Metrics are final calculations to be displayed in dashboards or used in reports that take into consideration all events for all customers. Usually metrics provide information out of data. Metrics can be either simple aggregates, or formulas to show percentages, uplifts, etc.
Metrics are used either as individual values in dashboards, where they can be further specified (date range, comparison with previous period, goal progress), or in reports to compare different datasets.
Among the most used metrics are: Total revenue, Unique visitors, Unique buyers, Conversion rate (buyers÷visitors), RPV (revenue÷visitors), Revenue Uplift when comparing performance of A/B tests ((RPV of variant A – RPV of Control Group) x visitors of variant A)
When a metric that was previously defined is shown in dashboard, it is usually to check performance. A basic metric can show total revenue this month, and comparison with previous period to have better overview what does the number tell, or with a goal that was set to see progress towards it.
Metrics are crucial part of reports. They fulfill similar role as in dashboards, but serve for comparisons between the different specified categories, e.g. date periods, segments, different attribute values, etc.
In reports you can use previously defined metrics, or define a new one just for the report using the plus sign.
In reports, metrics can be visualized in multiple ways, e.g. in table and/or different types of charts, using also different modifiers.
Creating a metric
Metrics are created separately in Data & Assets or in initiative, or can be created just for purpose of a report, and are not saved for other use.
Metric can be displayed in multiple ways, e.g as simple numbers, different time formats, or containing special characters, such as %.
Metrics specification consists of multiple basic fields.
- Data aggregation that should be used (sum, min, max, avg, count, exists)
- Sum, min, max and avg work with attributes and their values
- Exists provide 1 or 0 value, depending on if customer matches the condition or not
- Whether to look for customer or events or their attributes
- When selecting event, another drop-down is presented to choose the specific event
- Attribute that provides the value (except when count or exists is used)
- When an event is chosen, 3 choices appear
- All – takes into consideration all events that match the condition
- First – takes into consideration only first occurrence of event per customer within specified constraints (e.g. date filter)
- Last – takes into consideration only last occurrence of event per customer within specified constraints (e.g. date filter)
When creating a formula, multiple metrics are defined, only separated by math operators.
- When counting unique customers that fulfill some criteria (e.g. visitors of special URL, buyers, etc.) it is better to use count event and select first occurrence, and not count customer and specify it in customer filter.
- When creating a metric, don’t specify date filter unless necessary. Date filter can always be specified in report/dashboard, and the one in metric itself can be forgotten about, providing misleading data.
- When using metrics in dashboards, always use goal or comparison with previous period, otherwise it does not provide much information.
- Use understandable names.