Segmentation & Churn Analysis — Copilot Adoption

← Back to Prompt Library

Purpose

Classify Copilot users into behavioral segments based on usage intensity, track how users move between segments over time, and quantify churn — the rate at which users disengage from Copilot after initial adoption.

Audience

People analytics leads, Copilot program managers, digital adoption teams

When to use

After at least 8–12 weeks of person query data have been collected, when you need to understand not just whether people are using Copilot, but how their usage patterns evolve over time.

Required inputs

  • Person query CSV with Copilot metrics and HR attributes
  • At least 8 weeks of data (12+ weeks recommended for meaningful churn analysis)
  • HR attributes for segmentation breakdowns

Assumptions

  • Data is at person-week granularity
  • PersonId is a consistent anonymized identifier
  • MetricDate is a date field representing the start of each week
  • Rows with missing Copilot metric values likely represent unlicensed users
  • The vivainsights R or Python package is available in the environment

An HTML report or Jupyter/R notebook containing segment distribution charts, transition matrices, churn curves, and at-risk group analysis.

Prompt

You are a behavioral analytics specialist. Your task is to perform a user segmentation and churn analysis on Microsoft Copilot usage data from a Viva Insights person query export. The goal is to classify users into usage-based segments, track transitions between segments, and quantify churn.

LANGUAGE CHOICE
Choose R or Python based on what is already installed in your environment to minimize setup.

DATA LOADING AND PREPARATION
1. Load the person query CSV using `import_query()` from the `vivainsights` library (R or Python). This handles variable name cleaning and type parsing automatically.
2. Identify Copilot metric columns by checking for columns containing the word "Copilot" in their name. Reference the taxonomy at https://github.com/microsoft/viva-insights-sample-code/blob/main/examples/example-data/copilot-metrics-taxonomy.csv to classify and validate the detected metrics. Use `Total_Copilot_actions_taken` as the primary activity metric. Print detected columns and the date range.
3. Run `extract_hr(df)` from the `vivainsights` library to identify available HR / organizational attribute columns. Use the returned list for all segmentation breakdowns instead of hard-coding column names.
4. Classify each person-week:
   - "Licensed": at least one non-null, non-zero Copilot metric value.
   - "Unlicensed": all Copilot metrics are null or zero.
5. Filter to only licensed person-weeks for the segmentation analysis.
6. Fill any remaining NaN values in Copilot metric columns with 0 for licensed users (a licensed user with a null action count in one metric likely had zero usage of that specific feature).

USER SEGMENTATION
6. Use `identify_usage_segments()` from the `vivainsights` library to classify users into segments. This function uses `Total_Copilot_actions_taken` and classifies users based on both usage volume and consistency (habit formation over 12 weeks), producing segments such as Power Users and Habitual Users. It is preferred over manual percentile-based segmentation.

7. Also compute a "stable segment" for each user based on their most frequent weekly segment over the entire period (mode). This gives a single label per person.

SEGMENT DISTRIBUTION OVER TIME
8. For each week, count the number of users in each segment. Create:
   a. A stacked area chart showing the four segments over time (absolute counts).
   b. A stacked area chart showing the four segments as percentages of total licensed users.
   c. A summary table of the latest week's segment distribution.

SEGMENT BREAKDOWN BY HR ATTRIBUTES
9. For each HR attribute (Organization, FunctionType, LevelDesignation), compute the segment distribution using the "stable segment" label. Create:
   a. A stacked bar chart showing segment proportions by each HR attribute value.
   b. Only include HR attribute values with at least 5 licensed users.
10. Identify groups with the highest proportion of Power Users and groups with the highest proportion of Inactive users. Flag these as "leading" and "at-risk" groups respectively.

TRANSITION ANALYSIS
11. Build a week-over-week transition matrix. For each consecutive pair of weeks, count how many users moved from each segment to every other segment. Aggregate across all week transitions.
12. Normalize the matrix to show transition probabilities: given a user is in segment X this week, what is the probability they are in segment Y next week?
13. Visualize the transition matrix as a heatmap with annotations showing the probabilities.
14. Highlight key transitions of interest:
    - "Activation": Inactive → any active segment
    - "Churning": any active segment → Inactive
    - "Deepening": Light User → Regular User or Regular User → Power User
    - "Declining": Power User → Regular User or Regular User → Light User

CHURN ANALYSIS
15. Define "churn" as: a user who was active (in any active segment) for at least 2 consecutive weeks, then became Inactive for 2 or more consecutive weeks. The churn date is the first week of inactivity.
16. For each week, calculate:
    a. The number of users who churned that week.
    b. The churn rate: churned users / active users in the prior week.
17. Create a line chart of weekly churn rate over time.
18. Build a "time to churn" distribution: for users who churned, how many weeks were they active before churning? Show as a histogram.
19. Compute overall churn statistics:
    a. Total users who ever churned
    b. Percentage of ever-active users who churned
    c. Median active weeks before churn
    d. Re-activation rate: of those who churned, what percentage became active again later?

AT-RISK USER IDENTIFICATION
20. Identify currently at-risk users: licensed users who were active 4+ weeks ago but have been Inactive for the last 2 consecutive weeks (not yet meeting the full churn definition).
21. Summarize at-risk users by HR attribute: which organizations, functions, or levels have the highest concentration of at-risk users?
22. Create a table of at-risk user counts by Organization and FunctionType.

REPORT GENERATION
23. Compile all outputs into an intermediary document first, then export to HTML:
    - R: Create an RMarkdown file (.Rmd), then knit to a self-contained HTML file.
    - Python: Create a Jupyter notebook (.ipynb), then export to a self-contained HTML file. Keep the intermediary file alongside the HTML output for troubleshooting. The report should contain these sections:
    a. "Segment Definitions" — explain the four segments and thresholds used.
    b. "Segment Distribution" — charts from step 8 and summary table.
    c. "Segment Breakdown by Group" — charts from step 9, with leading/at-risk group callouts.
    d. "Transition Analysis" — heatmap from step 13, key transition highlights from step 14.
    e. "Churn Analysis" — churn rate trend, time-to-churn histogram, summary statistics.
    f. "At-Risk Users" — summary table from step 22.
    g. "Key Takeaways" — 3-5 bullet points synthesizing the most important findings.
    h. "Methodology" — brief description of data source, segment definitions, churn definition.

24. Use static charts (matplotlib/seaborn for Python or ggplot2 for R).
25. Save the report and intermediary file with descriptive filenames like "copilot_segmentation_churn_YYYYMMDD.html".

IMPORTANT NOTES
- Segments are assigned by `identify_usage_segments()` based on both usage volume and consistency (habit formation over 12 weeks). Print the segment definitions used so the reader can interpret them.
- Do NOT count unlicensed users in any denominator. They are excluded from the analysis entirely.
- Suppress any HR attribute segment with fewer than 5 users in the visualizations.
- The transition matrix should only include users who appear in both consecutive weeks.
- If the dataset has fewer than 8 weeks, note that churn analysis may not be reliable and adjust the churn definition (e.g., reduce the "2 consecutive inactive weeks" requirement).

Adaptation notes

  • The identify_usage_segments() function provides sensible default thresholds. If the segments do not produce meaningful groupings for your data, consult the function documentation for available parameters to customize segmentation criteria.
  • If your organization has a specific definition of “churn” (e.g., 4 weeks of inactivity instead of 2), modify the churn definition in step 15.
  • For organizations with multiple Copilot products, consider segmenting by product (e.g., Copilot in Teams vs. Copilot in Word) by using product-specific metrics.
  • Add a “New User” segment by tracking each user’s first active week and analyzing the onboarding trajectory separately.
  • If you want to predict future churn, add a note requesting a logistic regression or survival analysis model.

Common failure modes

  • Agent uses absolute thresholds instead of percentiles. The prompt specifies percentile-based thresholds, but an agent may default to arbitrary cutoffs. Verify that it calculates and prints the actual thresholds.
  • Agent counts unlicensed users as “Inactive.” The Inactive segment should only include licensed users with zero activity — not unlicensed users. Verify the filtering step.
  • Transition matrix includes users who appear in only one week. Ensure the matrix only counts users present in both the “from” and “to” weeks.
  • Agent conflates week-level segments with user-level segments. The report uses both: per-week segments for trends and a “stable segment” (mode) for HR attribute breakdowns. Ensure both are present.
  • Churn definition is too aggressive. Two weeks of inactivity may include holidays or vacation. Consider extending to 3–4 weeks for more conservative churn estimates.
← Previous: Executive Summary Next: ROI Estimation →