People represented

Observed M365 enabled population

M365 credits per person per week

Observed GitHub population

Consumption baselines by function

Baseline table

Population-level highlights

Product mix is the new baseline. The report now separates observed M365 Copilot and GitHub Copilot activity before relating either product to working patterns. M365 credits, GitHub AI credits and GitHub activity remain separate measures; no common unit or conversion has been confirmed.

Consumption is concentrated. Across the observed product window, the highest-volume 10% of people with positive Microsoft 365 credits account for 38.0% of all observed M365 credits. Concentration is reported within each product separately: M365 credits and GitHub AI credits are different units and are never pooled into a shared denominator. See the consumption detail page for the concentration cuts and the heavy-user pattern split.

Enablement is a people measure. M365 enablement is reported as the share of people with any observed enabled day, not as a percentage of possible enabled days.

Ways of Working comparisons use the joined product window. The Person query has 26 weeks of collaboration history, while product feeds cover 13 weeks of weekday daily records. Relationships in this report use the product weeks after aggregating all product files to a single Person × week panel.

How to read cost intensity

M365 credits per session is retained as an appendix-style intensity measure. It is not a productivity or value measure, and it does not establish that a group is more or less efficient.

Observed product mix

Product-mix baseline

M365 credit volume within observed product mix

Join strategy

The dashboard first aggregates every daily or dimensional product export to PersonId × MetricDate, where MetricDate is the Sunday week start. The joined panel then has one row per Person Query row, so downstream charts cannot accidentally multiply people by service, feature, language or model dimensions.

Collaboration volume by observed product mix

Collaboration mode mix

Network size by observed product mix

Reading the network measures

Network metrics describe collaboration breadth in the observed calendar and communication graph. They do not measure relationship quality, influence, delivery quality or business impact. The useful question is whether product-use segments have visibly different collaboration reach that deserves deeper analysis.

M365 credit volume and network

After-hours and weekend collaboration by product mix

Interpretation guardrail

After-hours collaboration counts meetings, email, Teams chats, calls and channels with at least one other person outside configured working hours, with overlapping activity deduplicated. It is not total work time. Flexible schedules, on-call duties and time zones may matter, but none is recorded here.

M365 cost intensity and workload

Profile partition withheld: one or more cells or complements are below the publication floor.

Credit concentration

Heavy-user consumption pattern

Reading the intensity measures

Volume and consistency answer different questions. Concentration shows how much of the observed credit total sits with the highest-volume people. The pattern split shows whether those people consume at that level habitually or in occasional bursts. A group can be concentrated without being habitual.

The window is 13 product weeks. Consistency is measured inside that window only. It is a lookback description, not a trend, and it cannot show whether a pattern is starting or fading. Unobserved weeks are excluded from both the threshold and the per-person counts, so a coverage gap is never counted as a quiet week.

Thresholds are parameters, not findings. Heavy user: top 20% of people with positive observed M365 credits, by average weekly credits. High-credit week: a person-week at or above the 90th percentile of all active person-weeks. Sustained: at least 50% of a person’s observed weeks are high-credit weeks. Changing any threshold changes the groups; narrower heavy-user cuts are permitted and will be withheld whenever a resulting group falls below the publication floor.

Service mix

M365 credit-volume groups

Cost-intensity profile availability

Profile partition withheld: one or more cost-intensity cells or complements are below the publication floor.

Profile definitions

The label high cost per session describes a measurable consumption pattern. It is not a wellbeing, productivity or value judgement.

Profile baseline

Profile baseline withheld: the complete ConsumptionProfile by CreditQuartile partition does not meet the publication floor.

Join and aggregation workflow

Core rules

Key crosswalk check

Simulation design

This dashboard is a demonstration of analytical possibilities. Every person, credit, session and Ways of Working value is synthetic.

The current source set represents 480 synthetic people. The person query spans 26 Sunday-start weeks. The M365 and GitHub credit feeds cover the final 13 weeks at weekday daily grain, which is why this report filters Ways of Working measures to the product window before relating them to credits.

Available-to-focus hours are simulated on a 40-hour weekly basis here, but that basis is configurable per tenant through metric rules. The metric describes calendar time left after meetings and scheduled Teams calls, and should not be read as coding time or psychological flow.

The following effects are deliberately modest:

• Higher consumption is associated with more chat and less email as a share of collaboration modes.
• Total collaboration volume is raised mainly by organisational composition.
• Internal network size has a small residual association after composition is considered.
• Cost per M365 session has a weak association with after-hours collaboration in some credit bands.
• External network size is a null.

Synthetic source assets

The uncompressed source-shaped CSVs sit under:

examples/utility-r/_data/
  person-query/PersonQuery.csv
  consumption-query/PeopleMetaData.csv
  consumption-query/PersonM365CreditsMetrics.csv
  consumption-query/PersonGitHubCreditsMetrics.csv

They are intended for demonstrations, workshops and adaptation of the template. They contain no customer data or real identifiers. The GitHub-query files in the same folder are not required for this consumption report.

Definitions and limitations

Credits per session means Microsoft 365 Copilot credits consumed divided by observed Microsoft 365 Consumption-query sessions. It is preferred over any token-normalised measure because token fields are not present in this export.

Available-to-focus hours means working hours remaining after excluding meetings and scheduled Teams calls. Uninterrupted_hours and Interrupted_hours partition available-to-focus time in this source. The working-hours basis is tenant-configurable; the 40-hour identity checked here is specific to this synthetic dataset.

Credit concentration uses only coarse positive-user M365 quartiles, shared with the credit-group panels. No individual curve, top-1% point, fine histogram, observed extrema or individual ranking is published. If any quartile or the no-observed-M365 group falls below the floor, the entire partition is withheld.

Credit activity segments are based on active product weeks and observed M365 credit volume. They are not the standard vivainsights action-based segments, because this source set does not include a Copilot actions metric.

PeopleHistoricalId is the documented join key between the Consumption activity files and PeopleMetaData. It must be taken from the Consumption export rather than derived from PersonId. A person-query export alone cannot be joined to PeopleMetaData; the bridge is an activity file that carries both PersonId and PeopleHistoricalId.

The dashboard shows associations. It does not establish that AI consumption caused a working pattern, and it does not observe output quality, task difficulty or business value. It also makes no wellbeing claims: uninterrupted, interrupted, available-to-focus, after-hours and weekend measures describe observed calendar and collaboration patterns only.

Replacing the synthetic data

1. Replace the four CSVs listed above with schema-aligned governed exports.
2. Validate the identity crosswalk and disclose the match rate.
3. Require independent, week-relevant eligibility and ingestion/coverage evidence before zero filling. Missing rows alone do not establish zero activity or no provisioning; M365 enabled days do not establish GitHub provisioning. Missing person-weeks remain unknown here. Recorded sums per analysis week are lower bounds when completeness is unknown, not complete-period consumption.
4. Preserve the daily service grain until after service-level totals are needed; aggregate services before computing person-week rates.
5. Keep person-level averaging so each person carries equal analytical weight.
6. Retain the minimum of 10 and whole-partition disclosure gates, including contributor complements and shared gates for related tables/charts. Blank charts, empty tables and NA values mean withheld (or undefined rates), not zero. Overlapping releases require a governed cross-report disclosure review; these within-report checks are not a certification of arbitrary real data.
7. Keep M365 Copilot credits and GitHub AI credits in separate panels and denominators unless an authoritative common-unit/conversion contract is established.