Choosing the right data structure for your custom metrics (NEW)

 

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 you're in the right place. If not, go to this article instead.

 

When you create a custom metric, in the step where you choose the Data structure, you determine the values to use from your historical data by selecting one of the following: Include all values, use latest values, or use latest values per period.  (See below.) Your choice determines the data points your metric will use when calculating the metric value for a time period.

The goal of this article is to help you understand which option is right for your data. If you’re still unsure which selection is best – don’t worry, you can go back and change your selection later by editing the metric.

Include all values

  • This option is applicable to data where every element, for example, every line in a table, is a unique transaction. Data begins from when you start collecting it and goes up until the present.
  • This type of data is collected in real time, so it usually includes a date and a timestamp (a specific time within the day).
  • Select this option if you want to include every value for a time period to create a total of values. For example, you may have a spreadsheet that includes all sales transactions for the month of January. In this case, you want to sum all values to get the total values for the month.

Use latest values

  • This option is applicable to snapshot data - the value of the metric right now.
  • Select this option if your data includes latest values for specific points in time and new values replace older values. For example, if you're tracking social media followers, each import (refresh) gives you the latest value, which is the total of all values for the time period.

Use latest values per period

  • This option is applicable to data that starts at zero and accumulates over a set period of time (for example, a daily or a weekly accumulation). At the end of the time period, it resets to zero.
  • When you select this option, you also select a time period over which the value applies to (hour, day, week, month, or year).
  • The Google Analytics "Sessions by Day" data, is an example of when you'd select this option and choose "day" as the time period. The value at any point in time is the summary up to that point in the day. The latest value is the complete summary for the day. Your data is collected again on the following day, beginning at zero. If you choose “Month” as the time range in your visualization, the metric’s value for the month is a sum of the last value for each day in the month.
Have more questions? Submit a request