The new Metric Builder is a powerful new way to build metrics. It allows you to quickly get insights, navigate your data, and use Dashboard Filters to slice and dice all your metrics at the same time from the dashboard.
After you have connected a data source and pulled in a report or dataset, here is the updated process of creating a new metric.
Here is a short demonstration of the process of building a new metric:
Define your data
Since the new Metric Builder allows you to change your grouping, filtering, and aggregations on the fly, we need to do those types of calculations on the Chart tab and not using Transforms in the Data Pipeline.
Best practice is to split transforms into two groups, Definition transforms, and Presentation transforms.
Definition transforms are used to define what your data actually is. How does your company define a sale? For example, you might use a Filter transform to remove refunds. Other definition transforms could be adding a Tag Transform to label sales teams, or Formatting a Date.
Presentation transforms are used to specify the data you want to see in the chart and how you want it displayed. These transforms usually limit, group and remove data. For example, using the Filter transform to only see 30 days of data. Other presentation transforms could be Grouping by Week, Sorting Desc, and Pivot tables to get sliced data.
Definition transforms reduce the amount of data your metric can work with. For example, if your metric is looking at new sales, but you data is a list of transactions that include sales, refunds and upgrades, a Filter transform to remove the refunds and upgrades is a definition transform and correct. You do not want your metric to use refunds. You are using that transform to define what a sale is.
Presentation transforms change how the data is displayed. If you add Presentation transforms in the Data Pipeline it reduces the amount of data the metric can use and can make it harder to get the insights you want. For example, if you use a Filter transform to only show 30 days of data, then when you want to see your metrics for the whole year, the metric does not have access to that data, since you filtered it out. Using the enhanced metric builder to show 30 days of data instead makes it so you can change the date filter on the fly from the dashboard.
Select the calculation you want on each Value column, the grouping, and the primary date column
In the Chart tab of the Metric Builder, we give you the ability to filter and group by date. So you need to tell Grow what date column to work off of, and how to group and calculate (aka Count, Sum, Average) for each column you want to display.
Here is how to set those two things in the Chart tab of the metric builder.
- For each of your Values, select how you want it to calculate when it groups. You can pick from Sum, Count, Average, Min, Max, Median.
- Select the column you want to use for the grouping (x-axis). The Metric Builder will automatically select the first date column in your data, but you can change that to any other column
- Go to Date Range and select the date column you want to use as your date range filter. This can be any date column, even if it is not charted.
Add any other customizations to your chart
You can also choose how the chart is displayed:
- Add any dynamic key values. These are also affected when filters are applied.
- Select what default date range is displayed. Each chart on the dashboard can show a different date range, and when a date filter is applied, they all update to show that filtered range.
- Slice the data by the values in another column
- Compare two date ranges in the chart.
- Sort the data in ascending or descending order, and limit how many rows display. For example, you could display your top ten lead sources, or your bottom five sales reps.
As you add the values and customization to the chart, you might notice that the data in the table doesn't change. If you want to see how the data is arrange to create the chart, you can toggle between the chart and table views at the top right of the chart preview area: