Understanding Your Score

MungePoint produces a Copilot Readiness Score for each SharePoint site it scans. The score is a composite number from 0 to 100 that reflects how well your content is positioned for accurate Microsoft 365 Copilot retrieval. Higher is better.

The score is built from five independent dimensions. Each dimension measures a specific category of content quality that directly affects how Copilot finds, ranks, and presents information to users.

The five dimensions

1. Duplicates

What it measures: the density of byte-identical files across your document libraries.

When the same file exists in multiple locations, Copilot has no reliable way to determine which copy is authoritative. It may cite any version, blend information from multiple copies, or return conflicting answers to the same question depending on which copy ranks highest at query time.

MungePoint uses content-hash matching to find byte-for-byte duplicates across the scanned site. Each duplicate set is grouped and scored by its potential for Copilot confusion. A duplicate onboarding checklist is lower risk than a duplicate contract or HR policy.

What drives the score down: high numbers of duplicate file sets, duplicates in high-traffic libraries, and duplicates of document types that Copilot is likely to surface (policies, procedures, templates).

2. Staleness

What it measures: the age distribution of files and the concentration of outdated content.

Copilot does not age-weight search results. A benefits policy from 2018 and its 2024 replacement are both indexed and both eligible for retrieval. If the older document matches a user’s query, Copilot will surface it without indicating that it has been superseded.

MungePoint evaluates file modification dates, identifies content that has not been updated in configurable time windows, and flags libraries with high concentrations of stale material.

What drives the score down: large percentages of files with no modifications in extended periods, especially in libraries that contain governed content (policies, procedures, compliance documentation) where currency matters most.

3. Naming quality

What it measures: how descriptive and consistent file names are across your libraries.

File names are a retrieval signal. Copilot uses them alongside content and metadata when ranking results. Files with cryptic, auto-generated, or misleading names reduce retrieval precision. Names like Document1.docx, Copy of FINAL v2.xlsx, or export_20231115_143022.csv give Copilot no useful context about the file’s content or purpose.

MungePoint analyzes file names for patterns that indicate poor naming: auto-generated names, excessive version suffixes, copy-of patterns, encoding artifacts, and names that provide no semantic signal about the file content.

What drives the score down: high percentages of files with auto-generated names, copy-of prefixes, version suffix clutter, or names that do not describe the file’s content.

4. Metadata completeness

What it measures: how well your files and list items are tagged with structured metadata.

SharePoint metadata (content types, managed metadata columns, department tags, owner fields) functions as a filtering and ranking signal for Microsoft Search and Copilot. Documents with rich metadata surface more accurately for the users and queries they are relevant to. Documents with empty metadata are semantically flat and may not surface at all for targeted queries, or may surface inappropriately.

MungePoint evaluates metadata fill rates across your libraries, checking both built-in fields and custom columns. It identifies libraries where significant gaps exist and estimates the impact on retrieval accuracy.

What drives the score down: low fill rates on content type fields, empty managed metadata columns, and high-traffic libraries with minimal tagging.

5. Index noise

What it measures: the volume of content in the index that adds no value and degrades retrieval quality.

Index noise includes empty folders, zero-byte files, orphaned migration artifacts, temporary files, system-generated stubs, and other content that occupies space in the SharePoint index without contributing useful information. This content dilutes the signal-to-noise ratio that Copilot operates on.

MungePoint scans for known noise patterns: empty directory trees, files below meaningful size thresholds, common migration artifact signatures, and content that matches system-generated patterns.

What drives the score down: high volumes of empty folders, zero-byte files, migration leftovers, and other content that has no business value but is still being indexed.

The composite score

The composite Copilot Readiness Score (0-100) is a weighted combination of the five dimension scores. Each dimension contributes to the total based on its relative impact on Copilot retrieval quality.

A score of 100 means MungePoint found no issues across any dimension. This is unusual for real-world SharePoint sites.

A score in the 70-100 range indicates strong readiness. Issues are minimal and unlikely to produce visible Copilot retrieval problems.

A score in the 40-69 range indicates moderate readiness. There are meaningful content quality issues that could produce incorrect or confusing Copilot answers. Targeted cleanup is recommended before or shortly after Copilot deployment.

A score below 40 indicates significant readiness gaps. The content estate has substantial quality issues that are likely to produce visible Copilot retrieval failures. Cleanup should be prioritized before Copilot deployment.

Dimension drill-down

Each dimension score links to a detailed findings list. You can drill into any dimension to see:

The specific files, folders, or conditions contributing to the score.
The estimated impact of each finding on overall readiness.
Recommended actions for remediation.
The projected score improvement if the finding is resolved.

This per-finding detail is what feeds the Cleanup Workflow, where you can review, approve, and execute fixes.

Score tracking over time

MungePoint stores historical scan data locally. After running multiple scans (especially after cleanup work), you can view score trends to measure improvement. This is useful for:

Demonstrating progress to stakeholders before a Copilot rollout.
Verifying that cleanup actions had the expected impact.
Identifying sites where content quality is drifting back after initial remediation.