Skip to content
Data report

What actually drives revenue in bars and restaurants

We classified every one of 13,063 guest reviews across 41 Texas venues, twice, using two separate frontier models, then checked the result against compulsory state alcohol tax filings. Some of what we found contradicts what the industry teaches.

Book a DemoRead the methodology
What the study is made of

13,063

guest reviews classified

Every review, not a sample.

41

Texas venues

Houston, Katy, Sugar Land and The Woodlands.

2

independent models

Classified twice, so agreement could be measured.

1

external cross-check

Results held against state alcohol tax filings.

These are counts of what the study contains, not results. A research page whose headline figures are outcomes is a marketing page wearing a lab coat. The outcomes are on the case studies, labelled as client reported.

Methodology

How it was done, and what it does not support

Collection

Every public guest review for 41 venues across Houston, Katy, Sugar Land and The Woodlands, taken in full rather than sampled.

What it does not supportReviewers are not a random sample of guests. People who write reviews differ from people who do not, in ways that matter and that this study cannot correct for.

Classification

Each review classified into themes twice, by two separately built frontier models, with no shared prompt or scoring rubric between them.

What it does not supportAgreement between two models is weaker evidence than a controlled experiment. It rules out one model's idiosyncrasy, not a bias they might share.

External check

Classified results held against Texas state alcohol tax filings, so the themes could be tested against something outside the text itself.

What it does not supportTax filings describe alcohol sales, not the guest experience. They constrain the interpretation; they do not confirm it.

Interpretation

Relationships reported between what guests wrote about and how venues were trading over the same period.

What it does not supportRelationships, not causes. A venue with better reviews and better trading has not been shown to have one because of the other, and this study was not designed to show that.

On this page
  1. Three findings that surprised us
  2. The caveat that matters most
  3. How it was measured
  4. The volumes
  5. Data ethics
  6. Frequently asked questions

Three findings that surprised us

Naming drives rating. Occasion drives revenue. Two different games, and conflating them is the central mistake an operator can make with this data.

Guests who name a staff member in a review are far more positive. That result replicated three separate times across our earlier volumes. It is also, for revenue purposes, close to irrelevant: naming an employee separates high-revenue venues from lower-revenue ones by a ratio of just 1.20x.

What does separate them is occasion language. Reviews that mention a reservation appear 3.40x more often at the highest-revenue venues. Celebration language runs 2.81x. Group and party-size language 2.25x. The guest who books is worth more than the guest who is delighted, because booking converts a discretionary visit into a scheduled obligation.

Negative reviews carried no measurable revenue cost r = -0.021 across 40 venues, measured on 23,538 consensus labels

This is the finding we most expected to be wrong, and it survived every robustness check we ran. The overall rate at which guests complain about something is statistically indistinguishable from zero as a predictor of receipts.

That does not mean complaints are free. Broken out by aspect, food complaints do track revenue (r = -0.489 on the consensus label set). It is the aggregate negativity that carries no signal. Read the caveat below before acting on this.

Only 22 percent of venues have one clear problem For the other 78 percent the top issues are effectively tied

Operators tend to look for the single thing to fix. In 32 of 41 venues, the top two complaint categories sat within 25 percent of each other on mention count. There was no single thing. That has direct consequences for how you prioritise, and it is the kind of result you only get by counting every review rather than asking someone to summarise a sample.

The caveat that matters most

Every venue in this study is a top-10 alcohol earner in its city. All 41 of them are winners already, so every correlation here is a comparison among winners and none of it is causal.

The finding that negatives carry no revenue cost applies inside that elite cohort. It says nothing about a venue with a 29 percent negative rate that is not already a top earner, because venues like that are absent from the sample precisely because they failed. If you take one caution from this page, take that one.

How it was measured

Three tiers, deliberately kept separate so that the exact parts stay exact.

  • State tax filings. TABC Mixed Beverage Gross Receipts, dataset naix-2893. Compulsory, audited, monthly, venue level, public.
  • Star-derived sentiment. Computed in Python with no model involvement. Identical across every model tested.
  • Aspect labels. All 13,063 reviews classified individually by two frontier models. 98.1 percent polarity agreement across 23,993 aspect pairs. All counting, percentages and ranking done deterministically in code.

The full method, including what we got wrong and had to rebuild, is here.

The volumes

The wider program runs to six volumes and roughly 63,000 words, covering the hospitality literature, guest and worker language, and city-level breakdowns for Houston, Katy, Sugar Land and The Woodlands.

Data ethics

Review scraping ran with personal data collection disabled throughout. No reviewer names, identifiers, photographs or profile links were retrieved or stored. Reviews are quoted by star rating and date only. Employee first names that appear are ones guests published publicly, and they are reported only as aggregate patterns, never attached to an individual's performance.

One venue's review corpus was excluded after we determined it was not authentic. Its tax filings remain in the receipts analysis; its review data is not used anywhere.

Frequently asked questions

Which Texas markets does the research cover?

Houston, Katy, Sugar Land and The Woodlands. The 41 venues are each a top-10 alcohol earner in their city, and together their guests left the 13,063 reviews the research classifies. That selection is also the study's main limit: every comparison is among venues that are already winning.

Can I cite or reuse the research?

Yes. The research dataset and the Texas bar guest-review benchmark are published under the Creative Commons Attribution 4.0 licence. Cite Delta Kinetics and link to the page you used.

Is the data available to download?

Yes, in aggregate. The Texas bar guest-review benchmark publishes the review data as a table by market and complaint theme, with a CSV download, and is refreshed quarterly. Every figure in it combines at least 3 venues, so no single venue can be identified.

See the same measurement on your own venue

The method behind this research is what CoreTAP runs continuously. Bring two weeks of data and we will show you your own numbers.

Book a DemoEmail Us