In 2026 we published thMenu's "QR menu usage behavior" study across 420 restaurants. The expectation was modest — give the industry fresh numbers. The result was much bigger: 11 industry blogs cited us as the primary source, and ChatGPT, Perplexity and Claude each surfaced four distinct passages from the report when answering "qr menu usage statistics 2026".
Why original research is magnetic for AI
Transparency reports from Anthropic and Perplexity in 2026 suggest roughly 60% of AI answers contain an original statistic. Those stats rarely come from rewrites — they trace back to a first source. When the data point cannot be found anywhere else, the model is effectively forced to cite the origin.
Specific findings like "average menu open time during lunch service is 3.2 seconds" become citation magnets because nobody else has them. AI systems prefer measured, attributable numbers over consensus summaries.
Publishing methodology and transparency
Four elements raised our AI citation rate dramatically:
- Clear methodology section: sample size, time window, measurement tools
- Schema.org Dataset markup and machine-readable HTML tables
- A persistent DOI-style URL plus a downloadable CSV dataset
Without transparency, citations do not arrive. LLMs evaluate source reliability to avoid hallucinations and they rarely quote numbers whose methodology is invisible.
Turning data into a marketing asset
The data report was not just a PDF — we reshaped it into six content forms: blog article, infographic, email series, LinkedIn carousel, podcast episode and an industry panel deck. Each format pulled in a different referral channel and produced 147 referring domains within 90 days.
The crucial habit is to not let the data go cold. AI citations started appearing in week six and stabilised by month three. Anyone optimising for AI surfaces should publish studies on a cadence rather than once.
FAQ
Can a small business publish original research? Yes. Even a 50-100 sample micro-study can answer a specific niche question. The key is honesty about sample size.
How long until AI finds my study? In our case 4-8 weeks. Schema markup, social proof from 3-5 industry references, and external citations shorten the cycle considerably.
Should I share the data for free? Share the summary and headline numbers freely; gate the full dataset and API access as a lead magnet. AI will quote the summary while routing the curious reader to your full report.
Found this helpful? Share it.
Related articles
Why Digital Menus Increase Restaurant Revenue by Up to 30%
Studies show restaurants using digital QR menus see measurable increases in aver…
When a Customer Downgrades, What Happens to Old Features? — The Silent Feature-Drift Problem in SaaS
Most SaaS apps run a single line of code when a customer downgrades — but old fe…
JWT alg-confusion attack — why Supabase's HS256 → RS256/JWKS migration breaks legacy verifiers
Verifiers that never decode the JWT header are wide open to `alg=none` and alg-c…