ROT Data Cleanup for Microsoft Copilot: Managing Redundant, Outdated, and Trivial Content

If it doesn’t add value, then it’s just noise.

Managing Your ROT (Redundant, Outdated, or Trivial) Data

We are approaching the final installments in our series on transforming Copilot chaos into clarity. One of the biggest factors in Copilot data readiness and long-term Copilot ROI is whether organizations have a clear strategy for ROT data cleanup.

Did you know that for an IT project to be called “successful,” it requires 20% ROI in year one? That increases by 10% to 20% every year, for 3 years.[ii]

That means you will need to have between $720,000 to $1m in proven ROI in year one to call your Copilot rollout successful.[iii]

This week, we will focus on the most crucial step to ensure you achieve ROI:

1. Getting quality responses from Copilot
2. Improving the quality and speed of Copilot responses.

Improving response quality starts with ROT data cleanup for Microsoft Copilot, ensuring Copilot is trained on current, trusted, and relevant content instead of redundant or outdated files.

Without proper cleanup, Copilot must evaluate multiple versions of the same document, which directly reduces Copilot response accuracy and user trust.

What is ROT Data?

ROT data refers to Redundant, Outdated, or Trivial content that no longer adds business value but continues to consume storage, increase risk, and slow down systems like Microsoft Copilot.

• Redundant data includes duplicate files, multiple versions of the same document, and content saved across OneDrive, SharePoint, Teams, and email without a single source of truth.

• Outdated data is content tied to closed projects, former employees, expired initiatives, or information that is no longer accurate or relevant.

• Trivial data consists of low-value content that was never intended for long-term use but was never cleaned up or archived.

In a Copilot-enabled environment, ROT data becomes more than clutter. When Copilot is asked to answer a question, it must evaluate all available versions of a file, including outdated and duplicate content. This directly impacts Copilot response accuracy, response time, and user trust.

In short, ROT data creates noise. And when noise outweighs signal, Copilot cannot reliably deliver value.

Do You Know How Much Data You Really Have?

It is estimated that in 2025, there is around 402.74 million TB of data created worldwide, every day.[iv]

So, with that said,

1. The average business user creates between 500 GB and 2 TB of content each year.
2. On average, over 80% of the content (Word, PowerPoint, Excel, and Emails) a user works with has been created in the past 90 days.
3. That number reduces to 15% of content created in the past 4-12 months and only 5% of content that is accessed is more than a year old.

This volume of unmanaged content is a primary reason organizations struggle with Copilot data readiness and inconsistent AI responses.

Ok, now follow the logic here:

Think about how many end users still:

1. Attach files to emails for internal use.
2. Open documents and then save local copies.
3. Do not understand the difference between copy and move to SharePoint, Teams, or other similar tools.
4. Do not use links to share content.

Now, we make all that content available to Copilot and ask it to look at 5 or even 10 versions of the same file across multiple OneDrives, SharePoint, and Email. How accurately do you think it will be able to choose which ones to use to respond to a user’s Copilot query?

Garbage in; garbage out. This is why ROT data cleanup for Microsoft Copilot is not optional if you want reliable answers and measurable ROI.

How to Reduce ROT Data and Improve Copilot Response Accuracy

Phase 1- Clean Up Your Data

The steps I recommend for this process: Export and Crawl, Import and Merge, Standardize, Deduplicate, and Verify.

Cloud Data- Remove any incomplete, incorrect, inconsistent, or duplicate data.

Teams/SharePoint- Remove any duplicate data. Plan to archive old, similar, orphan teams/sites. You may also find some that require splitting.

OneDrive/Exchange – Plan to archive any expired, archived, or closed projects as well as any legal holds, or ex-employee content.

This phase establishes the foundation for SharePoint data cleanup and long-term Microsoft 365 data governance.

Phase 2 – Increase Quality of Responses While Reducing Cost

Increase the quality of your remaining data.

Phase 2 focuses on improving Copilot response accuracy while controlling Microsoft 365 storage and infrastructure costs.

• Archive any email older than 12 – 18 months old.

• Archive any documents that are more than 3 years old and not accessed in the past 90 days.

Reduce SharePoint Storage Costs

Move the content that is no longer current and/or needed for legal hold to Azure Hot or Cold storage.

This approach not only helps reduce SharePoint storage costs; it also removes low-value data that negatively impacts Copilot performance.

Let’s use an example to illustrate potential savings. Let’s assume you have 3.2 TB of SharePoint data total (in excess of the included tenant and licensing storage allotments), but 1.5 TB is ROT data and can be migrated to Azure Cold Storage.

• Original Monthly SharePoint Online Storage Cost (all 3.2 TB)
  • SharePoint Online storage is $0.20 per GB per Month
  • Your estimated monthly storage cost will be ~$655.36/month
• After Migrating 1.5 TB to Azure Cold Storage:
  • New monthly SharePoint Online Storage Cost
    • 8 GB remains at $0.20/GB = $348.16 per month
  • New Monthly Azure Cold Storage Cost (using pay-as-you-go estimates)
    • 1536 GB at $0.0036/GB = $5.53 per month
  • Your estimated monthly storage cost will be ~$353.69/month
    • One caveat to keep in mind: this estimate is the data storage pricing only and does not include pricing for retrieval, etc.
• Using Azure Cold Storage strategically in this situation could save you up to ~$3,620 annually.

Using Azure Cold Storage strategically helps organizations lower costs while supporting cleaner data inputs for Copilot.

Improve Copilot Response Time and Accuracy

When proper data cleanup has been completed, here is an average of the results we have seen:

• Accuracy of Copilot responses increased by 83%
• Copilot response time increased by 62%

These results are directly tied to consistent ROT data cleanup for Microsoft Copilot and disciplined data lifecycle management.

Look, you know you need to reduce and clean your data. Like moving from one house to another, you can either pick up everything from house 1 and move it to house 2 (not recommended, especially if smaller), or you can donate, throw/give away, or store what you haven’t used or might never use.

It is time to remove the ROT data to save you money and improve the quality of your Copilot responses. This will improve the user experience with Copilot, further encourage use and increase productivity, and improve Copilot ROI.

The final step in sustaining Copilot ROI is applying strong Microsoft 365 data governance practices across SharePoint, OneDrive, and Teams.

Now you are ready for the final steps to improve the quality of Copilot responses:

• Choose which SharePoint site data is being crawled and which ones don’t need to be part of the Copilot searches i.e.: old and archived data not yet in Azure Cold Storage

• Apply appropriate data sharing, security, and governance controls in OneDrive and SharePoint to help monitor sharing habits.

• Create Microsoft 365 governance reports to look for irregularities in sharing habits.

• If you have E5 licenses, you can turn on Auto-labeling for content. This improves metadata and better manages security and discoverability of content.

• If you have E3 or E5, you can add-on Microsoft Syntex. This allows you to add any unique or industry-specific terminology to Microsoft Syntex to increase discovery and search accuracy. It is an additional cost.

• If enabled, don’t forget to leverage DLP and Purview to better manage content security.

• Use PowerApps to archive, tag, or move ROT content (old, out of date, expired or content that does not fit into that 3-month/90-day rule) to “keep your house clean.”

Removing ROT data improves Copilot data readiness, reduces storage costs, and significantly increases Copilot ROI by delivering faster, more accurate responses users can trust.

In the next post, we will dig into proven change management techniques to drive change and usage of M365 and Copilot to better achieve your ROI goals.

[i] Purchased and used with permission from marketoonist.com

[ii] Foster Capital – 2023 ROI Report

[iii] Based on a company with 10k seats of Copilot at a cost of $360 per user, per year

[iv] Gartner, Forrester, and WSJ – 2024,2023, 2025 

Get clarity on your Copilot data readiness.

ENow’s Copilot Center Organizational Readiness Reports show you where governance gaps exist, what data needs attention, and what to fix before Copilot usage scales.

Explore Copilot Center Readiness Reports