Creating custom metrics: Which configuration options should I choose? (for data sources)


We’re improving the metric creation experience. Our recent redesign lets you preview changes while you create and edit custom metrics. We’ve also combined raw and modelled data sources into a single object called a data feed.
If you see Data Feeds in the left navigation sidebar, go to this article.
These new features are being released gradually. If you don't see them in your account yet - no worries - they’re coming soon!


Not sure which configuration options to choose when creating custom metrics? This article describes some fundamental properties of metrics and data sources and explains how the two relate to each other. Understanding these basic concepts and applying the question/answer technique described in this article should give you the knowledge and tools you need to successfully create custom metrics.

This article includes:

What do I need to understand about metrics?

Metrics are data artifacts that capture changes to data over time. This is one of the primary reasons so many business leaders rely on metrics. Important decisions aren’t solely made based on data - they’re made based on monitoring trends in the data.

Each individual metric measures a single value and tracks how that single value changes over time.

Metrics collect and store data history. Each time your data source is updated, either manually or via automated refresh, new data enters into the metric and either adds to or replaces the previous data. This is one of the most powerful things about metrics. As time goes by, metrics evolve enabling you to do comparisons to previous time periods and gain insights from the trends in your data.

What do I need to understand about data sources?

Data sources don’t store data history. Each time a data source refreshes, all of its data is completely replaced. Data sources feed data into associated metrics so when the data source is updated with new information, the metric is also updated with the new information.

When you write a query, the resulting data source includes ALL of the data you requested. A metric includes a subset of that data. For example, a data source may be a table that includes everything from your customer database but a metric that measures your current number of customers would only refer to one column of that data - the customer name column.

Which questions and answers will guide me in my configuration choices?

Asking and answering the following questions and answers, based on the metric you want to visualize and the associated data source, will help guide you in making the right configuration choices in the custom metric wizard.

  • Question 1: What is the one thing you want to measure for this metric? Decide which single value you want your metric to track. Some common examples are “revenue”, “page views”, and “new customers”. (See below.) Note: The value isn’t always found in a column in your data source. For example, it could be a count of rows in your data source.

  • Question 2: When were the values recorded? Depending on the value you want to measure, this may be a date column in the data source or it may be the current date (the time when the records were fetched, also known as “import time”). (See below.)

  • Question 3: When your data updates, do new values add to or replace previous values? Answering this question will help you choose whether your data processes updates as transactional values, current values, or as a periodic summary of values. If new values add to your previous values, then your data is transactional. If new values replace your previous values, then your data is either current values or periodic summary. (See below.) Note: When you create custom metrics we do our best to select the appropriate data shape for you.

Learn more

Go here to find some examples for commonly created metrics, using the question and answer method described above.

Related articles

Have more questions? Submit a request