When you need a calculation

Before writing a formula, check whether the metric or attribute you want already exists. It usually does.

Before writing a formula, check whether the metric or attribute you want already exists. It usually does.

Check these first

1. The metric list. Zendesk ships a great many, including business-hours versions of every time measure.
2. The standard reports. Somebody has probably already answered this question.
3. A different aggregator. Median instead of average is a setting, not a formula.
4. A filter. Much of what people write formulas for is a filter on an existing metric.

When a calculation is genuinely needed

• Your own definition of something. What counts as an escalation here, which ticket types are "simple".
• Grouping values into categories. Twenty products into four families.
• A ratio between two metrics that is not offered directly.
• A flag marking tickets that meet several conditions at once.

Attribute or metric

An attribute produces a category to group by: "product family". A metric produces a number: "tickets solved within target". Deciding which you are making is the first step and is where most confusion starts.

Put it on the dataset, not in one report

A calculation defined on a cloned dataset is available to every report. The same formula retyped into six reports will diverge, and one of the six will be wrong.

Write down what it means

A calculation encodes a business decision. In a year the formula will still be readable and the reasoning will not, and somebody will change it without knowing what it was for.

Keep them few

Every calculation is something to maintain and something that can silently break when a field changes. Build the ones that earn it.

See also

• Creating a standard calculated attribute
• Creating a calculated metric
• Writing formulas