Search Readiness and Direct-Booking Infrastructure on Independent Restaurant Websites in the United Kingdom: A Cross-Sectional Audit of 1,000 Sites

Abstract

Background. Independent restaurants depend on being found and chosen online, yet little measured evidence exists on how their own websites present the information that search engines and diners use. Methods. We drew a seeded random sample of 1,000 independent UK restaurants with a listed website from OpenStreetMap-derived records covering 106 UK areas, excluding chains. Each homepage was audited on 11 October 2026 with a 16-check instrument covering crawlability, machine-readable description and diner actions, reading the HTML as delivered. On 12 October 2026 the 755 homepages that loaded were re-read at field level, one menu link was followed, and each page was rendered in a headless browser at phone width. Shares are reported with 95% Wilson intervals; subgroup differences were tested with Fisher's exact and chi-squared tests with Holm adjustment. Results. 17.9% of sampled domains no longer resolved, could not be reached or returned an error. Among loaded homepages, generic web hygiene was close to web-wide benchmarks (meta description present on 69.0%, mobile viewport on 95.4%), but the restaurant-specific layer was largely absent: 77.1% (95% CI 74.0–79.9) carried no usable Restaurant or LocalBusiness structured data, and among the 212 homepages that did, 26.9% declared a menu and 36.3% a cuisine. 21.9% linked to no menu and 10.4% linked directly to a PDF; of 471 menu web pages, 32.5% displayed at least three prices as text. 25.6% had no booking link, 18.4% sent bookings to a third-party domain and 54.7% had no telephone link. The median homepage did not pass 4 of the 16 core checks; 1.6% passed all. London homepages were more likely than those elsewhere to route bookings to a third party (odds ratio 2.89, 95% CI 1.98–4.21, Holm-adjusted p < 0.001) and less likely to lack a readable sitemap (OR 0.52, 0.35–0.78). Rendering JavaScript changed the structured-data result for 2.1% of pages without it. Conclusions. The main deficit is not general website quality but the absence of domain-specific, machine-readable restaurant information and of direct booking and menu pathways on the restaurant's own site. We release anonymised site-level data and analysis code. This deposit contains the preprint (PDF and LaTeX source), anonymised site-level data (no domains or names), a codebook, aggregate results, the analysis and figure scripts, and the source code of the measurement instrument and data-collection scripts.