% More/Less Likely than Benchmark

Definition

The % more/less likely than benchmark metric is a relative difference calculation.

Relative difference is a way of measuring the percentage increase or decrease from one value to another.

Calculation

The basic formula for relative difference is:

(final value - initial value) / initial value x 100%

Where:

final value = the values for the comparison segments

initial value = the values of the benchmark segment.

Use Cases for this Calculation in Decile

In Decile, we use this metric to understand the difference between segments relative to a benchmark. To help us answer questions such as:

• How much more likely are my male customers to purchase into a specific product category than my female customers?
• During what period am I more likely to acquire younger customers?
• Do high-value customers convert from channels that are different from my overall customer base?

Example

In the example below, we answer the first question above. Are there specific categories men are more likely to purchase into than females?

To derive the “% more or less likely than the benchmark,” we apply the relative difference calculation above, utilizing males as the benchmark (Initial value)segment and females as the comparison (Final value) segment.

Using this calculation, we can see that Women are more likely than men to purchase products in every category except “Body.”