A Teams rollout is considered successful when users have switched from email and older collaboration tools to Teams for their daily communication and collaboration. But how do you know when that switch has happened — and how do you identify where it hasn't?
Adoption metrics tell you whether users are engaging with Teams in the ways you intended, which user populations are lagging, and whether the features you've enabled are actually being used. Without this data, adoption efforts are guesswork.
Where to Find Teams Usage Data
Microsoft 365 provides Teams usage data through several channels:
- Microsoft 365 Admin Center > Reports > Usage: High-level summaries of Teams activity — active users, meetings, calls, messages. Available for 7, 30, 90, and 180-day windows. Good for executive reporting and trend tracking.
- Teams Admin Center > Analytics & reports: More detailed Teams-specific reports including per-user activity, device usage, live events, and PSTN usage. Better for operational analysis.
- Microsoft Graph API / Graph PowerShell: Programmatic access to the same data, plus additional metrics not available in the UI. Useful for building custom dashboards or scheduled reports.
- Microsoft Viva Insights (formerly Workplace Analytics): Advanced collaboration analytics that shows patterns of communication, meeting load, focus time, and more. Requires additional licensing.
- Teams created and abandoned: A rising count of teams with no second-week activity means people are trying the tool once. That is an onboarding signal, not an adoption success.
- External versus internal activity: Heavy guest traffic with thin internal channel traffic means Microsoft Teams is being used as an extranet while staff still coordinate elsewhere.
Core Metrics to Track
Monthly/weekly active users (MAU/WAU). The most basic adoption metric: how many users have logged into Teams and taken at least one action (sent a message, joined a meeting, made a call) in the period. Track this against your total licensed user count to calculate the adoption percentage. A target of 80%+ monthly active users is reasonable for a mature deployment; in the first six months after rollout, watch the trend, not the absolute number.
Active users by activity type. Break down active users by how they're using Teams: messaging only (low adoption), meetings only, messaging + meetings (good), messaging + meetings + channels + collaboration features (high adoption). The progression through feature usage indicates how deeply Teams has been embedded in daily workflows.
Channel messages vs chat messages. The ratio of channel messages to private chat messages indicates whether users are collaborating in the open (channels) or reverting to private conversations. A healthy Teams environment has significant channel activity — it indicates that Teams has replaced email for team communication, not just become a chat application. If all activity is private chat, users may be missing the collaboration model that makes Teams valuable.
Meeting minutes per user. The total minutes of Teams meetings per licensed user per week. This is a proxy for how much synchronous collaboration is happening through Teams vs. other tools. Track whether this number grows after rollout and compare across departments to identify where Teams is being used vs. where people are still using phone calls or in-person meetings without Teams.
Teams and channels created per month. The rate of new team and channel creation. A healthy trend is a reasonably steady rate — enough to indicate active new use cases, but not so many that sprawl governance is losing the battle. If team creation spikes and then stops, it may indicate an early-adopter wave that hasn't reached the broader organisation.
Guest user activity. How many active guest users are in your tenant, and what are they doing? This is an adoption metric for external collaboration: if guest access is enabled but guest activity is near zero, either external collaboration workflows aren't using Teams, or users are routing external collaboration through other channels.
Tracking Adoption by Department or Role
Aggregate adoption metrics hide variation across the organisation. The finance team may be at 95% adoption while the operations team is at 40%. Knowing which populations are lagging is essential for targeted adoption support.
The Microsoft 365 Admin Center usage reports can be filtered by licence or by user group in some cases, but detailed per-department breakdowns typically require either the Graph API (query usage data by department attribute from Entra ID) or Viva Insights.
For organisations without Viva Insights, a practical alternative is to export per-user activity data from the Teams Admin Center and then join it to your HR system's department data in a spreadsheet or BI tool. It's manual, but it gives you the department-level breakdown you need.
Setting Adoption Targets
Adoption targets should be set before rollout and reviewed regularly. A framework for target-setting:
- Awareness target: X% of users have logged in at least once within 30 days of licence assignment.
- Basic adoption target: X% of users send at least one message or attend at least one meeting per month.
- Deep adoption target: X% of users use Teams for channels, file sharing, and meetings (not just chat).
Targets should increase over time. A new deployment might aim for 60% basic adoption in month 3 and 80% in month 6. What constitutes "success" depends on the specific use cases Teams is being deployed to replace — replacing email for team communication requires different adoption levels than deploying Teams Phone to replace desk phones.
A Headline Adoption Number That Hid a Quiet Department
Monthly active users for Microsoft Teams is a tenant total. It will look healthy while an entire function ignores the product, as long as other functions are busy. The figure is worth keeping. It is not worth presenting alone.
A 1,600-person food distributor reported 82 percent monthly active users and called the rollout complete. Splitting the same activity export by department showed warehouse supervisors at 30 percent and the commercial team above 95 percent. The supervisors still ran shift handovers in a group text, because the shared devices on the floor signed in with a generic account that the usage report attributed to one user. The generic account made the floor look almost unused and made one "user" look extremely active. Both readings were artefacts of the identity design.
Fix the identity first if the metric is going to be used for a decision. Shared devices need a story: either a licensed user per person, or an explicit note that those devices are out of the adoption denominator. Then set the target on the population that was supposed to change behaviour. A goal of "80 percent of warehouse supervisors send a channel message in a week" can fail and be useful. A goal of "80 percent of the tenant is active" can succeed and describe a different company from the one that paid for the project.
Feature mix matters as much as the headcount. A population that only joins meetings has moved the call, not the collaboration. Track channel messages and meeting attendance as separate series for the pilot group. If meetings rise and channel posts stay flat for six weeks, the training emphasised the wrong task, or the channel design does not match the work. Adding more licences will not change that shape.
Publish the split numbers to the managers of the quiet group, with the comparison, and with one concrete action. Measurement that stays inside IT becomes a dashboard. Measurement that lands on the manager who owns the shift pattern becomes a conversation about why the handover is still on a phone.
Exclude disabled accounts and unlicensed shared mailboxes from the denominator before the percentage is shown to anyone. A 7-day window and a 30-day window answer different questions. Label the window on the chart. Do not compare this month to the rollout month without noting holiday weeks. A dip in late December is not a reversal. App usage sitting at zero for an app you deployed is a deployment fact. Include it beside the active-user chart. When a department is quiet, check whether they have a client installed and a licence before concluding they refused the tool. Keep the raw export. A percentage without the export cannot be rebuilt when the definition of active changes. Shared-device accounts should be listed next to the chart so the denominator is explainable. A department at the tenant average can still be failing its own target. Show the target on the same figure. Channel message counts without a unique-author count can be one enthusiastic person. Look at both. Reorganisations break department joins. Rebuild the join in the same week as the reorganisation, not at the next quarterly review. Adoption targets for meetings and for channels should be separate sentences. One blended target hides a stall. If a role has no desktop, measure the mobile client. A desktop-only metric will mark them as absent. Keep the export even when the chart looks fine. The quiet week is easier to explain with the file than without it. Present the quiet group to its own manager with one proposed change, not with a tenant-wide lecture. If the numerator and the denominator come from different days, say so. A licence count from Monday and activity from the previous month will not describe the same population.
