UK Supermarket Basket Report — Tesco, Asda, Sainsbury's, Morrisons, Aldi and Lidl

Introduction

Category and national account managers at UK grocery suppliers who need their own lines and competitors' priced week by week across six retailers. Own-label and sourcing teams benchmarking tier-for-tier against the discounters. Pricing analysts tracking how much of a headline price is loyalty-gated. Researchers who need a basket series with a frozen, published basket definition rather than a shifting one.

This answers the September inquiry pattern: recurring UK grocery price monitoring on a fixed basket, isolating the loyalty-price versus standard-price gap as a distinct captured value, and own-label versus branded benchmarking at unit-price level.

What this report measures

For a fixed basket of everyday items, collected weekly and reported monthly as a continuous series, the report measures what six grocers displayed online: the standard displayed price; the loyalty-gated price where the retailer runs one (Tesco Clubcard Price, Sainsbury's Nectar Price) captured as a separate field that never overwrites the standard price; the retailer's published price-per-unit exactly as printed with its stated basis; pack size and unit; promotional mechanic and wording; availability; and whether the item is listed at all.

It derives basket cost by retailer on both a standard-price and a loyalty-price basis; month-on-month movement of each series; unit-price comparison using the published per-unit fields; the own-label versus branded gap within matched categories and tiers; the loyalty-price gap where both prices show; and promotional mechanics prevalence.

What this report does and does not measure. It measures observed online prices, published unit prices, loyalty-price gaps, promotional mechanic prevalence, availability rate and basket cost. It does not measure market share, grocery sales, volume share, shopper numbers, till-roll inflation or any official price statistic. This is not a consumer price index and must not be read as one: the basket is ours, fixed and published, and is not weighted to national consumption. "Cheapest basket" means the lowest total for our basket, at our collection times, on the stated price basis. It does not mean cheapest shop. Own-label comparisons across retailers are category-and-tier equivalents, not identical products, and are labelled so on every figure. We publish no estimate of any retailer's sales or share.

Collection design — written as a repeatable monthly build

Sample at a glance

These figures are the planned scope derived from the basket, the postcode sample and the weekly cadence described below — not an achieved count. Captured volume runs lower, because a line permanently delisted at a retailer records delisted rather than a price, a line carried only in another pack size drops out of pack-price comparison, and a failed pass is excluded rather than filled. What share of the basket each retailer actually lists and has available is a reported result availability rate by retailer, and by retailer × category, and every edition prints its own pass and observation count as-of date, week count and observation count.

Sources. The online grocery storefronts of Tesco, Asda, Sainsbury's, Morrisons, Aldi and Lidl. Where a retailer's online range differs from its store range, or covers only part of its estate, that is recorded and stated. Store-varying fields — availability, and any store-specific price the site exposes — are collected against 12 postcodes spanning London, a northern metropolitan area, a Midlands city, a Scottish city, a smaller market town and a rural area. National price fields are collected once per retailer per pass.

The frozen basket — the core discipline of this series. 75 items, defined once, versioned, and not changed between editions except by a documented version bump. A monthly series whose basket quietly changes is not a series: month-on-month movement would then be partly an artefact of redefinition, and the report would be worthless. The rules:

• The definition lives in a single versioned file, basket-uk-v{n}.csv, with a version number, effective-from date and a changelog entry for every change.
• Each line specifies category, tier (branded, own_label_value, own_label_standard, own_label_premium), brand or own-label line, variant, net quantity and unit, and a per-retailer match key.
• No line is added, removed or respecified mid-series. Where an item is permanently delisted at a retailer, the line stays and that retailer's cell becomes delisted — never substituted, because substitution silently changes the basket.
• When a bump is unavoidable — a line discontinued nationally, a pack-size change across all retailers — the report publishes both series in parallel for at least three editions so readers can see the join, with the changelog stating what changed and why.
• The definition is published in full every edition, with its version number on the cover.

Monthly build sequence — the operational runbook, identical every month. Weeks 1–4, fixed weekday and window: weekly pass across all six retailers and all basket lines for national fields, plus the postcode pass for store-varying fields. After each pass: automated validation of line coverage per retailer, match-type distribution, unit-price field presence, and price-change outlier flags, with any item moving beyond a set week-on-week threshold queued for human review before entering the series. End of month, days 1–2: aggregate the four or five passes into the edition; monthly basket cost is the mean of weekly basket costs, with weekly values also published so within-month movement is visible. Days 2–3: compute month-on-month against the prior edition on the same basket version only — if the version changed, suppress the MoM figure for affected lines and use the parallel-series presentation. Day 3: regenerate figures, fill slots, update changelog, set as-of date. Day 4: review against the pre-publish checklist, refresh the landing page in place, send the edition email.

Matching. Branded lines match on brand, variant and net quantity, human-verified at basket definition and re-audited at each bump. Own-label lines match on category, variant, net quantity and tier, and are explicitly category equivalents. Where a retailer carries only a different pack size of a branded line, the observation records size_variant and is compared only on published unit price, never pack price. Multipacks and promotional bundles are excluded from basket totals and counted separately.

Loyalty price as a distinct captured price. Clubcard and Nectar prices sit in their own fields with their own flag. Basket cost publishes twice — standard basis and loyalty basis — and the two are never blended into one headline. Where a retailer runs no loyalty price, its loyalty basket equals its standard basket and the report says so explicitly rather than letting readers assume. The loyalty gap is a first-class finding, because the share of a basket that is loyalty-gated is among the most commercially significant things this series observes.

Data schema

Report sections

This month in one page. Standing opener, same shape every edition: basket cost on both bases, month-on-month movement, and the month's one structural finding.

Figure 1 — Basket cost by retailer, standard and loyalty basis. Grouped bars; x = retailer, y = basket cost, series = standard vs loyalty basis. Items × one measure across two series → bars. Matched-line count on the figure.

Figure 2 — Basket cost series over time. Line; x = edition month, y = basket cost, series = retailer. Level over time → line. The series is the point of this report, so this carries the full history available at the same basket version, with any version join marked on the axis.

Figure 3 — Month-on-month movement. Horizontal bars; x = change in basket cost, y = retailer. One measure across six items → bars. Suppressed for lines affected by a version change, with the suppression stated.

Figure 4 — Within-month weekly movement. Line; x = collection week, y = basket cost, series = retailer.

Figure 5 — Loyalty-price gap. Point plus whisker; x = retailer running a scheme, y = gap between standard and loyalty price, point = median, whisker = interquartile range. A distribution.

Figure 6 — Share of the basket that is loyalty-gated. Stacked bars; x = retailer, y = share of basket lines, segments = loyalty price shown / standard only / other promotion / unavailable. Parts of a whole that sum → stacked.

Figure 7 — Unit-price comparison, selected categories. Horizontal bars; x = published price per unit, y = basket line, series = retailer. Uses retailers' own published fields with the basis printed, stating where a retailer published none.

Figure 8 — Own-label versus branded gap. Grouped bars; x = category, y = own-label price as % of matched branded line, series = retailer. Category-and-tier-equivalent caveat on the figure.

Figure 9 — Own-label tier ladder. Grouped bars; x = retailer, y = price per unit, series = tier including branded. Shows each retailer's internal price architecture.

Figure 10 — Promotional mechanics prevalence. Stacked bars; x = retailer, y = share of basket lines on promotion, segments = mechanic. Parts of a whole → stacked.

Figure 11 — Availability rate. Grouped bars; x = category, y = availability rate, series = retailer.

The discounter comparison. Aldi and Lidl against the big four on a stated matched-line subset, with the coverage constraint given explicitly rather than buried.

Basket integrity note. Standing section: version in force, changes since last edition, delisted lines by retailer, excluded lines, and the matched-line count behind every basket figure.

How to use this. The loyalty gap as a commercial signal; using published unit price rather than pack price for cross-retailer benchmarks; why month-on-month is only valid within one basket version.

Methodology and limitations.

Appendices. A: full basket definition, version basket version in force. B: postcode sample. C: match-type counts, delisted lines, collection success rates. D: basket changelog, all versions.

Data slots

Methodology and limitations

Sampling limits. Six retailers, one 75-item basket, online storefronts, four or five weekly passes a month, and 12 postcodes for store-varying fields. A defined sample, not a census: retailers list tens of thousands of lines and our basket is a fixed window onto a small part of that. Figures read as "for our basket" throughout.

Not an inflation measure. Our basket is unweighted by consumption and self-selected. Movement in our basket cost is movement in our basket cost. It is not CPI and not a national grocery inflation rate, and the report states this on the page where the month-on-month figure appears, not only here.

Online versus in-store. We observe online prices, which are not guaranteed identical to shelf prices at any retailer, and the discounters' online ranges are narrower than their store ranges. Where that constrains a comparison, the constraint prints beside the figure with the matched-line count.

Loyalty prices. Captured as displayed on the site. We do not verify redemption, do not assume enrolment, and never blend loyalty and standard prices into one basket. A loyalty-basis basket is the cost to a shopper who is enrolled and for whom every displayed loyalty price applies — an upper bound on the benefit, not a typical outcome.

Own-label comparison. Own-label lines across retailers are different products with different formulations and often different pack sizes. Tier assignment uses the retailer's own range naming where it supports one and is marked unknown where it does not. These are category-and-tier equivalents, labelled on every figure, and must not be quoted as like-for-like.

Published unit price. We use the retailer's own published per-unit field rather than recomputing, because recomputation would hide inconsistencies in how retailers define the unit and would stop our figure matching what the shopper sees. The cost is that bases differ between retailers and some lines publish nothing; we normalise the basis for charting, print the original, and publish the null rate.

Basket integrity. The largest risk to a monthly series is silent basket drift. Controls: a versioned definition file, a changelog, no mid-series substitution, delisted rather than replacement, suppression of month-on-month across a version boundary, and parallel publication of both baskets for at least three editions at any bump. Every basket figure prints its matched-line count so no reader compares baskets of different sizes.

Outlier control. Any week-on-week item move beyond a set threshold is flagged and human-reviewed before entering the series, catching page-structure changes and mis-parses. Reviewed-and-rejected observations are excluded and counted, not silently dropped.

Observed versus inferred. Standard price, loyalty price, was-price, published unit price and basis, promotional wording, availability and pack size are observed. Basket costs, month-on-month movement, loyalty gaps, tier gaps, mechanic prevalence and availability rates are derived by stated arithmetic. Tier assignment is a judgement, flagged as such. Nothing is modelled or estimated, and no share, sales or inflation figure appears.

Missing data. Failed passes are reported, never interpolated. A month with fewer than four usable passes for a retailer publishes its pass count and flags its month-on-month figure as low-confidence.