Explore Archival Versus Purge in Data Cleanup
Learning Objectives
After completing this unit, you’ll be able to:
- Differentiate between archive, purge, and backup as data cleanup strategies.
- Identify archival or purge candidates using record age and business relevance.
- Describe how to access archived records to support compliance, analytics, and AI use cases.
Learn Why Archiving and Purging Matter
Data cleanup isn’t only about fixing values. It also includes deciding which historical data should remain operational, which should be accessible only when needed, and which should be purged.
At Northern Trail Outfitters (NTO), customer service leaders want faster, more reliable support, especially during warranty periods, while marketing leaders want an accurate view of total lifetime value (LTV). Luna sees a common risk—teams keep too much data just in case, which increases cost and complexity, yet still fail to preserve the right customer context for agents and analytics.
Luna starts by aligning cleanup decisions to NTO’s data retention policy and business needs.
Define Archive, Purge, and Backup
Organizations need to keep data available to run their business and follow rules and regulations. Data retention policies help decide what to do with data that is no longer needed every day. This data can either be archived, stored safely for later use, or deleted after it’s no longer required.
A backup is different from archiving. A backup creates a copy of data so it can be restored if something goes wrong. It’s important to create a backup before cleaning or moving data.
Archiving means saving a copy of the data and removing it from the main system to free up space while still keeping it available if needed.
Term | When to Use | Key Consideration |
|---|---|---|
Backup Keep a copy of data for recovery if something goes wrong. | Use before data cleanup, migration, or major changes. | Backup helps with recovery, but it does not replace archiving or retention policies. |
Archive Move older data out of daily use but keep it available. | Use when data must be kept for compliance, audits, or future use. | It’s important to preserve important details, like who, what, when, or amount, so the data can be understood later. |
Purge Permanently delete data. | Use when the retention period has ended and the data is no longer needed. | Deleting data too early can cause problems with audits, disputes, or reports. |
Remember: When you’re unsure, archive the data instead of deleting it—based on your retention policy.
To learn more about retention policies, check out the Archive Basics badge.
Detect and Decide: Identify Archival and Purge Candidates
Regardless of your backup and archiving solution, you need to make a series of decisions to establish and implement your data retention policies and procedures.
You can use this table to match what appears in your data with what to do next. It also shows how data profiling can help you check your assumptions and support each decision.
Scenario | How to Check | What to Do |
|---|---|---|
Records are old and rarely used but still must be kept. | Look at Created Date, Last Modified Date, and last activity (if available). Use data profiling to see how much data is both old and unused. | Archive the data so it’s still available if needed. Only delete it after the retention period ends and it’s no longer needed. |
Storage costs are high. | Check record counts and storage usage. Look at segments like “Cases older than 12 months” or “Orders older than 3 years” to estimate savings. | Start with the data that saves the most space. Use these numbers to align on scope and timeline. |
Teams need recent data but not the full history. | Start with the use case. Identify what data is actually used, like recent cases or LTV. Use data profiling to find which fields matter and which don’t. | Keep recent, high-value data active. Archive older details, but keep key summary fields like IDs, dates, and amounts. |
The same data exists in multiple systems. | Compare systems using data profiling, such as record counts, completeness, or value patterns. Look for differences or gaps. | Fix differences first. If the data is consistent, choose how to access it—for example, via a shared data layer—and proceed with archiving. |
A simple way to think about this is to group your data into three types
- Operational detail: Data needed for daily work.
- Historical detail: Data that is used less often but still needed.
- Long-term summary: Key data kept for reporting, analytics, and AI.
This approach helps you decide what to keep, archive, and purge.
Keep the Customer Context Without the Clutter
NTO doesn’t need all historical data forever. Luna proposes a minimum viable historical context approach.
- For Orders, she retains key details that support LTV and customer understanding, like order date, customer ID, order amount, and credits or returns.
- She preserves full order history only as long as it supports the business, such as 3 years for marketing use.
- For Cases, she retains what is required for compliance, but archives older records so they can be accessed only when needed.
This approach helps NTO understand customers for analytics and AI, without carrying extra data that slows things down.
In short, keep what you need, limit what you don’t, and archive the rest.
Ready to Move On?
In the next unit, you discover how Luna addresses unused fields and configurations.