BARB Weekly Top 50 Shows and Viewing Data

It is the flagship dashboard of the Barb audience currency - Broadcasters' Audience Research Board, the board behind the numbers Britain quotes about its own television. Every week it ranks the fifty most-watched programmes across all UK broadcast channels plus Barb-subscribing streaming services. Within Datadory's fifteen-dataset broadcasting slice it is the deep-dive record: where its sibling BARB Main Site - Viewing Data Hub is the umbrella over four products, this one drills a single ranking to programme-instance grain and scores 9 out of 10 on Datadory's quality rubric with a fully verified ten-field dictionary.

Two methodological facts make the rows unusually clean. First, since January 2026 the ranking runs on an As Viewed basis, measuring viewing wherever it happened - TV-owning and non-TV households alike - rather than the older schedule-driven As Broadcast method that counted from transmission logs. Second, audiences arrive as consolidated figures: seven-day (C7) or twenty-eight-day (C28) windows that sweep up catch-up and playback, optionally including pre-broadcast viewing up to thirty-five days before transmission and viewing on PC, tablet and smartphone alongside the TV set. Streaming titles are measured differently and flagged; split transmissions report only their highest occurrence per channel; anything shorter than five minutes is excluded. Get a sample of this dataset cut to your channels, weeks and programmes.

One row per transmitted programme instance per reporting week, already ranked into the fifty. Three straight from broadcast week 202631:

# delivered grain: one row per programme per broadcast week

week                    : 202631            # YYYYWW reporting period
Broadcaster             : BBC
Programme Title         : Ann Droid: Series 1, Episode 3
mbid                    : s86p2q            # Barb's id for this transmission
seven_day_value_all     : 5051.700          # thousands of viewers, C7 consolidated
twenty_eight_day_all    : 0.000             # C28 component, populated when consolidation widens
devices.Total           : 5051.700          # across selected device components
is_svod                 : false             # true marks a streamer average, not a broadcast figure

# same week, next two rows down the ranking
The Great British Sewing Bee: Series 12, Episode 3   BBC   s8zwzw   4248.800
Coronation Street: Series 67, Episode 147            ITV   s8t4zw   3679.200

Read the top row as three facts already decided for you. The audience arrives in thousands and consolidated, so "five million" is what the number says without division or window math. The title carries series and episode numbering natively, which makes season-level aggregation a prefix operation rather than a parsing project. And mbid - Barb's own identifier for this exact transmission - turns "join my title database to the weekly ranking" from fuzzy string matching into a keyed lookup. The is_svod flag then keeps the honest split visible: linear rows are consolidations, streaming rows are first-window averages, and any chart mixing them knows which is which before a footnote is read.

Ten fields define the row, every definition verified against live output during the August 2026 research pass, which is what earns the 9/10 quality score - only 145 of the 1,744 datasets in Datadory's catalog hold a nine or better. They divide into three jobs:

• Identity: week, Broadcaster, Programme Title and mbid say who aired what, when, with the join key carried natively rather than inferred from titles.
• Measurement: seven_day_value_all and twenty_eight_day_value_all hold the consolidated audiences in thousands, with devices.7D, devices.28D and devices.Total splitting the total into its consolidation-window components so a C7-only analysis never has to recompute one.
• Classification: is_svod separates streaming averages from consolidated broadcast figures at filter time.

Filter dimensions travel beside the fields rather than inside them: broadcaster-group versus individual channel selections, consolidation window (seven or twenty-eight days), pre-transmission inclusion and device-type scope all parameterise the same row shape, which keeps schemas stable while views change. The full dictionary follows in tabular form; anything not pinned below folds under additional fields on request.

• Geography - the United Kingdom as one national panel. There is no sub-national split in the currency itself, which is exactly why a figure from this ranking settles an argument about British television. For station-level structures in other markets, pair it with a registry product such as FCC TV Query System (LMS / CDBS TV Station Database).
• Temporal - current broadcast weeks from January 2026 onward on the As Viewed basis. A reported week lands once its consolidation has settled, roughly eight days after the week closes, and stays revisable as broadcasters amend transmission logs afterwards. Pre-2026 periods live in the sibling hub's archive back to 1992, so today's ranking can be set against nearly three decades of context.
• Granularity - one row per transmitted programme instance per week, capped at the top fifty by audience: the finest regular cut UK audience measurement publishes. Roughly 2,600 rows a year of current data, each carrying its period explicitly so incremental pulls key cleanly.

Set against the wider catalog - 1,744 datasets averaging 7.81, this broadcasting slice averaging 7.73 across fifteen products - the 9/10 reflects verified field documentation and a sample proven against live rows. Where it ranks: best broadcasting datasets.

API, files, or your warehouse. Daily, weekly, or hourly.

You pick the channel and cadence; period parsing, consolidation labelling and SVOD flagging stay our problem. Rows arrive normalised with the reporting week explicit on every record, so a delta pull appends without double-counting and a backfill joins history on period and channel rather than guesswork.

Hourly suits launch weeks, where a new series' trajectory moves fast enough to matter by afternoon. Daily suits ad-sales desks pricing against the freshest reported figures. Weekly suits trend work, because genre curves move slowly enough that snapshots lose nothing. Whichever you choose, the ten-field dictionary travels unchanged, and revisions to earlier weeks ride the same keys they were first delivered under. A sample ships first either way - real rows for your channels and weeks before any commitment.

Six jobs the weekly ranking settles outright:

1. Ad-sales pricing and yield - price spots against what comparable programmes actually delivered last week, across every rival channel at once, instead of against last year's averages.
2. Commissioning and scheduling - track genre performance week over week; a format's decay curve becomes visible in three or four rankings rather than a season-end post-mortem.
3. Competitive slate tracking - watch rivals' launches land in the fifty within days; see competitor tracking.
4. Media-market sizing and share analysis - anchor channel-share claims in the national currency itself; see market sizing.
5. Citation-grade verification - the number a headline quotes resolves to a specific programme, week and consolidation window; see citation grade research.
6. Streaming-versus-linear benchmarking - the SVOD flag puts platform averages and broadcast consolidations in one frame, the comparison the whole industry argues about.

Each job maps to a persona below, and the sample validates whichever one you came for.

Market Researchers & Consultants lead fit: the UK's own ratings currency is the reference class their media-market models want, and this record is its sharpest weekly edge - see broadcasting data for market researchers. Journalists, Academics & Students get citation-grade answers to "what did Britain watch", with the sibling hub's archive supplying the historical curve behind any headline - see broadcasting data for journalists, academics and students. Competitive Intelligence & Product Teams track rivals' hits and misses in near-real time; see broadcasting data for competitive intel product teams. Investors & Quant Researchers reading the advertising market find the demand signal here too. Persona fit has edges worth naming: sales-growth teams find no prospecting signal in audience rows, and e-commerce operators read them only indirectly as ad-market context - this measures watchers, not buyers. Persona-by-persona detail lives on the broadcasting data hub.

Three notes worth having upfront.

First, consolidation windows change the number, legitimately. The same programme can carry different audiences under C7 and C28, and pre-transmission inclusion adds another axis; pick the window your comparison needs before quoting, and keep it constant across a trend line.

Second, streaming titles are averages, not consolidations - the average audience over a title's first days of availability on TV sets, flagged is_svod. Mixing them with broadcast figures in one column without the flag produces exactly the wrong kind of chart.

Third, the ranking is a cap, not a census: fifty programmes a week means long-tail and minority-audience programming sits below the cut, and the sibling hub's archive products cover the periods before January 2026. Start there if your question reaches further back than the current basis. Then Get a sample of this dataset.

Because the raw artifact underneath is a dashboard built for one person reading one week in a browser, and every interesting question wants many weeks, many channels and both consolidation windows in one frame. Datadory hands over the corpus itself: the ten-field dictionary kept verified, the mbid join key delivered as a column so title databases link without string surgery, SVOD flags resolved at row level, and the result landed on the cadence your pipeline runs - with sample rows in hand before any commitment. Browse the rest of the slice on the broadcasting data hub, the best broadcasting datasets ranking, or see how it squares against a US registry in the BARB Weekly Top 50 vs FCC TV Query System comparison, or read the source profile at Barb - Broadcasters' Audience Research Board.