When a rights clearance workflow stalls, the cause is almost never the law — it is process leakage: items entering an undefined "unknown" state and never leaving it, dates being mis-typed, or decisions living in a spreadsheet that diverges from the catalogue. Fix clearance by triaging every item into a small set of well-defined buckets at intake, recording the evidence not just the verdict, and writing the result back as a machine-readable field. Below are the failures I see most and how to clear them.

Why does my backlog keep growing?

The classic symptom: a team clears items steadily yet the queue never shrinks. The root cause is usually a missing triage step. Without it, every item — even an obvious 1850 lithograph — flows into the same deep-research pipeline.

Fix it with a fast first-pass that routes items by a cheap test:

Intake triage (seconds per item):
  date of creation < 1880?            -> likely PD, light check
  staff-created / institutional?      -> cleared, internal licence
  has a deposit/gift agreement?       -> apply agreement terms
  everything else                      -> queue for research

In practice 60–80% of a typical archive resolves in the first three branches, leaving a manageable research queue instead of an undifferentiated mountain.

What is the single most damaging data error?

Mixing up creation date and publication date. UK term rules for many categories depend on publication and on the author's death — not on when the object was made. An unpublished manuscript can have a different term than a published one of the same age. If your spreadsheet only has one "date" column, term calculations silently go wrong.

Add explicit, separate fields and validate them:

date_created      (when the work was made)
date_published    (first publication, if any)
date_author_death (for p.m.a. terms)
date_basis        (which event drives the term decision)

How do I get out of "rights status: unknown" limbo?

"Unknown" is not a destination — it is a symptom of an unfinished diligent search. Define what "diligent" means up front so it terminates:

1. Check the obvious sources (catalogue, accession file, donor correspondence).
2. Search rights registries and the EUIPO Orphan Works database where relevant.
3. Try the named author/estate via two channels, with a deadline.
4. If unresolved after that fixed effort, promote the item to orphan with the search log attached — do not leave it as unknown.

The trick is the deadline: an open-ended search never ends, so cap it (for example, four weeks of elapsed time and two contact attempts) and move on with documentation.

Where should decisions be stored?

Not in a side spreadsheet. Write the outcome back into the collection management system as structured values:

Field	Example value
rights_statement	http://rightsstatements.org/vocab/InC-OW-EU/1.0/
licence	CC BY 4.0 or none
clearance_evidence	link to search log / agreement
cleared_by / cleared_date	e.reed / 2025-03-19
review_due	2030-03-19

Using a RightsStatements.org URI makes the value machine-actionable for IIIF and aggregators, and the review_due field stops "cleared" statuses being treated as permanent.

Common errors and their fixes

Symptom	Likely root cause	Fix
Backlog never shrinks	No intake triage	Add cheap routing rules
Wrong PD verdicts	Creation vs publication confusion	Split date fields, set date_basis
Endless "unknown"	Diligent search has no end condition	Cap effort, promote to orphan
Decisions contradict catalogue	Rights live in a spreadsheet	Write back to CMS as a field
Reuser complaints	No takedown route published	Publish a notice-and-takedown policy

How do I make the workflow auditable?

Every clearance decision should be reproducible by a stranger six months later. That means logging the inputs (which sources you checked, what you found) alongside the output (the status). A one-line free-text "cleared — OK" note fails this test. Treat the evidence log as the deliverable and the status as a by-product of it.

Key Takeaways

• Backlogs grow because of missing triage, not lack of effort; route cheaply at intake.
• The deadliest data error is confusing creation with publication dates.
• "Unknown" must terminate: define a capped diligent search, then promote to orphan.
• Store decisions as machine-readable fields in the catalogue, not a side spreadsheet.
• Use RightsStatements.org URIs and a review_due date so "cleared" is never treated as forever.
• Log the evidence so any decision is auditable and reproducible later.

Frequently Asked Questions

Why does my clearance backlog keep growing even though staff are working hard?

Almost always because items default to 'unknown' and never move. Fix it by triaging into clear buckets (public domain, in-copyright cleared, orphan, blocked) at intake, so most items resolve quickly and only the genuinely hard cases queue for research.

What is the most common single error in clearance records?

Conflating the date of creation with the date of publication. Term calculations hinge on the correct event, and using creation dates for unpublished works can give wildly wrong public-domain answers.

How do I handle an item where the rights holder simply never replies?

Treat persistent non-response as an orphan-works case: document your diligent search, record the dates and channels you tried, and either apply the relevant orphan-works route or assign a risk-managed status with a takedown commitment.

Can I batch-clear a whole collection at once?

You can batch the easy strata (clearly public-domain by date, or items under a blanket deposit agreement) but never blanket-clear the whole thing. The mixed-rights middle always needs item-level review.

Where should the cleared rights decision actually live?

In the catalogue record as a structured, machine-readable field (ideally a RightsStatements.org URI plus a licence), not in a separate spreadsheet that drifts out of sync with the collection management system.

Related reading

• Orphan works
• RightsStatements.org
• Determine the public domain
• Write a takedown and notice policy
• Metadata quality audit