The average adult male is 5’9″ tall. But, that doesn’t tell you anything about the average height of NBA players.
The average new car sold in the U.S. can drive around 25 miles on a single gallon of gas. But, that doesn’t tell you anything about the fuel economy of a Toyota Prius.
And, in your business, the average lifetime value (LTV) of a customer may be $500. But, that wouldn’t tell you that the LTV of your best 5% is $2,500 and your worst 5% is actually negative.
Employing averages across large heterogeneous groups can be very misleading, as it can obscure opportunities and risks.
Instead, try the following:
- De-average large populations to focus on more homogeneous one
- Use median, not mean (average) as your summary statistic
- Include the 5th and 95th percentile values to identify outliers