Now that you are more familiar with Auto Insights, you will face insights generated on averages and ratios. This article will guide you through how to understand the insights Auto Insights has called out.

Let's get the terminology out of the way first:

**Averages (metric)** - This is automatically calculated within Auto Insights by taking the sum of measure divided by a unique count of a segment. Some examples include Average Sales per Department, or Average Incidents per Category, etc.

**Ratios (metric) **- This is calculated within Auto Insights by taking the sum of one metric divide by the sum of another measure. Some examples include Average Price (Sales/Quantity), or Completion Rate (Completed_Incidents/Incidents), etc.

**Contributions **- They show the influence of each level (in a specific segment) on the total change in metric.

Within What Caused This section for averages and ratios are 3 main columns, the main metric, numerator changes, denominator changes.

When you take the sum of contribution *within each segment*, it will add up to the metric trend change.

This helps you visualise for each segment level (in this case - supermarket departments), which department positively contributes to the change in sales, and which one negatively contributes to the change in sales.

Essentially contribution takes into consideration the ranking of that department compared to the others in that period vs their ranking in the previous period.

**HOW DO I INTERPRET CONTRIBUTION WITH RESPECT TO MY METRIC?**

Let’s take a look at this example where we see Loyalist customers had contributed to a 0.31 increase in overall average price for that month (0.2 increase). Promiscuous customers on the other hand, had contributed to a 0.12 decrease in overall average price.

The main difference in contribution signs:

**(Loyalist vs Promiscuous)** is due to the shift in the base. So even though per item the average sales dollar had increased, overall, because of the large drop in the quantity of items purchased, this resulted in Promiscuous customers dropping in their average ranking compared to last month, hence a negative contribution.

**(Promiscuous vs First Time Buyer) **It is different for First Time Buyer as there’s no change its ranking compared to last month, thus a positive increase.

**WHAT ABOUT THE SCENARIO WHERE MY METRIC HAS DECLINED FROM LAST MONTH BUT SHOWS A POSITIVE CONTRIBUTION?**

Take the Deli Department in the earlier screenshot as an example, month on month, average price had declined by 0.25, but it still has a positive contribution.

Essentially, this dip in Deli's average price caused the proportion (ranking) of Deli sales to increase in the latest month compared to last month, Thus, the decline in deli's average price had a positive contribution with the increase in sales, pushing up the average price across all departments.

**CAN YOU PROVE THE CONTRIBUTION CALCULATION?**

Let's calculate the contribution number for meat department in this example (total quantity in Jul was 79500 with a -10,762** **change from last month):

Meat contribution = (4.34*11888)/79500 - (3.84*12080)/90262) = 0.65 - 0.51 = +0.14

If you have any questions, please contact us at support@hyperanna.com.