Back catalogue is the body of work a rights holder has already released and continues to distribute: podcast episodes from three years ago, television seasons that finished their network run, albums past their promotional cycle, articles still sitting on a publisher's domain. In advertising the term describes a supply source rather than an archive. Those files keep being played, and every play is a sellable impression, traded years after the budget that launched the work was spent.
The category exists because distribution costs collapsed while production costs did not. A drama series once earned money during its broadcast run and one syndication window. The same series now sits on cloud storage, gets scheduled into a free streaming channel, and earns whatever its ad breaks earn, indefinitely.
Where the line is drawn
No cross-industry threshold separates back catalogue from current release. Each sector sets its own, and the definitions conflict even inside single companies.
Recorded music is the most quantified. Luminate, the measurement firm behind the Billboard charts, defines current consumption as anything played within 18 months of release, catalogue as anything older, and deep catalogue as material beyond 60 months. Its own CONNECT platform glossary uses different bands: current at 0 to 36 months, catalogue at 36 to 72, deep catalogue past 72. Both are published by the same vendor, and figures built on them are not comparable. On the 18-month basis, catalogue represented 73.3% of total US album consumption in 2024, up from 72.7% a year earlier. In 2025, according to Luminate's year-end report, current music fell 1.6% by volume.
Film and television draw the line through accounting rather than the calendar. Lions Gate's 2014 annual report defines acquired libraries as titles released three years prior to the acquisition date, amortised against expected revenue over up to 20 years, against roughly three for newly completed films. Netflix made the same distinction a decade earlier, when it was still shipping discs. Its 2004 annual report records a change of estimate effective 1 July 2004, moving back-catalogue titles from a one-year accelerated amortisation schedule to three years, after the company concluded those titles had a significantly longer productive life than assumed. New releases stayed on one year. The revision cut cost of revenues by $10.9 million and added $0.17 to diluted earnings per share.
Nielsen avoids age entirely, splitting streaming titles into originals and acquired content by provenance. Podcasting has no formal threshold at all: back catalogue means every episode except the newest.
How a back catalogue becomes inventory
The conversion mechanism differs by medium, but the principle is constant. Something has to decouple the advertisement from the content file.
In audio that mechanism is dynamic ad insertion. Episodes carry markers at pre-roll, mid-roll and post-roll positions; the hosting platform or ad server stitches creative in at the moment of the request. NBC Sports adopted the Triton Ad Platform in 2026 for exactly this reason, a configuration that turns an entire back catalogue of sports episodes into live inventory carrying current campaigns rather than the sponsorships recorded at production. Libsyn's April 2026 exclusive hosting deals made the same argument, opening older episodes to brands as new downloads occur, and Audioboom used it as a rationale for acquiring Adelicious in 2025. According to the IAB, dynamic insertion accounted for more than 90% of United States podcast advertising revenue by October 2023, roughly double its share three years earlier.
Baked-in advertising works the opposite way, and the contrast defines the commercial argument. A recorded read ships with the audio permanently and accumulates impressions across the back catalogue with no expiry, no targeting and no way to retire the campaign short of re-editing the file.
Video moves through scheduling. A free ad-supported streaming channel begins as a playlist: playout software sequences files from a content library into a continuous feed, encodes it into HLS segments and publishes an XMLTV guide for aggregators to poll. That category exists because television libraries built for syndication lost their traditional buyers as cable subscriptions fell, and connected televisions offered a distribution route costing almost nothing to run. TriCoast Media built a supply-side and demand-side platform around a 5,000-title film and television catalogue it already owned. On the open web the mechanism is the ad server: an archived article from 2018 is trafficked exactly like one published this morning.
Origin and evolution
The phrase came from the record and book trades, where a catalogue was the list of titles a publisher kept in print and the back catalogue was everything not currently being promoted. Merriam-Webster's entry records the television usage arriving by 1995, citing a Hollywood Reporter piece describing programming drawn primarily from a company's back catalogue plus acquisitions.
Podcasting reached the same problem two decades later. The first season of Serial, released in 2014, carried host reads recorded into the audio, which forced sellers to guess demand before publication. Nieman Lab, reporting in September 2016, estimated that unlocking back catalogues could multiply available inventory several times over. Insertion technology did exactly that, and the resale of old episodes became the default. Nielsen counted more than 1.1 million video titles across linear and streaming by January 2024.
Why the category matters to buyers and sellers
Back catalogue is where much of streaming attention sits, which makes it a large addressable pool in connected television. In July 2023, Nielsen recorded acquired titles outpacing new originals, with Suits alone setting a record for an acquired series at almost 18 billion viewing minutes. Analysis of Nielsen rankers between January 2025 and February 2026, published by Forbes in March 2026, calculated that a combined Paramount and Warner Bros. Discovery would control 40% of the most-watched acquired series on streaming.
The pricing is the point. Library rights cost a fraction of commissioning, and the inventory they generate is continuous rather than concentrated around a premiere. Fremantle's multi-territory deal placing eleven Baywatch seasons on Prime Video illustrates the trade: several hundred hours of hour-long procedural scheduled against a service reporting 315 million global ad-supported viewers in the fourth quarter of 2025 and $19.8 billion in quarterly advertising revenue.
Attention data qualifies the enthusiasm. TVision's State of Streaming report, published in August 2025, measured 64% attention for original streaming content against 59% for library content, an 8.5% difference. Cheaper inventory performs slightly worse per impression.
Identification, measurement and the metadata problem
A buyer cannot value a back catalogue impression without knowing what the title is, and streaming supply has historically not carried that signal. Gracenote, Nielsen's content intelligence unit, covers more than 50 million titles across 80,000 channels and catalogues, and its persistent identifiers let a series be tracked consistently wherever it runs. Content Connect opened that metadata to agencies, brands, SSPs and DSPs on 4 December 2025. Xumo applied Gracenote and IRIS.TV signals across roughly 2,000 FAST channels on 22 June 2026, and Tubi Media Group agreed on 12 August 2026 to test Gracenote identifiers in programmatic bid streams. Peer39 analysis in March 2026 found that only 40% of connected television bid requests carried usable programme-level signals.
Audio has a different gap. Podcast delivery is measured from server logs rather than client beacons, so a baked-in advertisement inside an old episode has no separate request and its delivery is inferred from how much of the file was downloaded.
Limitations and disputes
Supply is expanding faster than demand can absorb it. DataBeat recorded connected television CPMs down 25.8% year on year in April 2026 while display and mobile rose, a pattern consistent with library-fed channel counts growing ahead of budgets.
Rights are the sharper exposure. Sony Music and nine affiliated labels sued Kroger on 21 August 2026 over at least 392 unauthorised uses of recordings in social advertising, an exposure ceiling of $58.8 million at the statutory maximum, having already granted the retailer fourteen separate licences between 2017 and 2025. The same theory was applied to DSW in August 2025. A back catalogue is an asset to its owner and a liability to anyone using it without clearance.
Text archives face a version of the same fight. Common Crawl's repository, which contains millions of paywalled articlesfrom major news organisations, supplies training data to AI developers, and the News/Media Alliance demanded removal rights and terms changes in 2026. Search behaviour cuts the other way: SISTRIX found in April 2026 that only 1.4% of news articles cited in AI responses remain in the citation set permanently, with evergreen material surviving better than news.
Disambiguation
Product catalogue means a merchant's live inventory feed, the structured file submitted to Merchant Center that supplies imagery and attributes to Demand Gen and Performance Max campaigns. It describes what is for sale now.
Library and acquired content are the video industry's near-synonyms, and they are not identical. Library covers everything a service can schedule, including its own older originals. Acquired describes provenance, content licensed from another owner regardless of age.
Evergreen content is a quality, not a date. An explainer published in 2015 that still answers a live question is evergreen and back catalogue at once; a match report from the same week is only the latter.
Remnant describes inventory that went unsold at auction. Back catalogue can sell at a premium or not at all.
Recent developments
Discovery mechanics are now being pointed at older material directly. Spotify's Prompted Playlist extension in April 2026 surfaces back catalogue episodes in response to a stated interest rather than recency. Gracenote's Q3 2026 Data Hub analysis, published on 20 August 2026, counted 2,172 FAST channels globally, and The Trade Desk became the first demand-side platform to take Gracenote show-level data directly in August 2026.
Timeline
- 1995: Trade press usage of "back catalogue" for television programming is recorded, in a Hollywood Reporter piece cited by Merriam-Webster
- 1 July 2004: Netflix extends back-catalogue DVD amortisation from one year to three, adding $10.9 million to that year's results
- 2014: The first season of Serial ships with host reads recorded into the audio
- September 2016: Nieman Lab estimates that unlocking podcast back catalogues could multiply available inventory several times over
- July 2023: Nielsen records acquired titles outpacing new originals, with Suits reaching almost 18 billion viewing minutes
- October 2023: The IAB puts dynamic ad insertion above 90% of United States podcast advertising revenue
- January 2024: Nielsen counts more than 1.1 million video titles across linear and streaming
- August 2025: TVision measures 64% attention for streaming originals against 59% for library content
- 4 December 2025: Gracenote launches Content Connect, opening programme-level metadata to agencies, brands, SSPs and DSPs
- 6 January 2026: Index Exchange and Gracenote introduce show-level transparency through Spectrum Reach
- March 2026: Peer39 finds 40% of connected television bid requests carry usable programme-level signals
- 22 June 2026: Xumo deploys Gracenote and IRIS.TV signals across roughly 2,000 FAST channels
- 12 August 2026: Tubi Media Group agrees to test Gracenote identifiers in programmatic bid streams
- 20 August 2026: Gracenote counts 2,172 FAST channels globally in its Q3 2026 Data Hub analysis
- 21 August 2026: Sony Music and nine affiliated labels sue Kroger over at least 392 unauthorised recording uses
Related PPC Land coverage
- Explaining baked-in - The permanent-insertion format whose economics turn on back-catalogue downloads, with the Nieman Lab estimate and the IAB share figures.
- Explaining host read - How a recorded read accumulates impressions across an archive with no expiry and no targeting.
- NBC Sports picks Triton Digital for podcast infrastructure - The ad-serving configuration that converts an audio archive into live inventory.
- Libsyn signs 5 exclusive podcasts as ad network inventory grows - Dynamic insertion framed explicitly as a way to make back catalogues commercially active.
- Audioboom acquires UK podcast leader Adelicious - Continuous monetisation of back-catalogue audio as an acquisition rationale.
- Explaining FAST - The channel format built from content libraries, with playout mechanics and April 2026 CPM movement.
- TriCoast Media bets on content flywheel to cut CTV ad costs - A 5,000-title catalogue used as the foundation for an advertising stack.
- Prime Video gains 11 Baywatch seasons in Fremantle multi-territory deal - Library depth treated as an inventory calculation rather than a programming story.
- TVision finds 9 of 200 daily viewing minutes are attentive ad time - Attention rates separating original streaming content from library content.
- Xumo brings contextual signals to linear streaming with Gracenote and IRIS.TV - Programme-level and frame-level metadata applied across a 2,000-channel footprint.
- Tubi and FOX One gain Gracenote metadata on 55 million titles - The August 2026 agreement covering discovery and a bid-stream identifier test.
- The Trade Desk gains show-level CTV data in Gracenote's first DSP deal - Title-level signals reaching the demand side, with the deployment chronology.
- Gracenote finds FAST sports games gain 37.5% as channels rise 13.8% - Q3 2026 channel counts and the argument for measuring volume rather than channels.
- Sony Music sues Kroger over 392 songs in $58.8M copyright claim - Recorded-music catalogue enforcement against a retailer's social advertising.
- Sony Music sues DSW for copyright infringement in social media posts - The earlier application of the same sync-licence theory.
- Common Crawl supplies paywalled content to AI companies despite publisher objections - How text archives entered training datasets without licensing.
- News publishers target Common Crawl, the AI training data backdoor - The demand letter seeking removal rights and terms changes.
- SISTRIX April 2026: AI citation drift and the death of keyword research - Persistence rates showing evergreen material outlasting news in AI citation sets.
- Spotify's Prompted Playlist adds podcasts for Premium users in seven markets - Prompt-driven discovery surfacing older episodes.
- Demand Gen feed use hits 35% as PMax advertisers face Q4 overlap - The product-catalogue sense of the word, and what a Merchant Center feed unlocks.
Summary
Who. Rights holders and their distribution arms create back catalogue supply: studios and production companies such as Fremantle and Lions Gate, podcast networks and hosting platforms including Libsyn, Triton Digital and Audioboom, record labels, and web publishers. Aggregators such as Prime Video, Tubi, Xumo and Roku turn it into channels. Gracenote, IRIS.TV, Nielsen and Luminate supply the identification and measurement layer buyers need to value it.
What. Back catalogue is previously released content that continues to be distributed and monetised after its launch window. In advertising it functions as an inventory class: impressions generated by material acquired or produced years earlier, usually cheaper to supply than new commissioning and traded through the same programmatic pipes.
When. The term entered television trade usage by the mid-1990s, acquired formal accounting treatment in the 2000s, and became an advertising category proper once dynamic ad insertion made publication date irrelevant to campaign delivery. The IAB placed dynamic insertion above 90% of United States podcast advertising revenue by October 2023.
Where. In free ad-supported streaming schedules, on-demand libraries, podcast feeds, music streaming services and publisher archives, and inside the ad servers, hosting platforms and metadata systems that attach a current campaign to an old file.
Why. Content that has already been paid for keeps generating consumption, and every consumption event can carry an advertisement. That arithmetic funds free streaming, sustains podcast networks between releases and gives buyers a large, identifiable and comparatively cheap pool of inventory. The unresolved parts are definitional inconsistency, the metadata gap that stops buyers seeing what they are bidding on, and the licensing exposure that follows any use of somebody else's catalogue.
Discussion