Why this page exists
Intelligence is only as useful as it is trustworthy.
RAYN combines venue performance data, public market signals, and AI-generated recommendations into a single weekly brief for premium hospitality operators. Every number in that brief has a source. Every comparison has a methodology. Every recommendation states its confidence.
This page explains how RAYN works — what we measure, how we compare, when we show a benchmark, and when we don’t.
How RAYN classifies data
Every data point in RAYN carries one of four labels. These labels appear throughout the dashboard and the Monday Brief. They are not cosmetic — they define how much trust to place in each output.
Data that comes directly from RAYN booking and operational records for that venue.
This is RAYN’s highest-confidence data tier. It includes bookings, covers, revenue, table selection events, guest visit history, channel source, no-show and cancellation rates, and all metrics derived from them.
When a metric is labelled Measured, the underlying data was captured by RAYN directly and has not been estimated or inferred.
Data extracted from public sources — official APIs, publicly accessible information, and standard web research methods.
Observed data includes Google ratings and review counts, TripAdvisor rankings, Instagram follower snapshots, publicly listed menu prices, booking platform presence signals, local event listings, and press mentions.
Observed data is directionally reliable but subject to source availability, refresh timing, and the inherent limitations of public information. Every Observed data point is labelled with its source and the date it was last collected.
Outputs generated by RAYN’s AI model, combining Measured and Observed inputs to produce directional intelligence.
Estimated outputs include demand forecasts, recommended actions, competitor spend estimates, and market pressure signals. Every Estimated output shows a confidence level — High, Medium, or Low — and states the inputs that drove it.
Estimated outputs are directional. They are not presented as facts.
When RAYN does not have enough reliable data to produce a meaningful output, it shows nothing.
RAYN suppresses outputs rather than fabricating them. An empty section means the data required to fill it has not yet been collected or does not meet minimum quality standards. The section will state what is needed and when it is expected to become available.
The six RAYN benchmark indices
RAYN measures venue and market performance through six core indices. These indices are stable, consistently calculated, and designed to be learned over time. The Monday Brief leads with them.
Each index is scored relative to a baseline — either the venue’s own historical performance or an anonymous peer benchmark. A score of 100 means at baseline. Interpretation varies by index and is always stated in context.
Demand Index
How current booking pace compares to the venue's own trailing average. In later network phases, this will incorporate anonymised peer data from comparable venues on the platform.
MeasuredPrice Position Index
How the venue is priced relative to its benchmark-eligible competitor set, based on publicly available menu data. A score above 100 means priced above comp set average. Context matters — this reflects positioning, not performance.
ObservedReputation Index
A composite of Google rating, review volume momentum, and review response rate, benchmarked against the comp set.
ObservedDigital Presence Index
Social and search visibility relative to the peer set. Combines Instagram follower growth, booking platform coverage, and posting frequency.
ObservedYield Index
How effectively the venue monetises its available covers, time slots, and table inventory, relative to its own peak performance baseline.
MeasuredChannel Health Index
How efficiently the venue converts booking intent into confirmed, honoured reservations, across all channels.
MeasuredHow RAYN constructs competitor sets
RAYN’s competitor identification engine evaluates venues using eight dimensions: occasion and experience overlap, price tier proximity, venue type, location, programming intensity, daypart overlap, physical capabilities, and cuisine.
Each venue is assigned a weight profile based on its category — fine dining, rooftop social, beach club, hotel brunch, or casual premium. These profiles adjust the relative importance of each dimension to reflect how guests actually make substitution decisions in that category.
Competitors are classified into four tiers:
Direct substitutes in most guest contexts. Included in benchmark averages at full weight.
Relevant in specific contexts. Included in benchmark averages at reduced weight.
Venues the guest may trade up to. Tracked for intelligence purposes but never included in benchmark averages.
Newly detected or borderline venues. Monitored until sufficient data exists to classify properly.
Competitor names are never shown to venue operators. Operators see anonymous peer-group comparisons only.
When RAYN shows a benchmark
RAYN applies strict eligibility rules before displaying any benchmark output. A number is only shown when the underlying data meets minimum quality standards.
Full benchmark
Shown when peer coverage and freshness meet strict minimum standards. All required sources refreshed within their defined freshness window.
Directional benchmark
RAYN shows a clearly labelled directional benchmark when peer coverage is thinner or data is older than the full benchmark standard. A directional benchmark is a signal, not a standard. It is always marked visibly.
Anomaly period
The comparison window overlaps a known anomaly period — Ramadan, public holidays, major local events. The benchmark is shown with a flag and a note. RAYN does not silently blend anomaly periods with standard trading.
Insufficient data
Minimum eligibility criteria are not met. No number is displayed. The output states what is required and when it may become available.
RAYN does not show a weaker number in place of a suppressed one. Suppression is not a failure — it is the system working correctly.
Confidence levels
Every Estimated output and every benchmark carries a confidence level.
Strong data completeness, fresh sources, stable comp set, no anomaly period overlap.
Partial data, some stale sources, recent comp set change, or anomaly period present.
Significant data gaps or low source reliability. RAYN may suppress the output entirely at this level.
Confidence is calculated automatically from data completeness, source freshness, and source agreement. It is not manually assigned.
Comp set governance
Venue operators can request changes to their competitor set parameters — radius, cuisine filter, price tier, venue type. All requests are reviewed by RAYN before being applied.
All comp set changes are logged with a timestamp and reason. When a benchmark moves, operators can see whether the movement came from market change, venue performance change, or comp set change. These are different signals and should not be conflated.
Manual overrides — including forced inclusion or exclusion of specific competitors — are applied by RAYN and logged permanently in the change record.
What RAYN does not claim
RAYN is transparent about the boundaries of its intelligence.
RAYN does not claim:
- Market-wide occupancy or revenue averages not derived from RAYN venue data
- Exact revenue or booking volumes of specific competitor venues
- Any metric presented as a verified industry standard before the platform reaches Phase 3 network scale
RAYN does claim:
- Precise measurement of all data flowing through the RAYN platform
- Directional intelligence from publicly observable market signals, clearly labelled
- AI-generated recommendations based on stated inputs, with stated confidence
- Anonymous peer benchmarking once the platform reaches sufficient network scale
Data sources
RAYN collects data from the following sources:
| Source | Data collected |
|---|---|
| Google Places API | Ratings, review counts, opening hours |
| TripAdvisor | Ratings, review counts, rankings |
| Weekly public follower snapshots | |
| Venue websites, Zomato, Talabat | Menu pricing and promotional offers (collected fortnightly) |
| Booking platform listings | OpenTable, Eat App, Zomato, TheFork (availability signals) |
| Time Out Dubai, What's On, Visit Dubai | Local event listings |
| Google News | Press mentions |
RAYN uses official APIs where available. Where official APIs are not available, RAYN collects publicly accessible information through standard web research and approved data collection methods.
No personal data about competitor staff or guests is collected. All data is used for venue intelligence purposes only and is not redistributed or resold.
Anonymisation
Competitor names are visible only to RAYN’s internal team.
Venue operators see anonymous peer-group comparisons — their own index scores relative to a defined peer group, and market average lines, without any competitor attribution. No competitor is identified, ranked, or named in any operator-facing output, including the dashboard and the Monday Brief.
Methodology versioning
RAYN’s benchmark methodology is versioned. The current version is v1.1, effective March 2026.
When methodology changes are made:
- Minor changes (data source additions, threshold tuning) are deployed without operator notification
- Major changes (formula changes, new indices, weight profile changes) are run in parallel with the previous version for 30 days before full deployment, and operators are notified
The version used to produce each benchmark output is stored alongside the output. Historical benchmarks are not retroactively recalculated when methodology changes.
Questions
If you have questions about RAYN’s methodology, data sources, or benchmark outputs, contact the RAYN team through the contact form at rayn-co.com.