You have probably sat through this slide. "Conversion rate up 20%." It looks good. Then you open the underlying data and find the rate moved from 0.5% to 0.6%. The 20% is arithmetically correct. It is also nowhere near enough information to decide anything.
Calculating a growth rate is one division. The hard part is never the calculation: it is reading the resulting number and reporting it honestly. Here are five failures that keep recurring, each with real numbers.
First: five names, one calculation
Growth rate, percentage change, rate of change, MoM, YoY, QoQ, WoW. They get used interchangeably in meetings, and the formula behind all of them is identical.
Growth (%) = (Current − Previous) ÷ Previous × 100
MoM (month over month), YoY (year over year), QoQ (quarter over quarter) and WoW (week over week) differ only in what you put in as the previous value. They are names for a comparison period, not separate formulas, which is why one growth rate calculator covers all of them.
If MAU went from 1,250,000 to 1,437,500: (1,437,500 − 1,250,000) ÷ 1,250,000 = +15%, an absolute gain of 187,500 users.
Trap 1: A percentage with no absolute number attached
That "conversion up 20%" slide is this trap. When the denominator is small, percentages get large for free. If one conversion out of 200 visitors becomes two, that is +100% growth: and if it drops back to one next week, that is −50%. You would be reporting noise as performance.
Always pair the percentage with the absolute change. Written as "+15% (+187,500 users)", the reader can judge the scale for themselves. When the sample is in the dozens, showing the raw counts alone is the more honest choice.
Trap 2: Assuming percentages are symmetric
A value drops from 100 to 80. That is −20%. What increase brings it back to 100?
Not +20%, but +25%. Adding 20% to 80 gets you 96, not 100, because the base changed underneath you.
This one costs real money when setting targets. "We lost 15% last quarter, so let's recover 15% this quarter" does not return to the starting point: the recovery actually needs 17.6%. Recovery targets have to be worked backwards from the original value, not from the size of the decline.
Trap 3: Mixing percent (%) with percentage points (pp)
Conversion moved from 2% to 3%. Both of these are valid descriptions:
- +50%: the relative increase (3 ÷ 2 − 1)
- +1pp: the absolute difference (3 − 2)
Both are true, and they leave completely different impressions. Choosing % when you want the result to look big and pp when you want it to look small is a real and common habit. For rate metrics such as conversion, churn or market share, report the change in percentage points and add the percentage in parentheses if it helps. The calculator will hand you +50%; deciding how to present it is your call, not the tool's.
Trap 4: Watching MoM while ignoring seasonality
December revenue is 40% above November. Is that growth?
For any business with a year-end peak, December beating November is what happens every single year. The question that carries information is how December compares with last December. If that is −5%, the +40% MoM slide is actively hiding the fact that things are getting worse.
Where seasonality exists, YoY is the primary view and MoM is supporting detail. The reverse holds for a young product with no seasonal pattern yet, where MoM and WoW say more about momentum.
Trap 5: Calling two data points a trend
A growth rate sees exactly two moments. Whether the path between them was a steady climb or a spike followed by a collapse is invisible in the number.
To summarize three or more periods as a single figure, use the CAGR calculator instead. With +100% in year one and −50% in year two, averaging the annual rates arithmetically gives +25%, while the value actually ended exactly where it started (CAGR 0%). Growth compounds; it does not add.
When not to use a growth rate at all
- When the previous value is 0: it is undefined. Going from 0 to 100 is "+100", not infinite growth.
- When the previous value is negative: a loss turning into a profit produces a percentage whose sign reads backwards and misleads. Report the absolute change.
- When the sample is tiny: with a denominator in the double digits or lower, the percentage carries almost no statistical meaning.
- When you need a forecast: growth rate summarizes the past. To project where that rate leads, move to the compound growth calculator.
The time-consuming part of growth reporting was never the arithmetic. It is choosing the right comparison period and attaching scale and context to the number. Hand the division to a tool and spend your time on the rest.