New minus old, over old
Take the change from last year to this year, divide it by last year's figure, and express it as a percent. It works for revenue, users, traffic, anything. It compares the same period one year apart, which is why it is trusted over month-to-month: the seasonal swings cancel out.
Reading the number
| YoY growth | What it usually means |
|---|---|
| −15% or worse | Serious decline — losing share or facing disruption |
| −5% to +5% | Flat; treading water |
| +5% to +15% | Healthy, above-inflation growth for a mature business |
| +15% to +50% | Strong growth |
| +50% and up | Rapid — normal for startups, hard to sustain |
What counts as good depends on the field: 10% is solid for an established retailer but disappointing for an early startup, where 30–50% is the target.
The trap: the low base effect
A huge YoY number can be an illusion if last year was unusually bad. Growing off a depressed base looks great but may not mean you have recovered.
- Example. A normal year was $1,000,000. A bad year dropped to $500,000. Recovering to $650,000 shows +30% YoY, yet you are still 35% below where you were two years ago.
- Fix. Look at a two- or three-year trend, not a single YoY figure, when the prior year had a one-off shock.
- To forecast forward. Multiply the current value by (1 + growth ÷ 100). At +25%, next year projects to 1.25× this year — assuming the rate holds, which is never guaranteed.
Common questions
How do I calculate year-over-year growth?
Subtract last year value from this year value, divide by last year value, then multiply by 100. Revenue of $1,250,000 this year versus $1,000,000 last year gives (1,250,000 - 1,000,000) / 1,000,000 x 100 = 25%.
Can YoY growth be negative?
Yes. If this year is lower than last year, the result is negative. $900,000 against $1,000,000 last year is a -10% YoY change, meaning a decline.
Why use year-over-year instead of comparing to last month?
Comparing the same period one year apart cancels out seasonality. December always beats November in retail, so month-to-month looks great every year even with no real growth. YoY strips that out and shows the underlying trend.
