Digital Preservation: A Practical Strategy for Protecting Digital Records

Digital preservation is the ongoing work of keeping digital files authentic, understandable, and usable as technology changes. It involves more than storing a backup: organizations need to know what they have, verify that files remain intact, preserve the context needed to interpret them, and plan for formats or software that may become obsolete. A digital preservation program combines policy, people, metadata, storage, and repeatable processes.

This guide explains the core practices, how digital preservation software can support them, and a step-by-step way to begin without trying to solve every future problem at once.

Why digital preservation matters

Businesses and institutions create records that may need to remain accessible for years: contracts, research data, product designs, financial records, photographs, training materials, and correspondence. A file can become unusable even if its bytes still exist. The application that created it may disappear, the format may be hard to render, or the organization may lose the metadata that explains who created it and why it matters.

The Library of Congress sustainability guidance discusses format factors such as disclosure, adoption, transparency, self-documentation, dependencies, and technical protections. The NDSA Levels of Digital Preservation offers a framework organizations can use to assess and improve a program over time.

Core parts of a preservation program

Inventory and selection

Start by identifying the records that need long-term care and why. Not every file deserves the same retention period or technical effort. Define ownership, business or legal value, access needs, and the event that triggers preservation. Link this work with records management and an approved data retention policy.

Integrity checks

A fixity check calculates a checksum for a file and later recalculates it to detect unexpected changes. Record the checksum, time, tool, and result. Checksums can show that a file has changed, but they do not prove that it was correct at the moment of ingest; capture fixity as early as possible and protect the record of each check.

Metadata and provenance

Preserve descriptive metadata that helps people find and understand a record, technical metadata that describes formats and dependencies, and provenance that explains how the file entered and changed within the repository. Keep metadata connected to the object and maintain clear identifiers. Without it, a valid file may still be impossible to interpret or trust.

Storage and redundancy

Use managed storage with monitoring, access controls, backups, and geographically or administratively separate copies when the risk warrants it. Redundancy reduces the chance that one failure destroys the only copy, but copies alone are not a preservation strategy. Verify them, test restoration, and make sure a compromised account cannot delete every replica.

Format planning

Assess format characteristics, tools, dependencies, and likely use. A widely documented format may be easier to render in the future, though no format guarantees permanent access. Preserve the original when authenticity matters and consider creating access derivatives for everyday use. Document any transformation and retain relationships between versions.

What digital preservation software should do

Preservation software can automate intake, metadata capture, fixity checks, format identification, storage replication, access controls, and reporting. Some systems manage digital objects in repositories; others focus on storage or workflow. Evaluate how components fit together rather than assuming one product covers the full lifecycle.

• Ingest files with stable identifiers and required metadata.
• Calculate and verify checksums, with alerts for failures.
• Track format identification, validation, and dependency information.
• Maintain relationships among originals, derivatives, and versions.
• Manage retention, access rights, and audit events.
• Export objects and metadata in documented, usable formats.
• Report on storage health, exceptions, and overdue actions.

Ask a vendor to demonstrate an end-to-end workflow using your own representative files, including an unsupported format, a duplicate, a corrupt object, and a restore. Verify that you can export your records if you change platforms. A preservation system should reduce operational risk rather than create a proprietary archive that is difficult to leave.

Build an implementation roadmap

1. Set priorities. Identify a small set of high-value content types and retention needs.
2. Assess current practice. Document existing storage, backups, formats, owners, and recovery evidence.
3. Define policy. Set ingest, access, integrity, retention, format, and incident rules.
4. Pilot a representative collection. Include normal cases and known problem files; test both preservation and retrieval.
5. Assign operational owners. Make clear who responds to failed fixity checks, storage alerts, and metadata gaps.
6. Measure and improve. Review results on a regular schedule and prioritize the highest-risk gaps.

The NDSA framework is designed for assessment and incremental progress. A small organization can use it to identify gaps without treating the highest maturity level as a universal target. Match controls to the value and risk of the records.

Security, privacy, and access

Long-term preservation does not mean broad access. Restrict records according to sensitivity, legal holds, contractual terms, and user roles. Encrypt transfers and storage where appropriate, protect administrative accounts, and log sensitive actions. Separate preservation copies from ordinary workspaces so routine sync errors or ransomware do not overwrite every retained copy.

Document how a record becomes available to an authorized person and how access is reviewed. Also plan for legal deletion or retention expiry. Keeping everything forever can create privacy, security, and cost risks; retention should follow a reasoned schedule.

Common mistakes

• Equating backups with long-term preservation.
• Choosing formats without considering how users will retrieve or render the content.
• Capturing no metadata or losing the link between an original and a converted copy.
• Running integrity checks but failing to alert an owner when they fail.
• Assuming a cloud provider is responsible for your retention and access policies.
• Failing to test restore and export processes.

Conclusion

Digital preservation is a managed lifecycle for keeping important files trustworthy and usable. Inventory and selection establish what matters; integrity checks, metadata, storage, and format planning help protect it. Begin with a focused collection, test the entire workflow, and improve controls as the organization’s needs and risks evolve.

Define what happens at each preservation stage

A documented workflow makes preservation repeatable. At intake, record the source, creator or submitting team, rights and restrictions, file inventory, and transfer date. Validate that files arrived intact, calculate fixity values, identify formats, and capture required metadata. If validation fails, quarantine the package and notify an owner rather than silently accepting incomplete data.

During ongoing storage, schedule integrity checks, review failures, maintain separated copies, and test restoration. When a format or dependency raises concern, document the risk and choose an action such as retaining the original, creating an access derivative, migrating a preservation copy, or using emulation. Record the tool, version, operator, date, and outcome for every transformation so future users can understand provenance.

Prepare for incidents and change

Write a response plan for missing objects, checksum mismatch, compromised credentials, corrupted storage, or a vendor outage. The plan should identify who can isolate a copy, compare replicas, restore a known-good version, and document the event. Run a restore exercise on a schedule. A backup that has never been restored is an assumption rather than tested recovery.

Review technology changes before they create urgent migration work. Track file formats, dependencies, storage costs, and access patterns. When a platform changes, test exports and metadata mappings on a sample collection before you migrate everything. Preserve the original object and transformation history when authenticity or auditability requires it.

Make progress measurable

Choose a small dashboard: collection inventory coverage, objects with fixity values, successful check rate, records with complete metadata, restore test results, and unresolved format risks. Assign an owner and target date to each gap. The NDSA assessment framework can help teams discuss maturity by functional area without assuming that the highest level is necessary for every collection.

Include sustainability and resource use in planning. More replicas, constant transformation, and indefinite retention all consume storage and energy. Preserve what has a documented value, choose controls proportionate to risk, and revisit the decision when access or retention needs change.

Preserve access without changing the record silently

Separate the preservation copy from the access copy when users need easier viewing or smaller downloads. A derivative may be converted, resized, or transcribed for accessibility, but the repository should link it back to the source and document the transformation. Keep a record of the software and settings used so another team can assess whether the derivative remains fit for purpose.

Metadata quality affects both trust and discovery. Minimum fields may include a persistent identifier, title or description, creator or source, creation and ingest dates, format, rights status, checksum, and preservation actions. The exact schema depends on the organization and collection. Make required fields practical enough that staff can complete them consistently.

Also define how users request restricted records. A file may be preserved but embargoed, confidential, or subject to rights restrictions. Store access rules with the record or in a linked policy system, and review them as dates and obligations change. Log access to sensitive items, but do not expose private information through public catalog metadata.

Finally, connect preservation to business continuity. Identify collections whose loss would stop a critical process, and test how quickly the organization can restore them in a clean environment. A recovery exercise should include credentials, decryption keys, software dependencies, and the instructions needed to render or interpret the data. The exercise often reveals missing context that routine checksum reports cannot detect.

Coordinate preservation across teams

Digital preservation touches records management, IT, security, legal, data owners, and people who need to use the content. Define a governance group or named coordinator who can resolve conflicts between retention, access, cost, and authenticity. The coordinator should not approve every ingest manually; instead, they maintain policy, exceptions, and review of high-risk collections.

Create a service catalog that tells staff what can be preserved, how to submit it, what metadata is required, and when an answer or access copy will be available. Include supported formats and a route to discuss unusual data. This makes the repository a reliable service rather than a storage location with unclear responsibilities.

Measure user outcomes as well as system checks. Can authorized employees find the record they need? Can they open it with supported software? Can an auditor follow its provenance? Are access requests handled within a useful timeframe? Use those questions to prioritize improvements. A repository with intact files but no practical retrieval path has not met its preservation goal.