The nice thing with metrics is that you define them in one central place and can then use them dynamically for both deep dives and overviews of your data. Once defined, a metric automatically gets an explore mode for everyone to freely analyze and you can quickly create widgets for metrics.
Metrics are quick to set up, easy to maintain, and scales well as you grow.
More about the metrics concept
Admins can create, edit and delete metrics
Create new metric
Create a new metric by choosing New metric on the metric catalog page. You can also do it quickly by clicking '+' next to 'Metrics' in the menu. On mobile, go to the Metrics tab and tap the New metric button.
| Select one of the templates: | |
|---|---|
| Sum | corresponds to the total sum of a column |
| Count | counts the number of rows in a table |
| Ratio | divides one column by another |
| Unique Count | counts the number of unique values in a column |
| Custom value | opens up for you to write a single SQL expression |
| Custom ratio | divides one custom value with another |
Define the basics
A metric must have Name, Table, Value and Time
Select a Table
Press the Table field and select a table or view from your database.
You can also create a join to use instead of a table.
Select Value
Press the Value field and select a column from your table. This will be used to aggregate and calculate your metric value.
If you selected Ratio as template before, you will instead get two fields here, Converted and Base, that work in the same way and calculates a percentage.
If you selected a custom template you will instead write a custom SQL expression here.
Select Time
Press the Time field and select a column. This will be used to plot and analyse your metric over time.
Joins
You can create a join to use for your metrics instead of a table or view.
All types of joins are available: inner, left, right, and full. Join any number of tables based on your needs.
Create a join on the fly when setting up a metric. Simply click on the Table field and select ‘Create new join’ to get started.
Existing joins can be reused across multiple metrics. Also, joins are protected from change to ensure that your metrics remain accurate and reliable.
Preview metrics
When the basics are defined you will automatically get a preview of your metric. You will get a number and a graph to help you validate that the metric is set up as intended. You will also get the query Steep generates to give you full visibility and control. The query can easily be copied so you can run it in any other tool you make use for SQL.
Any time you make a change to your metric definition and want to preview the results – just hit Refresh.
Metric properties
Beyond the basics basics, table, value and time, there are many more options for defining a metric.
Dimensions
Dimensions provide breakdowns for your metric, for example by market, channel or product category.
To set up Dimensions, press the Dimensions field and select one or more columns.
View dimensions in explore mode
Once you’ve defined a dimension, it can be explored freely by every member in explore mode.
Filter
A metric can have one or more filters. It can be used to filter out unwanted rows in your metric, for example a specific country with noisy data.
A filter affects the whole metric including its slices.
Filters are based on simple conditional statements consisting of:
- A column (eg ‘country’)
- A condition (equal to, not equal to, less than, greater than, in or not in.)
- A value that should be matched (eg ‘se’)
- Together these are combined into a conditional statement. (Example: country = ‘se’ )
Slices
Slices are like sub-metrics. Once set up they provide an easy way to slice up your metric in explore mode, for example to find your high priority market Europe quickly. The source of truth is still the same as for the main metric.
Slices can have their own widgets. Slices are based on simple conditional statements, just like filters, consisting of:
- A column (eg ‘country’)
- A condition (equal to, not equal to, less than, greater than, in or not in.)
- A value that should be matched (eg ‘se’)
- Together these are combined into a conditional statement. (Example: country = ‘se’ )
Private metric
By default all metrics are public. To limit access to specific metrics they can be set to private and connected to a private teams.
Read more about Private teams and metrics
Description
Sharing metrics with a wider audience will sometimes require some context. You can add a description to go with your metric to make it easy for anyone in your organisation to understand what they are looking at.
Custom templates
You can create custom metric definitions with SQL for when the built-in templates are not enough. This lets you add virtually any type of metric to Steep. Create a new metric and chose one of the templates - Custom value with a single SQL expression - or Custom ratio to divide one value with another.
A list of all available columns is there to guide you next to the SQL expression field. Preview mode will also be a good help to validate the result of your query matches your intent.
Edit metric
- Go to any metric.
- Pick Edit in the top right corner.
Editing a metric updates all widgets connected to this metrics across you workspace.
Delete metric
- Go to the metric you want to delete.
- Pick Delete in the bottom of your edit metric fields.
This action will also remove all dashboard widgets and comments that may have been created for this metric.
