How to create Metrics

Metrics (or Measures) are core to viewing data. Learn how to create them here.

If you are a One Model user, you have been exposed to Metrics. Metrics are quantitative measurements. The number of employees hired and the number currently employed at your company are both metrics.

While this article does not cover creating Dimensions, it's worth mentioning them here to avoid confusing the two. Dimensions are attributes of your data. Work location of employees is a dimension. When viewing data in One Model, you will almost always be viewing metrics that are grouped or filtered by dimensions.

To recap;

Metrics = quantitative measurements such as the number of employees hired.

Dimensions = attributes of your data such as the work location of employees.

Creating a Metric

To get to the Metrics screen where you will create a metric,

  1. Go to the Explore tab

  2. Click on Create / Edit next to Metrics in the left hand menu.

 

Metrics Screen

This will open the screen for creating a new metric.

3. To edit or copy an existing metric, select the pencil icon next to the existing metric in the left menu while on the Create / Edit screen.

 

 

Please note: It's almost always better to copy an existing metric since it gives you a good starting point. The risk of doing so however is that you unintentionally modify the metric you're copying rather than copying it. If you modify an existing metric and save it, there is no way to undo it.

 We strongly recommend that you download and save the Admin Report> Metrics before you make changes or create new metrics so you will have a reference to go back to if you make a mistake. 

Use caution not to unintentionally modify an existing metric!

Menu Bar

At first, the menu bar shows

  • Create a new metric

  • Cancel changes

  • Save metric

As you start making selections, the options in the menu bar expand to include

  • Delete metric

  • Copy metric

Settings

The configuration of the metrics is performed in the Settings section.

  • The first five settings control the administrative aspects of the metric. This includes what it's named, where it can be found, and who has access to it.

Name (required) - You can name your metric whatever you want but it's best to be consistent with your naming conventions.

Description - Provide information about how the metric is calculated. 

Category (required) - This is where the metric will live in the list of available metrics in Explore. A category is required when creating a metric. This can be a new category or an already existing one.

Subcategory - This is the second level of where the metric will live in the list of available metrics in Explore. Although a category is required when creating a metric, a subcategory is not required. This can be a new or existing subcategory.

Metric Permissions (required) - Here you select the roles that should have access to this metric. At a minimum, select your own role and the Admin role.

  • The settings from this point onward define how the metric is calculated.

Metric Definition (required) - A metric definition can be as simple as referencing a particular column but it can also get complex. Both Table Metrics and Calculated Metrics can be created. A table metric is derived from columns in the data model whereas a calculated metric is a metric built off of other metrics. Arithmetic calculations and static numbers are available to be used for all metrics. The DATEDIFF function can only be used for metrics based on tables where the dates are in the same table.

  • For a table metric, select a column from a table in the Tables section of the list of entities on the left. An example of a table metric is Hires.

  • A calculated metric is created by adding an existing metric or metrics from the Metrics section on the left list. An example of a calculated metric is Termination Rate, which is Terminations / Average Headcount.

  • Arithmetic calculations, static numbers, and the DATEDIFF function can be utilized in either type of metric. DATEDIFF allows you to capture metrics such as Tenure or Age.

  • NOTE: Unlike Excel, where you will get an error (#Div/0!) when dividing by a denominator that is zero (0), the One Model platform will show a zero (0) result for the calculated measure.

Measure Format / Precision / Decimals (required) - It's up to you what you select here, although if your metric is a count or sum of whole numbers, you won't need decimals.

Measure format options include: number, percent or currency. Precision is the number of decimal places the calculation will operate by when running the calculation, and is important for compensation-related metrics. Decimals control how many decimal places to be shown in the final number.

Aggregate Context (required) - This section only becomes visible once you create a Metric Definition based on a Fact Table. This section will not show calculated or derived metrics. Knowing whether you're creating a cumulative, average, or point in time metric is important to consider when selecting your aggregate context.

To better understand the different types,refer to : Difference between Cumulative, Average and Point in Time Input Measures

  • Aggregate Type - If your metric is being created from a non-numeric column, the only options here will be Count and DistinctCount. You might use DistinctCount to calculate the number of Job Applications on which there was activity during the selected time period. Since there can be more than one row per application, DistinctCount of Application ID prevents each one from being counted more than once. If in your metric definition, the value generated is numeric, you have significantly more options. Sum and Average are the most commonly used.

  • Period Filter - If None is selected here, it will use the time context you select in your queries in explore at face value. For event based metrics such as Hires or Terminations, None is the correct choice since you want to see how many happened throughout the selected time period. StartOfPeriod and EndOfPeriod are used for point in time metrics, such as Headcount. Period filter is more helpful for base metrics, not events.

Filters -Editing dimension filters in the Filters section will switch the items displayed in the left menu from Metrics and Tables to Dimensions. Choose the dimension(s) you wish to filter by and the values you would like to include. Currently, users cannot use dimension filters to exclude values.

Time Functions - Time Functions work in combination with Time Models in explore to create robust methods of exploring data across time. While more often than not, you'll be leaving None selected here, understanding how time functions work is useful. [Read our detailed document about these somewhat complex functions here.]

Restrictions - Only visible to One Model Admins so you will need to speak with your Customer Success Lead. 

Example Metrics

Input Metrics

Creating Headcount (EOP) Metric

To create the Headcount (EOP) metric, firstly select explore and click on create/ edit metric.

Step 1: In the Build a Base Metric Page, enter the name of the metric as Headcount (EOP).

Step 2: Provide a definition about what the metric is used for such as "The headcount at the end of selected time period".

Step 3: Select or create a category as "Headcount" and subcategory as "Headcount" for the Headcount EOP metric

Step 4 & 5: From the Tables select the "Employee Headcount". This will appear under the Metric Definition. This metric is an input and does not require an calculation.

Step 6: Select the Measure Format as "Number," Precision as "18" and Decimals as "0"

Step 7: Set the Aggregate Type as "Sum".

Step 8: Set the Period Filter as "EndofPeriod" .

Step 9: The filters should be set as Dimensions>Calculations>In Counts as it always applies to metrics that we want to count during periods that the employees headcount is 1.

Step 10: The Time Functions type should be set as "None".

Step 12: Click the Save button to save the metric

 

 

 

 

 

Note: that you will need to choose at least one Data Access Role under Metric Permissions to save this metric.

 

Creating Headcount (SOP) Metric

Similarly to the creation of Headcount (EOP), the user can follow the same steps however make the following changes to the steps above

Step 1: In the Build a Base Metric Page, enter the name of the metric as Headcount (SOP).

Step 2: Provide a define what the metric is used for such as "The headcount at the start of selected time period".

Repeat Steps 3 - 7 as above.

Step 8: Set the Period Filter as "StartofPeriod" .

 

 

External Hire Metric

The user can follow the same steps however make the following changes to the steps above

Step 1: In the Build a Base Metric Page, enter the name of the metric as Hires- External.

Step 2: Provide a define what the metric is used for such as "The total number of external hires during the selected time period".

 

 

Calculated Metrics

Creating Headcount (Average) Metric

The Headcount (Average) is a calculated metric based on predefined metrics. It is Calculated on a daily basis to provide a true average headcount as persons move through the organization.

Step 1: In the Build a Base Metric Page, enter the name of the metric as Headcount (Average).

Step 2: Provide a define what the metric is used for such as "The headcount in an organization during the selected time period".

Step 3: Select or create a category as "Headcount" and subcategory as "Headcount" for the Headcount EOP metric

Step 4 & 5: From the Metrics create calculation by selecting the metrics zCalculations>Headcount>Headcount (Total). Note: "Headcount (Total)" and "Days in Period" metrics need to have been created prior to creating the Headcount (Average) metric.

Step 6: Select the Measure Format as "Number," Precision as "18" and Decimals as "2"

Step 7: There should be no filters selected in the Filter section as well as no Time Functions selected.

Step 8: Click save to save the metric

 

 

Watch this short 4-minute video to see the steps in action - How to create a Metric (Video)

Was this article helpful?

0 out of 0 found this helpful

Comments

0 comments

Please sign in to leave a comment.