Introduction to Time Concepts in Metrics

Time is a concept applied to metrics to provide more accurate information and analytics. There are many different concepts from points in time to spans of time to averages, and it can be confusing to understand the difference between the different time concepts, and how they are applied to your metrics. This article aims to solve that by providing a detailed introduction to the basics of time concepts in metrics. We recommend starting with Introduction to Metrics in One Model before continuing. 

Time Concepts include:

• Point in Time
• Cumulative or Span of Time
• Average

Let's discuss each of these in turn starting with Point in Time. 

Point in Time

Point in Time helps to answer key questions such as:

• How many people worked in my organization as of the end of last year? As of the beginning of this year?
• What is the total service tenure for my organization today? 
• How many applications are currently active?
• How many goals are still actively being worked on in my organization?

Point in time measures traditionally count / sum data as of the start of a time period or end of a time period. You can apply Point in Time measures to years,  quarters, months or days, or any level of your time dimensions.

Tip > if you are building a metric from a period based fact table (prefixed prd) then you are likely to apply the Start of Period or End of Period “Period Filter” in the Aggregate Context section.

Cumulative or Span of Time

Cumulative measures sum data across a Span of Time and help to answer key questions such as:

• How many terminations have I had this year?
• How many applications did I have last month?
• How many goals have been created this year?  

Tip > if you are building a metric from an event driven fact table (prefixed evt) then you are likely to NOT apply the Start of Period or End of Period “Period Filter”.

Average

Average is similar to the Cumulative concept, but instead of summing data across a span of time, an Average measure will average the data across a span of time. These measures are most commonly used as part of a calculated metric, e.g., Average Headcount is used as a denominator in many calculated metrics like  Termination Rate, Hire Rate and Promotion Rate.

It is important to use an average in these metrics, rather than Point in Time because Point in Time are more likely to overstate or understate the final result.

Time Functions in Metrics 

These change the time context of a metric to serve different analytical needs. Typical time functions include the following:

• Rolling Months
• Time Projected
• Previous Period
• Same Period Last Year
• Year To Date 

Rolling Months (and number of months)

Rolling Months are months going backward, to the day.  

For example, suppose today's date is August 10, 2023, and you run a query for “Today” using a metric with a Time Function Type = Rolling Months and the Number of Rolling Months =12, it will sum the data going back to August 10, 2022. 

If you run a query with the same metric for 2022 and 2021, it will display the same values as a regular metric (without Rolling Months) would, since it's going back 12 months from the end of each selected period (year).

Time Projected

Projects the value for the selected time period based on the number as of the current point in that time period.  

For example, if today is the 15th of the month, running a query using a metric with the Time Function Type = Time Projected, will divide today’s metric result by 15 and multiply by the number of days in the month, quarter, or year (as set in the query) for the projected value. 

For past time periods, it displays the actual value since that time period is complete. Time Projected only works when viewing data by month, quarter, or year, and not for day.

Previous Period (and To Date)

This shows you the previous period to the selected reporting period.  

For example, running a query for this month using a metric with the Time Function Type = Previous Period, will display the result from the whole previous month.  The same goes for year, quarter, and day.  

If your metric has the Time Function Type = Previous Period AND the To Date field selected, then if today is the 15th January, and your query is set for this month, then the result will be the result up to 15th December only. 

Want your query to show the difference From Previous Period? Create a calculated metric like “[Headcount EOP] minus[ Headcount EOP - Previous Period]”

Same Period Last Year (and To Date)

This is the same time period as what's selected in the query, only for the prior year.  The To Date option only impacts the current time period.  

For example, if today is August 10, 2023 and you have selected 2023, then a metric with Time Function Type = Same Period Last Year is the total for all of 2022. If the metric has Time Function Type = Same Period Last Year AND the To Date field selected, the result will be for January 1, 2022 through August 10, 2022.

Year To Date and Annualise, and Restrict to current period to Date

Year to Date (YTD) provides cumulative year to date results . When looking at a yearly basis it will be the YTD through the current date, e.g. if it is currently May 13, YTD is Jan 1 to May 13, but when looking at it on a Quarterly, Monthly or Daily basis it will be the YTD through the period selected.  

The Year to Date Time Function is meant to allow users to easily compare YTD this year to prior years. At the beginning of the year, Last Month or Last Quarter will not work because those time periods do not exist in the current year yet.

• Annualise is a separate option, and is often selected in conjunction with Year to Date for meaningful analyses.  If Annualise is selected, it estimates the number based on the reporting period selected. For the current/incomplete year, given that it's annualising based on what data is available, the annualization may not equal the actual number of hires for the year. We have an interesting blog article about forecasting v annualization. 
• Bear in mind that YTD + Annualise will calculate differently from how Time Projected is calculated, specifically at more granular levels like Quarters or Months.  While Time Projected will only pull data from the selected Quarter or Month and calculate a projection with that given data, YTD + Annualise will always calculate a number on a yearly level.  
• NOTE: Applying the “Restrict to current period To Date” option with YTD will notably affect how data is displayed for previous years, as no data for time that hasn't passed in the current year will be available.  

For example, if the current date is November 12th, when looking at Terminations YTD unrestricted on a yearly level from the current year and the previous year, the figure for the current year’s Terminations will be the sum of all terminations between January 1st and the current date (November 12th), while the previous year’s figure will be the sum of all terminations that had occurred in that year (January 1st through December 31st).  When “Restrict to current period To Date” is applied, however, while the current year’s Termination YTD figure won’t change, the previous year’s figure will change to the sum of all terminations between January 1st and the current date last year (November 12th), thus providing a helpful comparative view. 

Read our article About Time Functions for a more detailed explanation. 

Tip > go to your Storyboard Library and search for “Time Function and Period Filter Examples”. This is a pre-built storyboard with lots of query examples. Not seeing it? Contact your Customer Success Lead to discuss.