← All writing

Europeana's quality tiers, explained: what contentTier and metadataTier actually mean

· 6 min read

If you’ve ever browsed the Europeana API response for a record, you’ve probably noticed two fields sitting quietly in the europeanaAggregation block:

"contentTier": "4",
"metadataTier": "B"

The Europeana documentation covers these across several scattered Confluence pages. This post brings it together in one place: what the tiers mean, how they’re calculated, and what a given combination tells you about a record’s quality.

Two independent scales

Content tier and metadata tier measure completely different things and are calculated independently.

Content tier is about the digital object: how good is the media file, and under what terms can it be reused? Scale: 0, 1, 2, 3, 4. Higher is better.

Metadata tier is about the descriptive record: how rich, structured, and multilingual is the metadata? Scale: 0, A, B, C. Later letter is better; C is the top.

A record can have any combination — contentTier 4 / metadataTier A (great media, thin metadata) or contentTier 1 / metadataTier C (low-res thumbnail with an exceptionally rich description) are both valid.

Content tiers: resolution × rights

The content tier algorithm has two inputs: media resolution and rights statement. Importantly, rights only enter the calculation once the media clears the resolution threshold for tier 3. Below that, the rights statement is ignored.

Tier Meaning Image resolution Rights
0 Does not meet publishing criteria — —
1 Minimum viable — thumbnail only ≥ 0.1 MP any
2 Usable in thematic collections ≥ 0.42 MP (~800×533 px) any
3 Reusable by individuals, educators, researchers ≥ 0.95 MP (~1200×800 px) restricted-reuse rights
4 Free reuse platform — top tier ≥ 0.95 MP open / public domain rights

The resolution thresholds shown are for images. Audio, video, 3D, and text types have their own equivalent criteria.

What “rights” means here

Tier 3 requires one of seven reuse-permitting statements:

  • Creative Commons: CC BY-NC, CC BY-ND, CC BY-NC-SA, CC BY-NC-ND
  • RightsStatements.org: NoC-NC, NoC-OKLR, InC-EDU

Tier 4 requires the fully open ones: CC0, Public Domain Mark, CC BY, CC BY-SA.

The logic is intentional: Europeana’s tier 4 is explicitly positioned as a free reuse platform. A high-resolution image under an all-rights-reserved statement can’t sit at tier 4 regardless of quality, because reuse isn’t permitted.

Reading a content tier

  • Tier 4: high-resolution media, open rights. The institution is sharing freely.
  • Tier 3: high-resolution, but with some reuse restrictions. Still very usable.
  • Tier 2: decent resolution, good for browsing and display, not necessarily for download/reuse.
  • Tier 1: thumbnail quality. Useful as a placeholder, not much more.
  • Tier 0: doesn’t meet the bar. May be a broken link, or metadata-only with no media.

Metadata tiers: language × elements × context

The metadata tier is more nuanced. It’s scored on three criteria, and the overall tier is the minimum across all three — a record only reaches tier C if all three criteria independently clear tier C. One weak criterion drags the whole score down.

The three criteria

1. Language tagging. How much of the metadata carries xml:lang attributes? Language-tagged values enable Europeana to display metadata in the user’s preferred language and improve multilingual search. Tier C requires ≥ 75% of relevant elements to carry at least one language tag.

2. Enabling elements. Beyond the mandatory EDM fields, there’s a defined list of enabling elements — optional fields that unlock specific discovery and display scenarios (faceted filtering, timeline browsing, map view, etc.). Tier C requires at least 3 distinct enabling elements covering at least 2 distinct user scenarios.

3. Contextual classes. This is the big one. EDM supports four contextual entity types that turn a flat record into a knowledge graph node:

  • edm:Agent — people and organisations (artists, publishers, workshops)
  • edm:Place — geographic locations with coordinates
  • edm:TimeSpan — time periods with structured begin/end dates
  • skos:Concept — subject classifications

These show up in the record JSON as the agents[], places[], timespans[], and concepts[] arrays. A contextual class counts toward the tier when it’s fully populated with minimum required fields or when it carries an owl:sameAs link to a supported LOD vocabulary (Wikidata, Getty, GeoNames, etc.).

Tier B needs ≥ 1 such contextual class. Tier C needs ≥ 2.

The tier table

Tier Language tags Enabling elements Contextual classes
0 below A — —
A baseline minimal none required
B higher coverage ≥ 1 discovery scenario ≥ 1 class populated or LOD-linked
C ≥ 75% of relevant fields ≥ 3 elements, ≥ 2 scenarios ≥ 2 classes populated or LOD-linked

Reading a real record: contentTier 4 / metadataTier B

This is a common combination for well-digitised collection items from larger institutions.

contentTier 4 tells you: this institution is sharing a high-resolution image under an open licence. You can download and reuse it.

metadataTier B tells you: the record is solid — it has language-tagged metadata and at least one contextual class (most likely a skos:Concept for subject classification or an edm:Place for location). But it’s missing something to reach C: either the language coverage doesn’t hit 75%, or there’s only one contextual entity rather than two, or the enabling elements don’t span two distinct discovery scenarios.

What B → C would take

Looking at the structure of a metadataTier B record, the most common gaps are:

  • The agents[] block is absent (no linked author/creator entity). Adding a well-populated agent with owl:sameAs pointing to a Wikidata URI would add a second contextual class.
  • Language tags are present on titles and descriptions but missing from subject terms or type labels.
  • Only one of the three contextual class types is populated.

The path to C runs through linked open data: connecting the record’s entities to external vocabularies does double duty — it satisfies the contextual class requirement and tends to pull in better language coverage via the linked vocabulary labels.

Why the tiers matter

The tiers aren’t just labels — Europeana uses them for:

  • Filtering and ranking in search results. Higher-tier records are more discoverable.
  • Feature eligibility: tier 4 content can appear in Europeana Stories, exhibits, and reuse campaigns in ways that lower-tier content cannot.
  • Feedback to institutions: the tier system is designed so data providers can see exactly which axis is limiting their score and what to fix.

If you’re building on the Europeana API — ingesting records, analysing collections, or building derivative applications — the tiers are a ready-made quality signal you don’t have to compute yourself.

Sources

EuropeanaMetadataLinked Open DataCultural Heritage

← All writing