The Restaurant Recovery - OpenTable Data

−100%

Peak collapse

vs 2019, April 2020

~2.5 yrs

Time to recover

Back to 2019 levels Oct 2022

60K+

Restaurants tracked

OpenTable global network

+20%

2026 vs 2025

US, Jan 2026 YoY

US seated diners - indexed to 2019 baseline (%)

vs 2019 baseline

pre-pandemic level

March–April 2020: The fastest demand destruction ever recorded in the hospitality sector. OpenTable reservations went from flat to literally −100% within weeks as lockdowns hit. Restaurant closures rose 32% in the US. The data was so dramatic it became a benchmark used by governments and economists worldwide to track the real-time impact of restrictions.

Mid-2020: Outdoor dining exploded - from less than 1% of seated parties pre-pandemic to over 12% by Aug–Nov 2020. The US hovered between −27% and −45% vs 2019 through the second half of 2020. OpenTable waived restaurant fees to keep its network intact during the trough.

2021: Recovery was lumpy. By April 2021 the US was back to ~97% of 2019 levels in some markets. Then Omicron hit in December - seated diners fell back to −33% vs 2019 within days. Washington DC dropped −53%, NYC −60%. The speed of the reversal was striking - from near-normal to deep contraction in under a week.

2022: "Revenge dining" drove the final push back to baseline. Premium dining (+50% and above per person) saw the biggest uplift. Bar seating up 225% vs 2019. Monday and Tuesday dining surged as hybrid work rewrote when people went out. By October 2022, US reservations crossed the 2019 baseline and held.

2023–2026: The dashboard shifted to YoY comparisons (vs prior year rather than 2019). Dining continued to grow - mid-week dining up 11% in 2024, solo dining up 10%, group dining (6+) up 8%. By January 2026, the US was running +20% YoY. OpenTable is now a general consumer confidence indicator, not just a restaurant metric.

Sources: OpenTable State of the Industry dashboard · Restaurant Dive · Axios · CoStar · OpenTable blog (2020–2026). 2019–2022 figures are vs 2019 baseline as reported. Post-2022 converted to approximate 2019-relative index from YoY figures.

Why this data matters

The real-time signal

Consumer confidence before the stats catch up

OpenTable data updates daily and leads official economic figures by weeks. People cancel discretionary restaurant spending before they cancel necessities - making reservations a leading indicator of consumer sentiment, not a lagging one. Governments and central banks used it in real time during Covid to calibrate policy.

By the time March 2020 GDP figures were published, OpenTable had already shown the full collapse six weeks earlier.

Goodhart's Law immunity

Nobody games a restaurant booking

The best proxies work because they measure frictionless behaviour - things people do without thinking about being measured. OpenTable captures genuine demand decisions: you either want to eat out or you don't. Unlike GDP or consumer confidence surveys, there's no way to optimise for the metric itself. This is what makes it durable as a signal.

Compare to PMI surveys - those can be gamed by sentiment and reporting bias. Bookings cannot.

The coverage caveat

Mid-market upward skew to keep in mind

OpenTable skews towards mid-to-upmarket restaurants that use its booking system. Cash-in-hand local spots, fast food, and lower-end eateries are largely invisible. This means it overweights discretionary spending by people who have disposable income - making it sensitive to middle-class economic anxiety rather than broad population behaviour.

The London data in 2022 showed −14% vs the rest of the UK at +18% - partly a composition effect from London's expensive restaurant mix.

Data access

The OpenTable State of the Industry dashboard is free and public at opentable.com/c/state-of-industry. Updated daily. Granular monthly/city data available. Historical Covid-era data also on Kaggle.

Methodology note

OpenTable changed its baseline in June 2024 to online reservations only (previously included walk-ins and phone). Pre-2024 figures are not directly comparable to post-2024 figures.

Part of a series

This visualisation accompanies a broader exploration of surprising economic proxies - indicators that are unexpectedly representative of something much larger than themselves.