Skip to main content
Intelligence is only available on Business tier plans.

Overview

A metric is a saved calculation over one event in your schema. It answers a question like “How much revenue did these members generate?” or “What share of this audience made a purchase?” Every metric combines three choices:
  • Event — The action you’re measuring, such as a purchase, a page view, or an email click.
  • Aggregation — How Hightouch combines those events into a number, such as counting them, counting the members who performed them, or summing a property like order total.
  • Filters (optional) — Which events count, such as only completed orders or only purchases in one product category.
You can build the same calculation inside any chart. Saving it as a metric means the definition lives in one place: anyone who picks Total revenue gets the same event, aggregation, and filters, and doesn’t have to rebuild them or guess how a teammate set them up. Each metric belongs to one parent model. You can use it in charts, experiments, and audiences built on that parent model.

Before you begin

  • Your workspace must have a schema with a parent model.
  • The parent model must have at least one related event model. Metrics only measure events; parent model columns, traits, and related models that aren’t events don’t appear as options.
  • You must have permission to create schema objects on the parent model. Parent models you don’t have permission for appear in the list but can’t be selected.

Create a metric

1

Start a new metric

  1. Go to Intelligence > Metrics.
  2. Click Add metric.
2

Select a parent model

Select the parent model whose members you want to measure. The metric is available only on charts, experiments, and audiences that use this parent model.
3

Define how the metric is calculated

  1. Under Event, select the event to measure.
  2. Under Aggregation, select how to combine the events. See Choose an aggregation.
  3. If the aggregation uses a property, select the property under Sum by or Count by.
  4. Under Formatting, choose how values display. See Choose a display format.
  5. Optional: under Filter conditions, add conditions on event properties to limit which events count. See Decide which events count.
  6. Click Continue. The metric definition form for a Revenue metric, with the Order Completed event, the Sum of property aggregation, USD currency formatting, and ORDER_VALUE under Sum by
4

Name and save the metric

  1. Enter a Metric name. Use the name people search for in a chart, such as Total revenue or Purchase rate.
  2. Optional: enter a Metric description that says what’s included and excluded, such as “Completed orders only, excludes refunds.”
  3. Click Finish.
Hightouch opens the metric’s page, where you can edit the definition later.

Save a metric from a chart

If you’ve already built the calculation in an Insights chart, you can save it without starting over:
  • From an event — In the chart’s Metrics panel, after you select an event and configure it, click the star icon (Save metric). Enter a Name and optional Description, then click Save metric.
  • From a modified saved metric — If you change the filters on a saved metric in a chart, the metric shows a Modified badge. Click Create metric to save the changed version as a new metric, or Clear changes to go back to the saved definition.

Choose an aggregation

The aggregation decides what the number means. Pick it by the question you’re answering: A few distinctions decide most choices:
  • Total events vs. Unique members. A member who buys three times counts three times in Total events and once in Unique members. Use Unique members or % of audience when you care about reach; use Total events when you care about volume.
  • Average of property vs. Sum of property per member. Average of property averages across events, so frequent buyers weigh more. Sum of property per member totals each member’s value first, then averages across members, so it answers “how much is a typical member worth?”
  • % of audience is the share of members in a chart’s segment who performed the event. Segment the chart by an audience to see that audience’s rate, such as its purchase rate.
The property-based aggregations need a property to calculate. Sum of property, Average of property, and the per-member versions show a Sum by field that lists only number properties on the event. Distinct count of property shows a Count by field that lists every property on the event.

Choose a display format

Formatting sets whether the metric’s values display as a number, a percentage, or a currency. It doesn’t change the calculation. Insights charts, experiments, and audience overviews show the raw value without this format, so a revenue metric set to USD still appears as a plain number in a chart.
  • Choose currency for money values, then select the currency code, such as USD or EUR.
  • Choose percentage when the property already stores a rate, such as a discount rate stored as 0.15.
  • Use the decimal place buttons to show up to four decimal places. The preview shows how a sample value looks.
The formats available depend on the aggregation. Count-based aggregations (Total events, Unique members, and Distinct count of property) display as numbers. % of audience displays as a percentage. Sum and average aggregations can use any of the three.

Decide which events count

Without filters, a metric includes every row of the event. Add Filter conditions when part of the event shouldn’t count toward the KPI, for example:
  • Only orders where status equals completed, so refunds and cancellations don’t inflate revenue.
  • Only page views where page_type equals product, so you measure product interest rather than all browsing.
  • Only email clicks where campaign_type equals promotional.
Put a filter in the metric when it’s part of what the KPI means: “revenue” that includes refunds isn’t revenue. Leave filters out when they describe a slice someone might want to look at, such as a single region. People can add those filters in a chart, or use Group by to compare slices side by side.

Example: track revenue from completed orders

Suppose your schema has a Customers parent model with a related Order Completed event that includes ORDER_VALUE and STATUS columns. Your team wants one revenue number everyone uses.
  1. Go to Intelligence > Metrics and click Add metric.
  2. Select the Customers parent model.
  3. Set Event to Order Completed.
  4. Set Aggregation to Sum of property, then set Sum by to ORDER_VALUE.
  5. Under Formatting, choose currency and select USD.
  6. Under Filter conditions, add a condition where STATUS equals completed.
  7. Click Continue, name the metric Total revenue, and add the description Sum of ORDER_VALUE for completed orders.
  8. Click Finish.
Anyone building an Insights chart on Customers can now select Total revenue, segment it by audience, and group it by a property such as state. To measure revenue per customer instead, create a second metric on the same event with Sum of property per member.

Use metrics across Intelligence

After you save a metric, it appears wherever you choose a metric for the same parent model:
  • Insights charts — Select the metric in the chart’s Metrics panel. The aggregation is fixed by the metric, but you can adjust the Measurement window and Filter events for that chart without changing the saved definition.
  • Experiments — Choose the metric as an experiment’s Primary metric, which decides whether the treatment won, or add it under Secondary metrics to track other outcomes.
  • Dashboards — Save a chart that uses the metric, then add the chart to a dashboard.
  • Audience overviews — Select the metric under Primary metrics in an audience’s settings to show it as a KPI on the audience’s overview page. When you create a metric, you can also pick audiences under Assign to audiences.
To see where a metric came from while you’re working in a chart, click the external link icon (View metric) next to it.

Tips and troubleshooting

An event is missing from the Event list

The Event list only includes event models related to the parent model you selected. If an event is missing, check that it’s set up as an event model with a relationship to that parent model in your schema. If it belongs to a different parent model, create the metric on that parent model instead.

A property is missing from Sum by

Sum by lists only number properties, because summing and averaging need numeric values. If a column such as order_total is missing, check its type in the event model. A column stored as text needs to be cast to a number, for example in the event model’s SQL.

A saved metric’s aggregation can’t be changed in a chart

A saved metric’s aggregation is part of its definition, so charts don’t let you change it. Edit the metric on its page in Intelligence > Metrics, or select the event directly in the chart and choose a different aggregation.

Metric edits apply everywhere the metric is used

Charts, experiments, and audience overviews that use a saved metric as is read its current definition, so a change to the event, aggregation, or filters affects all of them. If you need a variation for one analysis, create a new metric rather than editing a shared one. The exception is an experiment where someone changed the metric’s settings, such as its filters, while selecting it. That experiment keeps its own copy of the definition and doesn’t pick up later edits to the saved metric.

Deleting a metric can’t be undone

Deleting a metric also deletes its analytics and performance data. To delete metrics, select them in the Metrics list and click Delete, or open a metric and select Delete from its menu.

Common errors

Metric configured incorrectly

Cause: A required setting is empty when you click Continue while creating a metric. Every metric needs an Event, and property-based aggregations also need a property under Sum by or Count by. Solution: Select an event. If the aggregation uses a property, select one. Fields that still need a value show “Event is required” or “Column is required.”

Error in the form when saving a metric

Cause: You clicked Save metric on an existing metric’s page while a required setting is empty, such as a property-based aggregation with no property selected. Solution: Fill in the field that shows an error, then click Save metric again. To go back to the saved definition, click Discard changes.
Last modified on November 10, 2026