How the UK home & interiors benchmark is scored

The UK Home & Interiors Digital Growth Benchmark scores 75 UK home, furniture and interiors retailers on 54 outside-in criteria across six pillars, using 4,050 individual criterion observations. This page documents the rubric, the parity exclusions and the evidence rules, so any score in the league table can be checked against the method that produced it.

View the full league table

Methodology

Outside-in observable growth readiness, scored on one rubric.

This edition measures what is publicly observable about digital growth readiness. It makes no claim about internal profitability, CAC, LTV or conversion rate. Evidence was reviewed and frozen for the September 2026 edition.

Companies, one identical rubric

Criteria per company

4,050

Criterion observations

375

Parity exclusions across the cohort

Controlled-test parity

CV01, CV10, CV11, CV12, GI07 are uniformly excluded from scoring for all 75companies, because a controlled browser, checkout, performance and testing-stack pass was not consistently available across the full cohort. Excluding them for everyone prevents the original ten research subjects from carrying an unfair advantage. Excluded or missing evidence is omitted from the denominator rather than automatically scored poor, so a company is never penalised for evidence we could not observe.

Research parity

All 75 companies use the same 54-criterion rubric, scoring thresholds, evidence-status rules and criterion-level record structure. There is no 'first 10' quality tier in this edition.

Criterion universe

75 companies × 54 criteria = 4,050 criterion rows across Organic Discovery, AI Discovery, Conversion Experience, Commercial Proposition, Retention & CRM and Growth Infrastructure.

Controlled-test parity

CV01, CV10, CV11, CV12 and GI07 are uniformly excluded for all 75 companies because a controlled browser/checkout/performance/testing-stack pass was not available consistently across the full cohort. That removes 375 potentially biased comparisons.

AI Discovery

Every company is measured against the same 12 commercial and need-state prompts across ChatGPT, Gemini and Perplexity. Zero appearances are retained as zero visibility; no brands are selectively tested.

AI understanding accuracy

AI05 is excluded where a company received no recommendation appearance, because there was no output in which brand/product understanding could be judged.

SEO criteria

OD01, OD05 and OD10 use the same fixed public visibility/authority proxy for every company. They are directional outside-in measures; licensed share-of-voice data would strengthen a publication-grade edition.

Retention criteria

RT02, RT04 and RT05 are readiness criteria based on observable capture/account/review mechanics. The benchmark does not claim that an unobserved email, abandonment or post-purchase journey exists.

Not-applicable criteria

Store/local criteria are excluded from the denominator for genuine pureplays with no physical network; they are not scored as a weakness simply for having a different operating model.

Scoring

Assessable criteria are scored 0–2 and weighted within each pillar. Excluded criteria do not enter the denominator. Overall score is the equal-weight average of the six pillar scores.

Use

This is an outside-in digital growth-readiness benchmark, not a claim about internal profitability, CAC, LTV or actual conversion performance. Evidence was reviewed and frozen for the September 2026 edition.

Scoring, confidence and challenge

Each criterion is scored 0–2 against fixed thresholds, weighted into its pillar, and recorded with the source, evidence status, confidence level and method used.
Confidence is stated per observation. A directional public proxy is labelled as such and never presented as a measured internal metric.
Bands (Leading, Strong, Competitive, Developing, Exposed) are relative positions inside this cohort, not absolute quality judgements.
Anyone can submit public evidence for review. Scores change only after a re-observation pass under the same rubric, never silently.