How to evaluate a collections management system for your museum

Paul Gauguin, Nature morte au panier carré, 1899, oil on canvas, 61.5 x 73.5 cm, Nasjonalmuseet, Oslo. Image courtesy of Nasjonalmuseet/Børre Høstland.

Choosing a new collections management system affects how a museum records, finds, shares, and protects information about its collection. The decision therefore requires more than comparing feature lists or watching product demonstrations.

A collections management system, often abbreviated to CMS, stores and manages information about objects and related museum work. Depending on the institution and the system, this may include accessioning, locations, loans, conservation, exhibitions, images, rights, and public collection information.

Some systems include these functions directly. Others depend on optional modules or connections with separate tools. A digital asset management system, for example, manages image, audio, video, and document files. It may be integrated with the collections management system without being part of the same technology.

There is no single system that is best for every museum. The right choice depends on the collection, staff, existing data, institutional policies, technical capacity, budget, and work the system must support.

This guide presents a practical process for identifying those needs, testing possible systems, and documenting a defensible decision.

1. Start with the problem, not the software

Before researching vendors, write down why the museum is considering a change.

Possible reasons might include:

• Staff cannot reliably find collection information.

• Location records are incomplete or difficult to update.

• The existing system no longer receives technical support.

• Staff enter the same information in several places.

• Images and rights information are disconnected from object records.

• The system cannot exchange information with other museum tools.

• Public collection information must be prepared manually.

• Staff cannot work effectively from required locations.

• Current permissions do not adequately protect sensitive information.

Describe the problem as precisely as possible. “We need modern software” does not explain what must improve. “Registrars need to update an object’s location and retain its complete movement history” gives the evaluation team something it can test.

Also consider whether replacing the system is the appropriate response. Some problems may result from missing policies, inconsistent data, insufficient training, or an existing system that has not been configured well. Replacing technology will not correct these problems automatically.

Define what a successful change would accomplish. Success might mean reducing duplicate data entry, improving the reliability of location information, supporting a documented collections procedure, or allowing selected records to be published without re-entering them.

2. Build an evaluation group and map real work

The people who use collection information should participate in the evaluation. This reduces the risk of selecting a system that works well in a demonstration but poorly in daily practice.

Depending on the institution, the group may include:

• Registrars and collection managers

• Curators and researchers

• Conservators

• Rights, photography, or image staff

• Exhibition and loan coordinators

• Digital publishing or public-access staff

• Information technology and security staff

• Staff or advisors with accessibility expertise

• Someone responsible for procurement and budgets

Not every participant needs to attend every meeting. The important point is that the museum understands the requirements of frequent users, occasional users, administrators, and people who depend on information exported from the system.

Ask staff to describe real tasks from beginning to end. Examples include:

• Receive an object and assign an entry number.

• Formally add an object to the collection.

• Record and later update an object’s location.

• Produce a list of objects for an exhibition.

• Add a condition report with images.

• Prepare and monitor an outgoing loan.

• Record a copyright holder, licence, credit line, and restriction.

• Find every work associated with a person, place, or exhibition.

• Correct a name used across several records.

• Publish selected information to the museum’s website.

• Determine who changed a record and when.

The Spectrum collections management standard can help museums identify the procedures their systems need to support. Spectrum is a documentation standard, not a software product.

3. Audit the existing data and connected systems

A system cannot be evaluated properly without understanding the information that must move into it.

Document:

• The number and types of records in the current system

• Spreadsheets, card files, shared drives, and separate databases

• Images, documents, and other files connected with collection records

• Local terminology and controlled vocabularies

• Custom fields and reports

• Object, accession, loan, and other identifiers

• Relationships between objects, people, places, exhibitions, and events

• Rights and access restrictions

• Record histories and audit information

• Connections with websites, digital asset systems, finance systems, or other tools

Identify duplicate records, inconsistent terminology, incomplete fields, and information that cannot currently be exported. These findings affect the time, cost, and risk of changing systems.

Migration involves several different activities:

• Data cleaning: correcting or reconciling existing information

• Mapping: deciding where information belongs in the new system

• Migration: moving the information and files

• Validation: checking that nothing was lost, shortened, disconnected, or changed incorrectly

Ask potential vendors to test a representative sample of the museum’s actual data. Include difficult records, attachments, relationships, special characters, long text, uncertain dates, and local fields. A perfectly clean demonstration dataset will not reveal how a system handles the museum’s real information.

The UK National Archives similarly recommends auditing the current system, connected databases, custom fields, and technical environment before procuring a replacement.

4. Turn the museum’s work into ranked requirements

Convert the workflows and data audit into clear requirements.

Use three priority levels:

• Essential: the museum cannot adopt a system that fails this requirement.

• Important: the requirement has substantial value, but a documented alternative may be acceptable.

• Optional: the feature would be useful but should not determine the decision.

The Canadian Heritage Information Network’s collections management system checklist contains full and abridged sets of criteria that museums can adapt. Its categories include collections procedures, search, reporting, terminology, data exchange, security, support, and training.

Do not adopt every possible criterion. Select and adapt the requirements that follow from the museum’s own work.

Common evaluation areas include:

• Object identification and status

• Locations and movements

• Acquisition, accessioning, and removal from the collection

• Loans and exhibitions

• Conservation and condition reporting

• Images, documents, rights, and permissions

• Search and reporting

• Controlled terminology

• User permissions and change histories

• Public access and data sharing

• Import, export, and connections with other systems

• Training, documentation, and technical support

Write each requirement as a result that can be tested.

Instead of:

The system has a conservation module.

Write:

A conservator can create a condition report, attach three images, record the examiner and date, revise the report, and see who made each change.

A long feature list should not hide a failure in an essential workflow. One feature that the museum cannot work without may matter more than twenty attractive extras.

5. Test every system with the same scenarios

Ask every shortlisted vendor to demonstrate the same tasks using the same instructions.

Whenever possible:

• Supply anonymized examples of the museum’s records.

• Ask vendors to import a representative data sample.

• Let staff complete tasks instead of only watching a presentation.

• Include an unusual or difficult record.

• Test both routine and occasional work.

• Record which functions require configuration, customization, or additional payment.

• Distinguish functions that exist now from those planned for a future release.

• Confirm important answers in writing.

For each task, record one of four results:

• Works as required

• Works with additional configuration, work, or cost

• Does not work

• Not tested

This produces comparable evidence. Without a common test, one vendor may present attractive public pages while another demonstrates detailed registration work. The museum cannot compare them fairly because they have answered different questions.

Ask for references from institutions with collections, staffing, and technical capacity similar to your own. Ask those institutions what required more time than expected, which workarounds remain, how support operates, and what they would do differently.

Test accessibility directly

“Is the system accessible?” is too broad to produce a reliable answer.

Test whether staff can use essential functions with a keyboard, screen magnification, and relevant assistive technology. Check form labels, error messages, focus order, contrast, and the ability to enlarge text. Test staff interfaces separately from public collection pages.

The Web Content Accessibility Guidelines provide a framework for evaluating web-based interfaces. Applicable legal requirements will depend on the organization and jurisdiction.

Remote access is a different question. If staff need to work from several locations, test how they log in, whether a secure connection is required, and what happens during an internet or system outage.

6. Examine security, continuity, and the ability to leave

A collections management system holds information for which the museum remains responsible. Ask where that information is stored, who can access it, how it is protected, and how the museum can recover it.

Questions should address:

• Which organization owns the records and uploaded files?

• Where is the information hosted?

• Who can view, create, edit, export, and delete information?

• Can the museum require a second verification step at login?

• Does the system retain a record of important changes?

• How are updates and security problems handled?

• How often are backups created?

• Has restoration from backup been tested?

• What happens after an outage or security incident?

• How and when will the museum be notified of a problem?

• What happens if the vendor closes, changes ownership, or discontinues the system?

Security evaluation should involve the museum’s technical staff or an appropriate advisor. The NIST Cybersecurity Framework provides a general framework for identifying and managing cybersecurity risks, although the museum will need to translate it into requirements appropriate to its own circumstances.

Require a usable exit

The ability to download a spreadsheet is not necessarily a complete exit plan.

Confirm that the museum can retrieve:

• Collection records

• Stable identifiers

• Names and controlled terms

• Relationships between records

• Images and documents

• Rights and access information

• Relevant revision or audit information

• Documentation explaining the exported structure

Ask which formats are available, how long an export takes, what it costs, and whether media files and their metadata remain connected. If an application programming interface, or API, is part of the plan, ask about access limits, documentation, and additional fees. The Digital Preservation Coalition’s procurement guidance recommends addressing access, preservation, security, portability, and termination responsibilities when external providers manage digital material.

7. Compare the full cost of ownership

The subscription or licence price is only one part of the cost.

Request a written estimate covering:

• Initial setup

• Data cleaning, mapping, migration, and validation

• Configuration and customization

• Connections with other systems

• Storage for records and files

• User licences

• Training

• Accessibility testing and remediation

• Technical and editorial staff time

• Support packages

• Updates and upgrades

• Additional modules

• API or data-transfer charges

• Renewal increases

• Export and termination costs

• A future move to another system

Compare costs over a meaningful period, such as five years. Ask how the price changes when the museum adds users, records, storage, public access, or additional functions. A lower initial price may be offset by migration, customization, staff work, or recurring support. A higher price is not evidence of a better fit. Costs need to be evaluated alongside the museum’s essential requirements and capacity to operate the system.

8. Score the evidence and document the decision

Create one evaluation record for every important requirement. A narrow, vertical format will remain readable on both desktop and mobile.

Requirement or task:
What must the museum be able to do?

Priority:
Essential, important, or optional

Evidence from the test:
What did staff or the vendor demonstrate?

Result:
Works as required, needs additional work or cost, does not work, or was not tested

Additional cost or condition:
Does the result depend on configuration, customization, another system, or a higher subscription?

Risk or follow-up:
What remains uncertain or must be confirmed?

Responsible person:
Who will obtain the answer or approve the requirement?

Do not rely only on an average numerical score. A high overall result can conceal failure in an essential area such as migration, security, accessibility, or complete data export.

Before signing a contract, confirm that the museum has:

• An agreed list of requirements and priorities

• Results from task-based demonstrations

• A tested sample migration

• A written implementation and training plan

• A complete cost estimate

• Written support and security terms

• Evidence that the museum can retrieve its information

• References from comparable institutions

• Named responsibility for implementation and acceptance testing

Document why the chosen system meets the museum’s most important needs and what compromises were accepted. This record will help future staff understand the decision and evaluate whether the system continues to serve the institution.

Choose a system the museum can sustain

No collections management system will remove the need for clear policies, consistent documentation, trained staff, and continuing oversight. The strongest choice is not the system with the longest feature list. It is the one that supports the museum’s essential work, preserves the meaning and relationships in its information, can be used by its staff, and remains financially and technically manageable over time. If the institution is selecting technology specifically for a catalogue raisonné, its requirements may extend beyond an institution’s own holdings. Our guide to choosing a digital strategy for a catalogue raisonné examines purpose-built platforms, assembled systems, and custom development for that different use case.