# Cube Calculation vs Cube Metric Accounts

This article includes suggestions and workarounds. Content may not be accurate for all use cases or represent best practices for the latest release.

## Question

What is the difference between a cube calculation account and a cube metric account?

## Answer

Cube Calculation accounts and Cube Metric accounts primarily differ in the way in which they roll up data. There are also performance benefits to Cube Calculation accounts, since they will not attempt to evaluate in every intersection automatically, whereas a Cube Metric account will automatically evaluate for all possible intersections.

**Cube Calculation** accounts roll up data in the same way that standard custom accounts do: by default they sum at level, dimension, and time rollups. However, it is also possible to change their roll-up type as needed to "average" or "weighted average".

**Cube Metric** accounts roll-up in the same way as standard Metric accounts, and are often used for financial ratios. In roll-up periods, they re-evaluate the formula rather than summing the calculation in each contributing month and level.

For example, lets say the formula to evaluate is as below:

**div(ACCT.A, ACCT.B)**

In January, February, and March, ACCT.A = 1, and in the same time periods ACCT.B = 2.

1. For a **Cube Calculation** account, the quarter will sum the evaluation in each individual month:

- Jan = 1/2, Feb = 1/2 , Mar = 1/2
- Q1= .5+.5+.5 = 1.5

2. For a **Cube Metric** account it will run the evaluation in the Q1 rollup separately, taking the values in each month combined for both accounts and then running the calculation:

- Jan = 1/2, Feb = 1/2, Mar = 1/2
- Q1 = (1+1+1)/(2+2+2) = 3/6 = .5