Adoption Report

  • Updated

Adoption Report: questions and answers

Adoption Report, two main pages

v.3.0.0.0 ยท 19/11/2026

A reference for explaining the report and supporting its users. Each answer is short and states the rule the report follows.

Sections

1. About the report

What is this report for?๐Ÿ”—

It shows whether the application is being used and kept up to date. It answers: how many people create and change records over time, who is registered but not using it, which records and projects have gone quiet, and whether usage is growing or slipping.

What changed in version 3?๐Ÿ”—

Version 3 (v.3.0.0.0) shows two main pages instead of the earlier set: Adoption overview (are people using the application?) and How Power Framework is being used (what are they doing?). They are made for an account manager to run an adoption review with a client, and are described in section 2. The earlier pages are hidden, not deleted, and still work: Insight (v2), Adoption Overview (v2), Users, Activity, Data Upkeep, Feature Usage & Logins and Change Log. The Instruction page is now one screen. The counting rules did not change.

What changed in version 2?๐Ÿ”—

Version 2 (v.2.0.0.0) renames the report from "Audit Report" to "Adoption Report". It adds the Insight page as the first page and moves the Instruction page to the end. The Change Log page is the former Audit report.

Where does the data come from?๐Ÿ”—

From Dataverse only: the audit log (who changed what and when), the user table, and the current state of the tracked records (portfolios, programmes, projects, risks, issues, milestones, status reports and so on). Nothing comes from Power BI usage metrics or Microsoft Purview.

How current is it?๐Ÿ”—

The header shows the last refresh time, for example "Refreshed 14/10/2026 08:00 UTC". Figures are as of that refresh, not live. Under it the header shows the Dataverse environment and the report version, for example "orgname.crm.dynamics.com , v.3.0.0.0".

What time zone is used?๐Ÿ”—

UTC everywhere. An event made late in the evening local time may fall on the next UTC day.

How far back does the data go?๐Ÿ”—

The audit log starts when auditing was switched on, and older rows may have expired under the retention setting. For records with no audit history, the report uses the record's own created and last-modified times, once each. Earlier edits to those records are therefore understated.

Who is counted as a person?๐Ÿ”—

Human accounts only. Service accounts, application users, built-in Microsoft accounts and placeholder team members are left out. A person with a normal and an "(ADM)" admin account counts once.

Does the report monitor individual staff?๐Ÿ”—

It shows who made changes and when they last changed something or logged in (Users page, Change Log). It does not record reading, time spent or which screens were opened.

2. The two main pages

Version 3 shows two main pages, made for an account manager to run a review with a client on a shared screen. Page 1 asks: are people using the application? Page 2 asks: what are they doing in it? Both end with the same three lines: what is working well, what is worth discussing (written as a question) and a next step to offer.

How do I read Page 1, Adoption overview?๐Ÿ”—

From the top: a headline sentence (how many registered people used Power Framework, with the number for the period before), a bar of registered people (changing data, read-only, inactive), a trend label for people changing data with the movement behind it (New, Came back, Stopped), the last six months by month, and the three-line box. The footnote says what the login data can and cannot show.

Page 1, Adoption overview (demo data)

Page 1, Adoption overview (demo data)

How do I read Page 2, How Power Framework is being used?๐Ÿ”—

Three panels and the same three-line box as Page 1. "Where are people working?" shows the changes made in each module. "Are projects kept up to date?" counts the open projects that changed and the open projects with a status report. "What is due for a review?" shows, for each record type, the share of records that changed. The footnote says what is left out.

Page 2, How Power Framework is being used (demo data)

Page 2, How Power Framework is being used (demo data)

What do the five terms mean?๐Ÿ”—

Term Meaning
Registered A person with an enabled human account. Not the number of licences.
Active Used the application in any way in the period: changed data or logged in.
Changing data Created, updated, deleted, assigned or associated a record.
Read-only Logged in but changed no data.
Inactive Neither logged in nor changed data.

They add up: Registered = Active + Inactive, and Active = Changing data + Read-only.

Why is "Active" here different from the rest of this guide?๐Ÿ”—

On the two main pages Active means used the application in any way. In the older sections (Insight, Adoption Overview, Users) Active means changed data. What those pages call Active, the two main pages call Changing data, and what they call Engaged is Active here. The measures keep their names, so both sets of pages stay correct.

What do the 30 and 90 day buttons do?๐Ÿ”—

They choose the period: the last 30 or the last 90 days, ending today (UTC). Each is compared with the period of the same length just before it. Both pages follow the same button, and 30 days is the default. The six-month chart does not change. The hidden follow-up pages (Users and Data Upkeep) have their own date slicer.

Why do Read-only and Inactive say "not known"?๐Ÿ”—

Logins are recorded only from the date user access auditing was switched on. When that date falls inside the chosen period, people who only read cannot be told apart from people who are inactive. The page then counts only people who changed data, shows the rest as "not known" (a grey part of the bar), and the first thing to discuss is whether user access auditing is on. A 90 day period shows this for a while after the auditing started. The comparison with the period before is left out when that period is not covered either.

What do Rising, Steady, Falling and Early days mean?๐Ÿ”—

People changing data in the period against the period before. Up or down by 10% or more is Rising or Falling; otherwise Steady. "Early days" means the period before had fewer than 5 people changing data, so one person can move the trend. The page then gives the numbers and does not judge. Both limits are settings (the report owner can change this).

What are New, Came back and Stopped?๐Ÿ”—

  • New: their first change ever falls in the period.
  • Came back: changed in the period, did not in the period before, but had changed earlier.
  • Stopped: changed in the period before and not in this one.

Change in people changing data = New + Came back - Stopped, so the three numbers always explain the change. They count registered people. The older "Returning" also counts people active in both periods, so it is a different number.

What does the monthly chart show?๐Ÿ”—

People changing data, and from the first full month with login recording the read-only people, for each of the last six months, whatever period is chosen. The current month is marked "(to date)" because it is not complete.

How are the three lines at the bottom chosen?๐Ÿ”—

Each line is the first rule that is true.

  • Page 1, working well: more people changed data than before; at least half of registered people used the application; otherwise a plain fact.
  • Page 1, worth discussing: logins not recorded for the period; fewer people changed data (judged only when 5 or more did before, otherwise a note that the numbers are small); a third or more of registered people inactive (saying how many never used it and how many used it before); read-only people as many as or more than those changing data; otherwise "which area to grow".
  • Page 2 uses four checks, each a share: modules in use, projects changed, projects with a status report, and records changed. Working well is the highest share (a plain fact when all are under 25%). Worth discussing is the check that fell most against the period before (a fall of 10% or more, with 5 or more items changed before); if none fell, the lowest share; a share of 90% or more means nothing stands out. A check based on fewer than 5 items is not judged.

There is no pass mark, so clients do not all get the same answer.

Why is there no red and no overall status?๐Ÿ”—

The report is there to help clients adopt the application, not to grade them. When it was designed, all five clients seen showed Red on the old early-warning panel, which read as having failed; the Red came from the worst of four indicators and from small numbers. On the two main pages a gap is a question, small numbers are not judged, and the colours are purple, orange and grey. The early-warning panel is still on the hidden Adoption Overview (v2) page (section 10).

What is the "next step to offer"?๐Ÿ”—

A suggested action for the gap in the line above it, such as an onboarding session, a refresher or a status-report routine. The wording is in the model table Next Steps, one row per gap, and can be changed (the report owner can change this).

Why are no people named on the two pages?๐Ÿ”—

They are for a shared screen, where a list of inactive people can read as blame. The buttons "Who?" and "Records" open the hidden Users and Data Upkeep pages, for the client's admin after the meeting. Each has a "Back to the report" button.

What do the Users and Data Upkeep pages look like?๐Ÿ”—

Users shows who is registered, who is active and who has gone quiet. Data Upkeep shows which records and projects have not been changed for a while. Each has its own date slicer. The pictures use demo data, and the names on the Users page are covered here.

Users, opened by the Who? button (demo data, names covered)

Users, opened by the Who? button (demo data, names covered)

Data Upkeep, opened by the Records button (demo data)

Data Upkeep, opened by the Records button (demo data)

What does "Where are people working?" show?๐Ÿ”—

Records created and updated in each module in the period, stacked, one bar per module. It shows where people make changes, not time spent: the audit log has no time-in-application data. Modules with less than 2% of all changes are named in the line under the chart as barely used. A module with no changes is not drawn.

What does "Are projects kept up to date?" show?๐Ÿ”—

Two numbers for open projects: how many had a change in the period, and how many have a status report created in the period. A column chart shows the open projects with a status report in each of the last six months.

Why does a project count as changed when only a risk changed?๐Ÿ”—

A project record is often set up once and then left alone, while the work goes on in its risks, issues, milestones and status reports. A portfolio, programme or project counts as changed when it or anything under it changed (a programme includes its projects and their records). A risk, issue or other record is judged on its own changes only. Bulk edits and service accounts do not count.

Which projects are left out as closed?๐Ÿ”—

A project whose Project Closed Date is today or earlier. If it has no closed date, a project whose stage starts with "Clos", such as "4. Closure". Projects whose status reason is not Active are not loaded at all. Closed projects and their records are left out of Page 2, because no one is expected to keep them up to date. The closed state is today's, even for the period before.

What does "What is due for a review?" show?๐Ÿ”—

For each record type, the share of active records that changed in the period and the share that did not. Status reports are left out. A record that did not change is "due for a review", not wrong: it may be reviewed another way, or be ready to close. The line under the chart gives the totals. Records loaded in bulk and not changed since often explain a low share.

3. Pages and filters

What is on each page?๐Ÿ”—

Page Answers
Adoption overview Are people using the application? Headline, bar of registered people, trend, last six months, three-line box (section 2).
How Power Framework is being used What are people doing? Modules, projects kept up to date, records due for a review, three-line box (section 2).
Insight (v2) (hidden) One insight, four numbers (changed data, read only, total, logged in) and a weekly chart, each compared with the previous period.
Adoption Overview (v2) (hidden) Is use growing or slipping? KPI cards, weekly active people, early-warning panel.
Users (hidden; follow-up page) Who is active, reading, quiet or dormant.
Activity (hidden) What is created, updated and deleted, by week, type, module, weekday and person.
Data Upkeep (hidden; follow-up page) Which records and projects go stale.
Feature Usage & Logins (hidden) Which modules are used and who logs in.
Change Log (hidden) Field-by-field detail of each change.
Instruction One screen: the five words, how they add up, and the limits to know.

Hidden pages stay current with the model and can be opened in Desktop, and Users and Data Upkeep open from the buttons on the two main pages. A hidden "option sets" page supports the Change Log and is not for viewers.

How do the filters work?๐Ÿ”—

  • Period: the two main pages have two buttons, Last 30 days and Last 90 days, and no date slicer (section 2). The hidden pages have the date slicer described next.
  • Date: one slicer, linked across the hidden report pages. It starts at the last 90 days (the last 30 days on the Insight page). Choosing dates on one page applies to the others. The Change Log has its own filters.
  • Filter pane: Person Name, Item Type (record type), Entity Level (portfolio, programme or project), Project Name and Module. Nothing is selected by default, except Is Human Account, which is set to True on every page so the person lists show people only, not service accounts. Clear it to see service accounts.

Why did logins or some deletes disappear when I filtered by project or record type?๐Ÿ”—

Logins and some deletes belong to no record. A project or record-type filter works through the record, so those events drop out. This is expected.

What does "prior period" mean?๐Ÿ”—

The window of the same length just before the selected dates. For a week it is the previous week; for 90 days it is the 90 days before. The period end is capped at today. The Insight page compares in the same way.

4. The Insight page

What is the Insight page?๐Ÿ”—

The version 2 first page, hidden in version 3 as "Insight (v2)". It gives a quick read on whether the application is being used: one insight sentence, four numbers and a weekly chart for the selected dates. Each number is compared with the previous period of the same length.

What are the four numbers?๐Ÿ”—

Number What it counts
Changed data People who created, updated, deleted, assigned or associated at least one record in the selected dates.
Read only People who logged in but changed nothing.
Total write and logged in Changed data plus read only: everyone who used the application.
Logged in Everyone with a recorded login, including people who also changed data.

Each person counts once. Hover over a number to see a short explanation.

Why is Total write and logged in not the same as Logged in?๐Ÿ”—

Total is changed data plus read only. Logged in counts only people with a recorded login. Some people change data without a recorded login, for example before login auditing was switched on, so Logged in can be lower than the total. For instance, if 9 people changed data and 2 only read, the total is 11. If only 6 of them have a recorded login, Logged in shows 6.

How is the comparison made?๐Ÿ”—

The selected dates are compared with the previous period of the same length, just before them. For the last 30 days (16 Sep to 15 Oct) the previous period is 17 Aug to 15 Sep. Each card shows the previous figure, its dates and the difference, such as "โ–ผ 3 fewer than previous 30 days". Changed data and the total are green when up and red when down. Read only is never coloured, because more or fewer readers is not good or bad by itself.

Why does a card say "not recorded" or "No comparison yet"?๐Ÿ”—

Logins are recorded only from the date user access auditing was switched on. Before that, the previous period's Read only, Total and Logged in figures are "not recorded", so no comparison is made until a full previous period has login data. Changed data compares from the first data.

Why do Read only and Logged in show 0?๐Ÿ”—

When logins are not recorded in the selected dates, for example because user access auditing is not switched on, the two numbers show 0 and the note at the bottom of the page says logins are not recorded. Read that 0 as "none recorded", not "nobody logged in". The previous-period lines still say "not recorded".

How is the insight sentence chosen?๐Ÿ”—

It compares the people who changed data in the selected dates with the previous period: "More people", "Fewer people" or "The same number of people are changing data". The line below gives the numbers and how many registered people that is. If nobody changed data in either period, or there is not enough history, it says so.

What does the weekly chart show?๐Ÿ”—

People per week as stacked columns: changed data and read only. They are separate people, so the height is everyone who used the application that week. The first and last week can be partial, because only days inside the selected dates are counted. Read only appears only from the date login recording started.

Can I change the dates?๐Ÿ”—

Yes, with the date slicer at the top right. This page starts at the last 30 days; the other pages start at 90 days. The slicer is linked across pages, so choosing dates on one page applies to the others.

5. People

What are the kinds of people the report shows?๐Ÿ”—

Term Meaning
Registered A person with an enabled human account. All the rates below are measured against this group.
Active Created, updated, deleted, assigned or associated a record in the period. A login alone is not active.
Logged in Has a recorded login in the period.
Engaged Active or logged in. Shown as "Total write and logged in" on the Insight page.
Reader Logged in but changed nothing in the period.
New Their first-ever change falls in the period.
Returning Active in the period and had changed things before it.
Lapsed Active in the prior period, not active in this one.
Quiet Registered, no change and no login in the period, but seen within the last 30 days.
Dormant Registered, no change and no login for 30 days, or never.
Never seen Registered with no login ever recorded.

How are registered people counted?๐Ÿ”—

Distinct people (names without the "(ADM)" suffix) whose Dataverse account is enabled and human, and was created on or before the period end. The enabled flag is today's, so someone disabled later is not counted in earlier periods. Past registered numbers are therefore approximate. "Registered" means an enabled account exists, not that the person holds a licence or a particular role.

What is the difference between Lapsed and Dormant?๐Ÿ”—

  • Lapsed compares two periods and looks at changes only: active before, not active now. A lapsed person may still log in.
  • Dormant looks back from the period end and checks for any change or login over the last 30 days. It covers registered people only.

Someone who stopped editing but still logs in is Lapsed, not Dormant. They show as a Reader.

What are the status icons on the Users page?๐Ÿ”—

๐ŸŸข Active (changed something), ๐Ÿ”ต Reader (logged in only), ๐ŸŸก Quiet (no change or login in the period, but seen within 30 days), ๐Ÿ”ด Dormant (nothing for 30 days).

What is the activation rate?๐Ÿ”—

Registered people who changed something, divided by registered people. The engagement rate is the same with "logged in or changed something", so it is never lower.

Why are there far fewer active people than registered people?๐Ÿ”—

Active means changed a record. Many users only read, review or report. They show as Readers where logins are recorded, and otherwise look inactive. A low activation rate is a prompt to ask why, not a verdict by itself.

Why is a person missing from the Users list?๐Ÿ”—

The table lists people who are registered or have ever been active. A person who is not registered (for example a service account) and never changed anything does not appear.

6. Activity

What is a change event?๐Ÿ”—

One create, update, delete, assign or associate by a human account. Logins and the "cascade updates" Dataverse writes after a delete are not counted. One save of a record is one event, however many fields changed.

Why does the Change Log show many more rows than the Activity page shows events?๐Ÿ”—

The audit log stores one row per changed field. Creating one record can write dozens of rows. The Activity page counts events (one per save), the Change Log shows the field rows.

What is the difference between Records Created, Records Updated and Change Events?๐Ÿ”—

  • Records Created: records created by human accounts. Bulk loads by service accounts are left out.
  • Records Updated: different records updated. A record updated ten times counts once.
  • Change Events: every create, update, delete, assign and associate. A record updated ten times counts ten.

What is "events per person"?๐Ÿ”—

Change events divided by active people in the period.

What is a bulk edit?๐Ÿ”—

One account changing 10 or more different records of one table, with one event type, within 15 minutes (a mass update or import). Bulk edits count as activity, because they are real work and the person is active. They are ignored for stale records, the maintained rate and dormant projects. The size (10) is a model setting.

What is the rolling 28-day line?๐Ÿ”—

Active people over the 28 days that end on the last date of the selected period. It smooths weekly swings. The trend flag compares it with the 28 days before.

What is stickiness?๐Ÿ”—

People active in the last 7 days divided by people active in the last 28 days. It shows how many of this month's active people were also active this week.

Why does a period show 0 in one place and blank in another?๐Ÿ”—

A period with no activity shows 0. A period before the data starts, or in the future, stays blank. Login figures stay blank before user access auditing was switched on, so a gap reads as "not recorded".

7. Logins

How are logins counted?๐Ÿ”—

From user-access events in the audit log. The report counts login days (a person on a given day), not rows, because one visit can write several rows.

From when are logins recorded?๐Ÿ”—

Only from the date user access auditing was switched on. The Insight and Adoption Overview pages state the date. Earlier periods show no login figures. That is "not recorded", not "nobody logged in".

What does "last seen" mean?๐Ÿ”—

The time of the last recorded access. It is not an exact sign-in time: reloads and navigating away can write extra events.

Does a login mean the person looked at a record?๐Ÿ”—

No. A login is session access. Opening a record or viewing a report is not stored in Dataverse.

What is a reader?๐Ÿ”—

A person who logged in during the period but changed nothing. Logins are the only way the report can show people who use the application without editing.

8. Modules

What is a module?๐Ÿ”—

A type of record in the application, such as Risks, Issues, Milestones or Finance. Each module is one Dataverse table, except Finance and Project Model, which group a few tables each. There are 15 business-data modules. Configuration, system and access tables are not counted as modules in use.

What does "modules used" mean?๐Ÿ”—

The number of modules where at least one human account made a change in the period. Reading a module is not recorded, so a module people only look at does not count as used.

What is the breadth rate?๐Ÿ”—

Modules used in the last 28 days, divided by modules ever used. A module the client never adopted does not count against them. Only a module they used before and then stopped lowers the rate.

9. Data upkeep

What is an active record?๐Ÿ”—

A record in the Active state, excluding status reports. Status reports are periodic snapshots, not records that are maintained.

What is a stale record?๐Ÿ”—

An active record with no change by a person for 30 days, with bulk edits ignored. A record created or last changed only in a bulk edit more than 30 days ago counts as stale.

What is the maintained rate?๐Ÿ”—

The share of active records that are not stale.

What is a dormant project?๐Ÿ”—

A project with no change by a person, on the project or any of its records, for 30 days, with bulk edits ignored.

What is the status report cadence?๐Ÿ”—

Projects with a status report created in the last 30 days, divided by all projects. The project table has no state column, so closed projects are included.

Why are stale records so high when people are active?๐Ÿ”—

A common reason is that one person edited many records in a bulk update, which made them look recently maintained. Because bulk edits are ignored, the report shows upkeep by people working on the records. The stale-by-type chart shows which record types are affected.

Why does a record type show an empty bar?๐Ÿ”—

Some categories, such as (Blank) and Statuses, have no active records to count. An empty bar means zero.

10. Early-warning status

The early-warning panel is on the hidden Adoption Overview (v2) page. The two main pages do not show an overall status; they use the same four ideas as a gentler three-line box (section 2).

How is the overall status decided?๐Ÿ”—

It is the worst of four indicators: Red if any is Red, otherwise Amber if any is Amber, otherwise Green. Indicators with no data are ignored. It is deliberately not a weighted score.

What are the four indicators and their limits?๐Ÿ”—

Indicator What it measures Green Amber Red
Reach Registered people who changed something in the last 28 days 60% or more 30% or more below 30%
Trend Active people over 28 days against the 28 before fall under 10% fall of 10% or more fall of 20% or more
Maintenance Active records with a recent human change, bulk edits ignored 80% or more 60% or more below 60%
Breadth Modules used in 28 days out of all ever used 70% or more 40% or more below 40%

Are those limits agreed targets?๐Ÿ”—

No. They are proposed defaults, held in the model, and are to be confirmed with each client. Treat the status as a prompt for a conversation, not a verdict.

What does each Red mean in practice?๐Ÿ”—

Red on Likely reading Question to ask
Reach Few registered people changed anything Who is registered but not editing, and do they need to?
Trend Fewer people active than last month What changed: leavers, holidays, a project ending?
Maintenance Many records not touched recently Are the registers still the working tool?
Breadth Modules dropped out of use Which module stopped, and why?

Why is the status "No data"?๐Ÿ”—

The period does not have enough history, for example no activity in the 28 days before, so the indicator cannot be judged.

11. Change Log

What is the Change Log?๐Ÿ”—

The former Audit page: one row per changed field, with the old and new value, the person, the record and the time. Use it to find what changed on a record. Do not use it to count activity (see section 6). It follows the Is Human Account filter, so changes made by service accounts are hidden unless you clear that filter in the filter pane.

Why do some values show as a number?๐Ÿ”—

Pick-list values are translated to labels through the option-set tables. A number remains for the two Microsoft system tables (app action, flow machine group) and for any new pick-list column that has not been mapped yet. The report owner adds the mapping.

12. Why does this look wrong?

Active people is lower than expected.๐Ÿ”—

Check the date range and the filter pane first. Then remember that only changes count: readers do not. Last, check that the person's account is classed as human (service-style names with "#", "system" or "service account" are not).

Registered people is lower or higher than the user list in Dataverse.๐Ÿ”—

Disabled users, application users, built-in accounts and placeholders are excluded, and an "(ADM)" account joins its owner. The count uses today's enabled flag.

A project is "dormant" but people say they work on it.๐Ÿ”—

Their changes may be bulk edits (ignored), or reads and logins, which are not changes. Look at the person's activity on the Activity page.

A project is not "dormant" but nobody seems to work on it.๐Ÿ”—

Any change by a person in the last 30 days, on the project or any of its records, resets the clock. That includes a small edit to a single risk. Only bulk edits are ignored. Check the last upkeep change date in the Data Upkeep project table.

The numbers changed by one or two since yesterday.๐Ÿ”—

Stale records and dormant projects are measured against today's date, so each day some records cross the 30-day line. New data arriving at the refresh moves totals too.

Logins are blank for an earlier period.๐Ÿ”—

User access auditing was not switched on then. See section 7.

Total write and logged in is higher than Logged in.๐Ÿ”—

This is expected when some people changed data without a recorded login. See section 4.

A sudden jump in change events.๐Ÿ”—

A mass update or import by one person can add hundreds of events in minutes. Check changes by person and by week on the Activity page. A jump like that counts as activity but not as upkeep.

Past registered numbers do not match what we remember.๐Ÿ”—

Registered uses today's enabled flag and the account creation date, so people who left and were disabled are not counted in earlier periods.

13. What the report cannot tell you

  • Record views, report views and time spent. Dataverse does not store reads in the audit table, and nothing from Power BI usage or Microsoft Purview is used.
  • Exact sign-in times. Logins are session-level audit events.
  • History before auditing began, except the created and last-modified times of the records themselves.
  • Licences, contract position and agreed interventions. They are not in any loaded source. A contract page was put aside, and interventions are a future idea.
  • Why usage is low. The report shows the pattern, the conversation explains it.

Was this article helpful?

1 out of 1 found this helpful