Module 1 — Lesson 4 of 8

Health Scores

Understand how Performance Health Scores are calculated, what each score range means, and how to use scores for decision-making.

📚 Overview

🔧 Deep Dive

🛠 Hands-On

✅ Check

🏆

Score Dimensions

🔢

0-10

Score Range

⚖

25%

Boot Time Weight

📅

14+

Days to Baseline

Health Score Composition -- Weighted Components

A Performance Health Score is a composite number (0-10) assigned to each endpoint, combining CPU, memory, disk, boot time, and app stability into one rating. Think of it like a credit score for a computer -- higher is healthier.

Health scores enable prioritization (find the worst machines fast), trend tracking (is the fleet improving?), stakeholder communication ("your department averages 6.2, down from 7.8"), and data-driven hardware refresh decisions.

Score Ranges at a Glance

✅

8.0-10.0

Excellent -- No action needed

👍

6.0-7.9

Good -- Monitor for trends

⚠

4.0-5.9

Fair -- Investigate & plan

🚨

0.0-3.9

Poor -- Immediate action

Simulated Health Score Dashboard

Tanium Console -- Health Score Overview

8.0

CPU Health

7.5

Memory Health

9.0

Disk Health

6.0

Boot Time

9.5

App Stability

Composite Score: 7.8 Good

(8.0 x 0.20) + (7.5 x 0.20) + (9.0 x 0.20) + (6.0 x 0.25) + (9.5 x 0.15) = 7.825

How Composite Scores Are Calculated

Each endpoint receives a sub-score (0-10) in five dimensions. The composite score is a weighted sum:

            Health Score = (CPU x 0.20) + (Memory x 0.20) + (Disk x 0.20) + (Boot x 0.25) + (Stability x 0.15)
          

Dimension	Weight	What Drives It	High Score Means
CPU Health	20%	Avg utilization, peak frequency, sustained high use	Rarely exceeds 50% CPU
Memory Health	20%	Memory pressure %, page file use, near-exhaustion events	Ample free RAM available
Disk Health	20%	I/O queue length, throughput, free space	SSD with ample free space
Boot/Login Time	25%	Total boot duration vs. baseline	Boots under 2 minutes
App Stability	15%	Crashes and hangs per day	Zero or rare crash events

Weights Are Configurable

The percentages above are defaults. Mercury Insurance can adjust weights based on priorities -- e.g., increase boot time weight to 30% if that is the biggest employee complaint. Changing weights recalculates all historical scores retroactively.

Score Ranges in Detail

● Excellent (8.0 - 10.0) -- No Action Needed ▼

The endpoint is performing optimally. CPU and memory have ample headroom, disk I/O is fast, boot times are short, and applications are stable. These machines provide a great employee experience.

● Good (6.0 - 7.9) -- Monitor for Trends ▼

Performing acceptably with minor areas for improvement. One or two dimensions may be trending toward "Fair." No immediate action required, but watch for declining trends.

● Fair (4.0 - 5.9) -- Investigate and Plan ▼

Noticeable performance issues. The employee likely experiences occasional slowness, longer boot times, or intermittent hangs. Investigate root cause and plan remediation -- software fixes or hardware upgrades.

● Poor (0.0 - 3.9) -- Immediate Action ▼

Significant performance problems impacting productivity. Persistent slowness, frequent crashes, or extremely long boot times. These machines need urgent remediation or replacement.

Baselining Your Environment

Day-one scores are not yet meaningful for decisions. Health scores gain power from context and comparison -- you need baseline data first.

Week 1: Daily Patterns

Capture how endpoints behave at 9 AM (heavy login) vs. 2 PM (steady work) vs. 11 PM (idle).

Week 2: Weekly Patterns

Identify recurring weekly peaks (e.g., Claims processing heaviest on Mondays).

Collect Boot Events

Endpoints aren't rebooted daily. Need 2+ weeks for 2-4 reboot data points per machine.

Scores Stabilize

After 14+ days, transient outliers are filtered and scores become reliable for trending and decisions.

Pro Tip: Communicate to Stakeholders

During the first two weeks, treat data as informational only. Do not make hardware refresh decisions or escalate to leadership until you have 14+ days of baseline. Communicate this upfront so early fluctuations don't erode confidence in the tool.

Comparative Analysis

Tanium Console -- Score Comparison

Average Health Scores by Department

IT Dept

8.2

Underwriting

7.1

Customer Svc

6.8

Claims

5.4

Trend Analysis (Claims Dept -- 90 days)

Period	Avg Score	Change	Key Factor
Jan 1 - Jan 31	6.8	--	Baseline period
Feb 1 - Feb 28	6.1	-0.7	Windows update impact
Mar 1 - Mar 15	5.4	-0.7	Disk space + aging HDD

Using Scores for Hardware Refresh

Move from age-based replacement ("every 4 years") to performance-based refresh:

Identify Persistently Poor

Filter for endpoints with 30-day average below 4.0.

Try Software Remediation

Some poor scores are fixable: runaway processes, full disks, misconfigured startups.

Flag Hardware Limitations

Machines still poor after remediation have genuine hardware limits: old HDD, insufficient RAM, weak CPU.

Prioritize by Impact

Cross-reference with department criticality. A poor-scoring claims adjuster machine has more business impact than a conference room PC.

Present to Leadership

"148 endpoints averaged below 4.0 for 90 days. Here's the department breakdown and productivity impact."

Key Insight

Health scores transform the refresh conversation from "we replace machines every 4 years" to "we replace when objective data shows they can't deliver acceptable experience." Some machines last 5 years; others need replacement after 3.

Deep Dive: Score Weighting Methodology ▼

Boot/Login Time (25%) gets the highest weight because every employee experiences it daily. Studies show employees with boot times over 3 minutes rate IT satisfaction significantly lower -- even if the machine is fast once booted.

CPU and Memory (20% each) both contribute to "snappiness" when clicking, switching windows, or loading files.

Disk Health (20%) captures I/O performance and capacity. The HDD-to-SSD transition often has the most dramatic score impact.

App Stability (15%) is lower because crashes are intermittent, while constant slowness is worse for overall productivity.

When to Adjust

VDI fleet? Reduce boot weight, increase CPU/memory
Known app stability issue? Temporarily increase stability weight
Fleet fully on SSDs? Disk health becomes less of a differentiator

Score Calculation Walkthrough

Given these sub-scores, calculate the composite health score and determine the category.

Score Calculator -- Endpoint CAEI445521

Dimension	Sub-Score	Weight	Contribution
CPU Health	6.0	0.20	1.20
Memory Health	4.0	0.20	0.80
Disk Health	3.0	0.20	0.60
Boot/Login Time	5.0	0.25	1.25
App Stability	8.0	0.15	1.20
Composite Health Score			5.05

Category: Fair -- Investigate and plan remediation. Disk and memory are dragging down the score.

🤔 What Would You Do?

An endpoint has a 30-day average health score of 3.5. Investigation reveals: disk is 95% full (HDD, not SSD), memory pressure at 88%, and boot time of 7 minutes. You clear 15 GB of temp files, but the score only improves to 4.1 because the HDD is the fundamental bottleneck.

What is the appropriate next step?

Mark as resolved -- 4.1 is in the "Fair" range now Keep clearing temp files until the score improves more Flag for hardware refresh -- the HDD is the root cause and cannot be fixed with software remediation alone Wait 6 months and check again

Correct! Software remediation helped slightly, but the aging HDD is the fundamental limitation. This endpoint is a strong candidate for hardware refresh -- ideally an SSD upgrade or full machine replacement.

Not quite. When software remediation only provides marginal improvement and hardware (aging HDD) is identified as the root cause, the endpoint should be flagged for hardware refresh.

Match the Score to Its Category and Action

Drag each score on the left to the correct response on the right.

Score: 9.2

Score: 6.8

Score: 4.5

Score: 2.1

Poor -- Immediate action required