< All Topics

Relative User Activity Thresholds

Relative User Activity Thresholds are for grouping users into Low, Medium, High and Very High usage categories

  • They are good for​;
    • Comparing departments, groups, countries​
    • What % of their user Low, Medium, High and Very High​
  • They enables report such as;
    • “Where are the highest Calls activity users located?” ​
    • “How does London usage compare to France usage for Team Chat?”​
  • Category boundaries based on global usage, per modality, for the day/week or month
  • Boundaries change as global usage trend changes
  • Activity grouped into classifications based on percentiles​
  • The usage thresholds for each classification will change over time and for each modality.
  • Zero usage users are not included in classifications

Classification Boundaries

We group users into either Low, Medium, High and Very High usage categories using the following percentiles.

These percentile classifications are based on Modality System previous experience with actioning usage data to drive improved adoption. ​

ClassificationFrom PercentileTo PercentileDescription
Low0th<= 10thUsers which need the most training and assistance
Medium> 10th<= 60thUsers should be encouraged via indirect adoption methods such as posters, emails, self-service training etc.​
High> 60th<= 90thContains “Teams Champions” and those that can help others achieve
Very High> 90th<= 100thExtreme users which may be candidates for good news stories, case studies etc.​

Ranges are inclusive of the upper bound, but exclude the lower bound. For example, the Medium range can be read as “greater than the 10th Percentile and less than or equal to the 60th percentile”

Worked Example of Calculation

Consider the following table of usage for Team Messages;

UserTeam Messages
Adam Barr1
Alan Steiner8
Chris Norred22
Christine Koch6
Dan Jump14
Dan Park1
Diane Tibbot3
Frank Martinez2
Jeff Hay7
Karim Manar16

Percentiles for usage are calculated using the PERCENTILE_CONT function in SQL, the equivalent in Microsoft Excel is PERCENTILE.INC. The percentile thresholds from this data are calculated as follows;


This results in the following classification of those users;

UserTeam MessagesClassification
Adam Barr1Low
Dan Park1Low
Frank Martinez2Medium
Diane Tibbot3Medium
Christine Koch6Medium
Jeff Hay7Medium
Alan Steiner8High
Dan Jump14High
Karim Manar16High
Chris Norred22Very High
Table of Contents