Black Friday sale: an extra 40% off every course until Monday, November 30.New Year sale: an extra 30% off every course until Friday, January 1.Memorial Day sale: an extra 25% off every course until Monday, May 31.SysAdmin Day sale: an extra 25% off every course until Sunday, August 1.Back to School sale: an extra 25% off every course until Sunday, August 22.Labor Day sale: an extra 25% off every course until Monday, September 6.Black Friday sale: an extra 40% off every course until Monday, November 29.New Year sale: an extra 30% off every course until Saturday, January 1.Choose your certification →

Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.

6.2.2. Endpoint Analytics: Scores, Startup, and Reliability

💡 First Principle: "The laptops feel slow" isn't actionable; "boot score 41 against an all-organizations median of 55, driven by three hard-disk models and one startup process" is. Endpoint analytics turns user experience into scores from 0 to 100 and compares them with a baseline, so you can see where you stand, which fix is worth the most points, and whether a change made things better or worse.

Getting data flowing. Endpoint analytics supports Windows Pro, Pro Education, Enterprise and Education on devices that are Intune-managed, co-managed or Configuration Manager–managed through tenant attach, and Entra joined or hybrid joined. The Connected User Experiences and Telemetry (DiagTrack) service must be running, and Intune-managed devices must reach https://*.events.data.microsoft.com. To start, open Reports > Endpoint analytics and choose Collect device data from All cloud-managed devices, Selected devices, or I'll choose later. Intune assigns a configuration profile called Intune data collection policy. Data can take up to 24 hours after a restart to appear, startup performance needs at least one restart, and a score appears only once at least five devices report.

Scores and baselines. Every score runs from 0 to 100, and lower means more room to improve. The overall Endpoint analytics score is a weighted average of three category scores: Startup performance, Application reliability and Work from anywhere. The built-in baseline is All organizations (median), shown as triangle markers. You can save up to 20 baselines of your own, for example before a Windows feature update. Metrics that regress beyond the default 10% regression threshold turn red as "regressed". Insights and recommendations lists fixes by how many points each would add, and Device scores and Model scores tabs help you target individual machines and plan hardware refresh. Per-device scores also appear on each device's User experience page.

ReportWhat it measuresKey details
Startup performanceBoot score (power-on to sign-in) and sign-in score (credentials to responsive desktop)29-day event retention; insights flag hard disk drives, Group Policy delays and slow boot or sign-in
Startup performance: Startup processesProcesses keeping CPU above 50% after the desktop rendersListed only if they affect at least 10 devices
Startup performance: Restart frequencyRestarts over 30 days, split into normal and abnormalAbnormal: Stop errors, long power-button press, unknown
Application reliabilityPer-app crashes, as mean time to failure, plus usage14-day rolling window; replaced by Device timeline when Advanced Analytics is licensed
Work from anywhereCloud readiness: Windows version, cloud management, cloud identity, cloud provisioningWindows 11 hardware readiness is shown but doesn't affect the score
Advanced Analytics reportsResource performance, battery health, anomalies, device timelineExtra licensing; taught in Phase 3

Reading startup performance. The startup score combines a boot score and a sign-in score. The sign-in score ignores first sign-ins and sign-ins right after a feature update, which are slow for reasons you can't fix. Hard disk drives typically boot three to four times slower than SSDs, which is why model scores drive hardware decisions.

Reading restart frequency. The ideal is about one Update restart per device per month. Fewer suggests devices aren't being patched, more means too many update restarts, and Restart (no update) should be near zero. Rising Stop errors or long power-button presses signal a stability problem.

Reading application reliability. An app's score depends mainly on mean time to failure: total usage divided by total crashes over a 14-day window. An app that ran 1,400 hours across the estate and crashed 70 times has an MTTF of 1,400 ÷ 70 = 20 hours. Apps are included when they're foreground apps with measurable usage, active on more than 5 devices or 2% of enrolled devices, whichever is larger. At most 10 crashes per app, per device, per day count, so one badly broken machine can't swamp the score.

Scope and permissions. To view reports you need a role such as Read Only Operator, Help Desk Operator or Endpoint Security Manager, or the Entra Reports Reader role. Device-level tabs honor scope tags, but aggregated scores use all enrolled devices. If a regional admin needs scores for only their region, that takes Advanced Analytics device scopes.

⚠️ Exam Trap: When a scenario says Endpoint analytics shows no startup data for newly onboarded devices, don't pick "assign a Remediation" or "wait for the next compliance check". Startup performance needs at least one restart, data can take up to 24 hours after that restart, and no score appears until five devices report. The other usual culprits are the DiagTrack service not running or blocked access to the events endpoint.

Reflection Question: After a Group Policy change, your startup score drops from 62 to 51 while the all-organizations median is unchanged. Which tab would you open first, and how would a baseline you saved last month help you prove the cause?

See how it connects
Alvin Varughese
Written byAlvin Varughese
Founder•26 professional certifications earned