UK ONS Retail Sales Index - Reference Tables Data

It is the full monthly statistical release behind Britain's Retail Sales Index, produced by the Office for National Statistics and delivered by Datadory as rows instead of a workbook. Each edition is one 10.7 MB file carrying 28 worksheets: four CPSA tables for the value of retail sales at current prices and four matching KPSA tables for chained volume measures, both families seasonally adjusted with year-on-year, three-months-on-year, month-on-month and three-months-on-three-months percentage changes; twelve Table 1-4 variants across annual, quarterly and monthly frequencies; Tables 5-6 on a coarser commodity grouping; and the IDEF/ID1/ID2 identifier sheets that map every column to its series code.

The index measures retail trade in Great Britain only - England, Wales and Scotland - with Northern Ireland, the Isle of Man and the Channel Islands outside its perimeter by construction. It is compiled from the Monthly Business Survey - Retail Sales Index, in which roughly 5,000 retailers representing about 75.7% of known retail turnover respond; the average response rate was 61.9% in 2024. Figures publish around 20 days after the reference period ends: the 21 August 2026 edition carried July 2026, and the next was scheduled for 18 September 2026.

Real prints from the August 2026 edition, exactly as the series reads:

table   : CPSA # value of retail sales at current prices, seasonally adjusted, % change y/y
2026 May : All_Retailing_Incl_Fuel +3.4 | All_Retailing_Excl_Fuel +4.8
2026 Jun : All_Retailing_Incl_Fuel +3.8 | All_Retailing_Excl_Fuel +5.0
2026 Jul : All_Retailing_Incl_Fuel +1.6 | All_Retailing_Excl_Fuel +2.3

table   : CPSA # predominantly food stores (AGG1), % change y/y
2026 May : Predominantly_Food_Stores +1.3 | Non_specialised_Food +1.9 | Specialist_Food -6.1 | Alcohol_Tobacco -1.7
2026 Jun : Predominantly_Food_Stores +0.7 | Non_specialised_Food +1.0 | Specialist_Food -4.9 | Alcohol_Tobacco +7.1
2026 Jul : Predominantly_Food_Stores -0.6 | Non_specialised_Food -0.9 | Specialist_Food +0.2 | Alcohol_Tobacco +10.6

table   : CPSA # predominantly non-food stores (AGG12), % change y/y
2026 May : Predominantly_Non_food +7.0
2026 Jun : Predominantly_Non_food +6.5
2026 Jul : Predominantly_Non_food +3.8

# column header metadata riding with every table
Predominantly_Food_Stores: sales_2023 £201,285m | weight 37.57 | code AGG1 | identifier EAPT

Read the middle block once. In July 2026 the food story split in two: the predominantly food aggregate fell -0.6% year on year, supermarkets inside it fell -0.9%, but specialist food stores turned positive at +0.2% and the small alcohol-and-tobacco category jumped +10.6% after a -1.7% May. Meanwhile total retailing including fuel slowed sharply from +3.8% to +1.6%, and ex-fuel from +5.0% to +2.3%, with predominantly non-food cooling from +7.0% to +3.8%. Three months of prints, five divergent category stories - that separation is the reason the store-type breakdown exists.

Fifteen documented field groups, verified against the released structure during the research pass - nothing inferred from column headers alone. A time-period key anchors every row, formatted 2026 Jul for months, 2026 QN for quarters or plain years. The headline pair is all retailing including automotive fuel (AGG21) and excluding it (AGG21X), which makes the fuel contribution visible by subtraction. Food retailing splits three ways: predominantly food stores overall (AGG1, 2023 sales £201,285m, weight 37.57), supermarkets within it (SIC 47.11, weight 34.57) and specialist food stores (AGG26, weight 2.26), with alcoholic drinks, other beverages and tobacco (AGG27) broken out alongside.

The non-food side mirrors it: predominantly non-food stores (AGG12, weight 38.33) with non-specialised stores, textile clothing footwear and leather, household goods, other specialised non-food, and non-store retailing split into mail order and other non-store. Every column header row also publishes the metadata that usually has to be hunted down separately: 2023 sales values in pounds, AGG/SIC codes, percentage weights against the 2023=100 base, and stable dataset identifier codes (EAPT for predominantly food stores volume, for instance). In the non-seasonally adjusted tables those columns repeat three times over for All, Large and Small businesses.

Geography: Great Britain as a single national market - England, Wales and Scotland. There is no regional or store-level cut inside the index; anyone needing sub-national resolution should pair it with a different source rather than expect it here.

Temporal: rows run from 1988 to the latest reference month - July 2026 in the current edition, roughly 2,000 rows in the monthly tables alone. Annual, quarterly and monthly frequencies coexist across the worksheet set, and the growth tables publish four horizons: year-on-year, three-months-on-year, month-on-month and three-months-on-three-months. Two calendar facts matter for planning: figures land about 20 days after the reference month closes, and the release schedule is fixed in advance (the edition after 21 August 2026 was scheduled for 18 September 2026).

Granularity: monthly national cells by SIC 2007 division 47 store type, crossed with adjustment basis (seasonally adjusted and non-seasonally adjusted) and, in the unadjusted tables only, business size band (All / Large / Small).

One caution worth its own paragraph: the August 2025 release was delayed to September 2025 to correct a seasonal adjustment error affecting January-May 2025. Any historical extract should be confirmed as post-correction before it feeds a model.

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

You choose the channel and the cadence; the field dictionary above travels unchanged across all three. Full-history loads suit teams calibrating once against nearly four decades of British retail cycles. Store-type-scoped feeds suit dashboards that track grocery or non-food momentum month to month. Warehouse delivery suits analysts running UK consumption queries in SQL beside their own transaction tables. Cadence changes are a settings conversation, not a re-integration project.

1. Grocery demand planners and investors read supermarket volumes straight from the source instead of waiting for retailer trading statements - July 2026 showed the food block at -0.6% year on year with supermarkets at -0.9%, while specialist food stores turned positive at +0.2%.
2. Macro and rates strategists treat total retailing ex-fuel as the fastest official read on British consumption: +4.8% year on year in June 2026 decelerating to +2.3% in July is a signal that arrives weeks before national accounts.
3. Channel and category analysts watch migration between store types with matched definitions - alcohol and tobacco swinging from -1.7% to +10.6% in two months while non-food cooled from +7.0% to +3.8% is channel behaviour no survey panel reproduces.
4. Pricing and deflator desks pair current-price CPSA values with chained-volume KPSA measures to derive implied price effects per category.
5. Academics and journalists cite the official record, with the survey's own coverage statistics - 5,000 retailers, 75.7% of turnover, 61.9% response in 2024 - standing behind every figure.

Investors and quants convert monthly store-type indices into features for consumer-discretionary, grocery and retail-property models, backtested to 1988 with official revision discipline intact. Market researchers and consultants size UK categories using the government's own weights rather than vendor arithmetic. Data scientists and ML engineers load tidy period-by-measure rows keyed so the next monthly edition diffs cleanly, with adjustment basis labelled on every column. Developers building data products ship UK retail dashboards off one reconciled series vocabulary instead of hand-parsing 28 worksheets a month. Journalists, academics and students ground any claim about British shopping habits in the statistical record itself.

Within Datadory's catalogue, the neighbouring retail slices answer different questions. The Retail Industry slice of the same index isolates internet sales by store type - the where-it-was-bought view, where these reference tables are the what-the-sector-did view. The automotive fuel and non-store time series goes deep on fuel volumes and implied deflators, the residual this release folds into its headline. The Retail Sales Bulletin carries the narrative commentary over roughly 34 store types if you want prose attached to the numbers. And for a completely different altitude, the Federal Reserve Survey of Consumer Finances comparison sets household-level balance sheets against aggregate till receipts.

Four honest caveats. First, three column schemas live in one edition: Tables 1-4 use the store-type layout, Table 5 collapses to four commodity groups, and the CPSA/KPSA sheets run their own ten-column format - parsers must handle all three, which is precisely why the flattening exists. Second, adjustment basis matters: seasonally adjusted and non-seasonally adjusted series sit side by side, and differencing across that line produces noise that looks like signal; the business-size splits exist only in the unadjusted tables. Third, the 2025 correction: January-May 2025 figures were revised when the delayed September 2025 edition shipped, so extracts should always be confirmed post-correction. Fourth, geography: Great Britain only - Northern Ireland, the Isle of Man and the Channel Islands never appear, so UK-wide claims need a companion source.

Because the raw artifact is a 10.7 MB workbook built for economists reading tables, with three incompatible layouts, header metadata welded into merged cells and identifier codes living on separate sheets. Datadory flattens it into tidy period-by-measure rows, carries the AGG/SIC codes, weights and adjustment basis through as fields, keeps consecutive editions reconcilable in your warehouse, and cuts delivery to the store types and date range you name. Browse the rest of the industry on the Consumer Finance data hub, or everything the Office for National Statistics releases.