sum change, not a series of cash flows. CAGR only needs a beginning value, an ending value, and a time period. It has no way to account for money moving in or out mid-period - additional investment, dividends, or distributions. For that, you need IRR, which weights each cash flow by its timing and size.

Common Mistakes

1. Averaging yearly growth rates instead of compounding them. As shown above, this consistently overstates the true growth rate - sometimes by 20+ percentage points on a volatile series.
2. Treating CAGR as a per-year forecast. CAGR is a smoothed abstraction. A company with a 58.7% CAGR did not actually grow 58.7% in each of the three years - it grew 200%, then −16.7%, then 60%.
3. Applying CAGR across a sign change. Running the formula on a metric that goes from negative to positive (or vice versa) produces a meaningless or undefined result.
4. Cherry-picking the window. Measuring CAGR from a trough or to a peak, whether deliberately or by accident, materially changes the answer without changing the underlying business.
5. Confusing CAGR with IRR. CAGR ignores interim cash flows entirely. Using it to describe a private equity or real estate investment with capital calls and distributions along the way will misstate the actual return - use IRR or XIRR instead.
6. Extrapolating a historical CAGR indefinitely. A 58.7% CAGR driven by a small revenue base rarely persists once a company scales - sanity-check any multi-year CAGR-based projection against total addressable market size and comparable companies' actual growth curves at similar scale.