Skip to main content
Gainsight Inc.

Bionic Rule Trend Calculation

The trend calculation function enables comparison of data between any two defined periods. This function is available with Bionic Rules Transformation Task, on clicking +Formula. An example use case is finding trends in usage metrics or support cases.

For more information on Formula Builder, refer to Formula Builder in Bionic Rules.

Trend Template (Non Percent):

The following list provides descriptions of all arguments in the Trend function:

  • Period (inside the bracket): The period type such as days, weeks, months etc..
  • Period Length: The length of the period.
  • Source Field: The metric which is compared for the selected periods.
  • Trend Indicator: The date identifier - this date is considered based on the Show fields (Date field) selected in the Transformation task for which trend is calculated.

Notes:

  • Rule Run date is referred to identify the date record that qualifies as current period record.
  • This is not the current date.
  • Aggregation Method: The method used for aggregation such as average, sum etc..
  • Prior Period: The number of earlier periods to be included for comparison.
  • Adjust Missing Data: Set this to True for adjusting the missing data or False otherwise. Missing data is either absence of the record or a null value or empty value for the metric selected. When set to:
    • True - the record for missing data is excluded by the Aggregation Method
    • False - missing data or null value is considered as 0 (zero)

Example

Consider the following sample setup with rule date - (1/9):

Trend(Period ( DAYS , 2 ), Clicks , Data Date , AVG , 3 , false )

Data Set

Data Date (Current Date to Older)

Clicks

1/8 1
1/7 0
1/6 5
1/5 3
1/4 8
1/3 null
1/1 6

Trend calculation for the provided data set.

Past 2 Days AVG minus Prior 6 Days AVG = Trend

Period Length 2 x Prior Period 3 = 6 Prior Days.

 

For Past 2 Days (1/8 - 1/7, as the rule run date is 1/9):

1 + 0       1

--------  = ----- = . 5

  2           2

For Prior 6 Days (1/6 - 1/1):

5 + 3 + 8 + 0 + 0 + 6        22

---------------------------- = ------- = Approx. ~ 3.67

              6                        6

As the record for 1/3 holds a null value and 1/2 is missing, 0 is considered instead. ​​​​​

Trend

.5 - 3.67 = -2.17

Alternative Options

The following examples use the same sample data used in the previous section:

  • Use Trend Percent,  (( Trend / Prior Period ) *100) (In this example, it’s prior 6 days)

-2.17 / 3.67 = -.5913 * 100 = - 59.13%

  • Set 'Adjust Missing Data' to “True”:

For Past 2 Days (1/8 - 1/7):

1 + 0       1

--------  = ----- = . 5

  2           2

For Prior 6 Days (1/6 - 1/1):

     5 + 3 + 8 + 6               22

---------------------------- = ------- =  5.5

              4                        4

As Adjust Missing Data is set to 'True' and the record for 1/3 holds a null value and 1/2 is missing, this record is excluded from Trend calculation.

Trend = .5 - 5.5 = -5.0

Trend Percent = (-5.0 / 5.5) * 100 = -90.9%

  • Was this article helpful?