Library Digitization Services: What Libraries Should Require Before Outsourcing a Collection

Library Digitization Services: What Libraries Should Require Before Outsourcing a Collection

Libraries digitize collections for many reasons: protecting fragile originals, expanding researcher access, improving discovery, supporting teaching, and ensuring that important material remains usable long after physical copies deteriorate.

But successful library digitization is not simply a matter of putting books on a scanner.

A professional library digitization workflow can involve collection assessment, preservation-aware capture, image quality control, OCR, structural processing, metadata, file-format planning, repository ingest and long-term preservation requirements.

For libraries considering an external partner, the real question is therefore not:

“Can this company scan our books?”

It is:

“Can this company create trustworthy digital assets that our library can discover, preserve, manage and reuse for years?”

What Are Library Digitization Services?

Library digitization services convert physical library and special-collection materials into structured digital assets.

Depending on the collection, this can include:

• Books and monographs
• Journals and periodicals
• Newspapers
• Manuscripts
• Theses and dissertations
• Photographs
• Maps and oversized materials
• Archival documents
• Rare books
• Ephemera and special collections

The process can include far more than image capture.

A complete workflow may look like:

Collection audit → preparation → capture → image QC → OCR → structural processing → metadata → preservation files → access derivatives → repository ingest → final QA

Government and cultural-heritage guidance similarly stresses that digitization decisions made during planning—including image quality, metadata and documentation—can affect whether digital objects remain useful in the future.

1. Start With the Collection, Not the Scanner

Before digitization begins, libraries should understand what is actually in the collection.

A collection audit should identify factors such as:

• Material type
• Number of volumes or items
• Page counts
• Physical condition
• Binding type
• Oversized or unusual objects
• Languages and scripts
• Illustrations, maps and photographs
• Existing catalogue records
• Copyright or rights restrictions
• Preservation priority
• Expected researcher use

This information determines the capture method, resolution, handling procedures, metadata requirements and level of post-processing required.

A fragile nineteenth-century bound volume should not necessarily move through the same capture process as a modern loose-leaf document.

Ask the vendor

Before awarding a large project, ask:

How will you segment our collection by material type and condition before production starts?

A credible workflow should not assume that every object can be processed identically.

2. Define Image Quality Before Production Begins

Digitization quality should be defined in the project specification before the first production batch is scanned.

For cultural-heritage materials, one important reference is the Federal Agencies Digital Guidelines Initiative (FADGI).

FADGI provides technical guidelines for digitizing still-image cultural heritage materials including textual documents, manuscripts, maps, photographs and similar collections. Its current still-image technical guidelines are the third edition published in 2023.

Libraries may also work to other national, institutional or project-specific requirements.

The important point is that the vendor and library should agree on measurable capture requirements.

These may include:

• Resolution
• Colour accuracy
• Tonal response
• Cropping
• Skew
• Completeness
• File naming
• Bit depth
• Master-file format
• Derivative formats

Important clarification about FADGI

FADGI publishes guidelines and measurement approaches, but it does not certify or endorse commercial digitization companies or products.

So a procurement team should ask a vendor to explain how its workflow measures against the relevant FADGI criteria, rather than relying on an unsupported “FADGI certified” marketing statement.

3. Decide What the Preservation Master Will Be

The file a researcher downloads does not necessarily need to be the same file preserved by the institution.

Large digitization projects often distinguish between:

Preservation master

A high-quality file retained for long-term preservation and future reuse.

Access derivative

A smaller or more convenient format provided for viewing, download or distribution.

OCR or text derivative

Searchable text or structured output generated from the image.

For some still-image digitization workflows, TIFF remains an important preservation master format. The Library of Congress notes substantial use of TIFF within its collections and references TIFF as a primary/master-file format for multiple content categories under its digitization practices.

The correct format, however, depends on the collection and repository requirements.

A library should therefore specify outputs before scanning starts, not decide them after thousands of items have already been processed.

4. OCR Should Make Collections Searchable—not Merely Digital

A page image preserves visual information.

OCR adds another layer: machine-readable text.

That can allow researchers to:

• Search within books
• Discover names and terms
• Index full text
• Copy relevant text
• Run text analysis
• Build digital scholarship projects
• Retrieve content through repository search

But OCR quality can vary dramatically.

Historical typography, damaged pages, unusual fonts, multi-column layouts, marginalia, mathematical expressions and non-Latin scripts can all make recognition more difficult.

That is why OCR should have its own QA process.

A robust workflow should distinguish among:

Capture quality → OCR quality → structural quality

A page can be beautifully scanned while still containing poor OCR.

What should libraries ask?

Ask the vendor:

• How is OCR quality measured?
• How are low-confidence pages identified?
• Is human correction available?
• How are headings and logical structures preserved?
• How are tables handled?
• How are multilingual collections handled?
• What happens when OCR confidence falls below the agreed threshold?

Gentize's current digitization workflow includes OCR confidence sampling and additional checking of low-confidence material as part of its digitization QA process.

5. Metadata Is What Makes Digitized Collections Discoverable

Scanning creates digital files.

Metadata helps users find and understand them.

For library projects, metadata may describe:

• Title
• Creator
• Date
• Publisher
• Subject
• Identifier
• Language
• Rights
• Format
• Collection
• Relationships between objects

The exact schema depends on the repository and institution.

MARC 21

MARC 21 remains widely used for exchanging bibliographic information. The Library of Congress maintains the official MARC 21 bibliographic format, with its latest online documentation incorporating Update No. 42 from May 2026.

Dublin Core

Dublin Core provides broadly applicable metadata properties for describing resources and remains widely used across digital repositories and interoperable metadata systems.

A library digitization vendor therefore needs to understand more than scanning.

It should be able to preserve, create, validate or transform metadata according to the institution's required schema.

6. Plan for Repository Ingest Before Digitization Starts

A common mistake is treating delivery as:

“Here is a hard drive containing all your scans.”

That is not necessarily the end of a digital-library project.

The library should determine early where the collection will ultimately live.

For example:

• Institutional repository
• Digital asset management system
• DSpace
• Islandora
• CONTENTdm or equivalent platform
• IIIF-enabled infrastructure
• Custom digital library
• Preservation repository

This affects file naming, metadata mappings, identifiers, folder structures and packaging.

Gentize's current library digitization service supports repository-oriented delivery, including workflows involving DSpace, Islandora, MARC21, Dublin Core, OAI-PMH and IIIF-ready content.

Why IIIF can matter

IIIF provides standardized APIs for working with image-based digital objects.

Its Presentation API can describe the structure and layout of digitized objects such as books, manuscripts, newspapers, maps and other image-based resources for use in compatible viewers and systems.

For institutions planning interoperable digital collections, these downstream requirements should be included in the specification at the beginning.

7. Quality Assurance Must Cover More Than Missing Pages

A useful QA process checks multiple layers.

Capture QA

Verify:

• Every required page was captured
• Pages are in the correct sequence
• Images are not unintentionally cropped
• Orientation is correct
• Required image-quality specifications are met

OCR QA

Review:

• Recognition confidence
• Character errors
• Missing text
• Reading sequence
• Language-specific problems

Metadata QA

Validate:

• Required fields
• Identifiers
• Naming consistency
• Schema conformity
• Relationships between records and files

Delivery QA

Confirm:

• Correct master files
• Correct derivatives
• Correct filenames
• Correct directory/package structure
• Repository compatibility

A strong project specification should also define what happens when an item fails QA.

The answer should be something like:

Identify → correct → recheck → document

not simply:

accept the error and continue production.

8. Protect Originals Throughout the Workflow

When materials leave the library, physical control becomes an important procurement consideration.

Library and Archives Canada specifically notes that outsourced digitization requires consideration of record security during transportation and digitization, alongside ongoing communication with the vendor.

Libraries should therefore ask about:

• Inventory controls
• Item-level tracking
• Packing procedures
• Secure transport
• Restricted-access collections
• Storage environment
• Handling requirements
• Damage escalation procedures
• Chain-of-custody documentation
• Return reconciliation

For particularly fragile, rare or restricted material, on-site digitization may also be worth evaluating.

9. In-House vs Outsourced Library Digitization

There is no universal answer.

In-house digitization can make sense when:

• The collection is small
• Digitization is ongoing
• Materials cannot leave the institution
• The library already has suitable equipment
• Skilled staff are available
• The institution wants to build internal expertise

Outsourcing can make sense when:

• Volumes are large
• Production deadlines are tight
• Specialized scanners are required
• Multiple material types are involved
• OCR or metadata work is substantial
• The library lacks production capacity
• Large-scale QA is required

Library and Archives Canada's guidance notes potential outsourcing benefits such as avoiding some upfront technology investment, gaining economies of scale and accessing specialist expertise, while also stressing vendor capability, standards and security.

A hybrid model is also possible.

For example, rare materials could remain on-site while more stable collections are processed externally.

10. Library Digitization Vendor Checklist

Before signing a contract, evaluate whether the provider can clearly answer the following.

• Collection assessment — How will different formats and conditions be evaluated?
• Handling — What happens with fragile, rare or bound material?
• Capture standard — What measurable image-quality requirements will be followed?
• Equipment selection — How will capture equipment change according to material type?
• Preservation masters — Which formats will be delivered and why?
• OCR — What recognition and correction workflow is included?
• Metadata — Which schemas can be supported or mapped?
• Quality assurance — What is automatically tested and what receives human review?
• Security — How are originals tracked and protected?
• Repository integration — Will the files and metadata arrive ready for your target repository?
• Error handling — How are failed items corrected and reprocessed?
• Reporting — Will you receive an item-level production and QA record?

The more precisely these questions are answered before production, the less ambiguity remains once thousands of objects enter the pipeline.

From Physical Collection to Searchable Digital Library

The best digitization programmes do more than produce scans.

They transform a physical collection into a searchable, structured and reusable digital resource while preserving enough quality and metadata to support future use.

That requires coordination across:

Capture → OCR → Metadata → Quality Assurance → Preservation → Discovery → Repository Delivery

When these requirements are established before production, libraries reduce the risk of rescanning collections later because files, metadata or documentation are inadequate.

Government cultural-heritage guidance makes the same planning point: early decisions about resolution, documentation, metadata and intended use can determine whether digital assets remain suitable for future needs.

How Gentize Supports Library Digitization

Gentize Innovations provides end-to-end library digitization for public, academic, national and special-collections libraries, with services spanning bulk capture, OCR, metadata and repository-ready digital delivery.

The current Gentize library workflow includes capabilities around:

• Collection auditing
• Preservation-oriented capture
• OCR and structural processing
• MARC21 and Dublin Core metadata
• Archival masters and access derivatives
• OAI-PMH-oriented discovery workflows
• IIIF-ready delivery
• Repository ingest
• Item-level quality control

Gentize's broader production capability also supports large-scale digitization programmes; its current company information states a capacity of 1,000 books digitized per day and 450,000 pages scanned per day.