Owner's Read: Leakage & Control Review

Method and coverage, stated plainly

The Standard Read is a public-data review: no credentials, no system access, no internal figures. This sample was produced across four parallel workstreams in a single working day: a website and direct-booking review with controlled-browser price checks; an OTA and channel mapping; a reputation and control-signal analysis; and an AI-assistant visibility run across several assistant engines. Prices are recorded as shown to an anonymous shopper on the dates checked, and may differ by device or login. Where a number could not be derived from public data, the register says so rather than inventing one.

The verdict in three sentences

The product and its reputation are strong: the best review scores in its competitive set, on every platform, a few months after opening. The selling machinery around that product leaks: the direct channel is the most expensive and least flexible way to book the property everywhere a guest can compare, a season-long OTA discount is running against it, and a mid-five-figure annual sum (illustrative band) is flowing to OTA commission on demand the brand itself creates. Most of the top fixes are free and inside two weeks' reach, and two need owner decisions this week.

Executive findings, the headline leaks

A leading OTA shows the identical room roughly 25–30% cheaper gross for the same dates, and sells free cancellation while the direct site is prepay-only. Every guest who rate-checks, which is every guest, is priced and termed off the owner's channel onto a commissioned one.

Whether anyone authorised it is exactly the kind of question an owner-side read exists to ask. The register flags an owner decision this week: audit the OTA promotions and set a sign-off rule.

Named in a low single-digit percentage of valid assistant answers for its own market, and in none on one major assistant, while three named competitors take the recommendations despite the property outscoring all of them on review rating.

A global review site has the property filed in the wrong region and near-invisible in its home market, and the maps and business profile, its strongest public surface, shows an ownership-claim prompt. Both are free fixes that also feed the AI answer layer.

OTA and direct-channel economics

The single largest leak domain for most independent properties. The channel mix is measurable from public listings and, in a paid Read, confirmed against the property's own reservation data. The commission overlay on top is an estimate, clearly flagged as one, because public data cannot read the real commission line.

Read this as a shape, not as one property's accounts: roughly three of every four revenue units arriving through a channel that takes a margin the owner never sees on a statement, while the owner's own channel is priced above it. In a Pro Read, the same estimate is rebuilt from reservation sources and commission invoices, and on the group property that step made the leak larger, not smaller.

Reputation and demand signals

Reputation is not the lead here, it is one of six leak domains. On this property it is a strength being under-used rather than a weakness.

• The review lead is genuine and fragile. Best ratings in the competitive set, but on a small review count against competitors with hundreds or thousands, and with no established review-reply habit before it is ever needed.
• One public defect is on display to every booker. A single low connectivity sub-score sits against otherwise excellent category scores, shown to every prospective guest in a market with heavy remote-work demand.
• The same rooms carry several naming schemes across channels. Price comparison breaks in the OTAs' favour, bed-type claims disagree between listings, and one booking-engine heading is misspelt.
• A whole unit class and the food-and-beverage trade have no online demand path. A placeholder page with no bookable unit, and a dining room guests praise in reviews but that has no findable public presence for non-residents.

Demand signal, in one line: strong word of mouth is not reaching the surfaces where new guests and AI assistants actually look, so the property under-earns on a reputation it has already paid for.

The profit-leakage register

The core of the deliverable: a ranked register across the six leak domains, ordered by severity, money and fixability. In a real Read each row is a full finding object (statement, evidence, severity, confidence, estimated exposure, what would prove it properly, recommended action, owner decision required, domain). Here the evidence and exact figures are redacted.

* "Not quantifiable" means not derivable from public data. It is a real finding state, not an omission: the register names the private data that would size it and carries it into the fix order and the recurring control layer.

Prioritised interventions, the fix order

The register ends in a fix order: what to fix first, what the money is, what to build next, and what not to touch yet. Free, no-trade-off fixes cluster at the top. Owner decisions are surfaced as questions, never made silently.

This week, free, no trade-offs

1. Audit the OTA promotions and knowingly kill or re-approve the season-long discount. Set an owner sign-off rule for any future OTA promotion.
2. Claim and correct the global review-site listing, including its geography.
3. Claim, or verify claimed, the maps and business profile.
4. Fix the direct-path friction: the click-to-call number, the engine's default calendar view, the enquiry form fields.

Next two weeks, the money (needs the owner's parity decision first)

1. Decide the parity policy, then reload rates so direct is never the most expensive public price, and align cancellation terms.
2. State the direct-booking reason on the site and engine, and deep-link room pages with dates.
3. Publish one owner-approved product map: the same room names, plans and terms across site, engine and OTAs.

This month and next; and what not to touch yet

• Give the unlisted unit class and the dining room a public demand path, once the owner confirms intent.
• Start the review-ask flow and the reply habit; re-run the AI-visibility check monthly on a fixed method.
• Not yet, deliberately: the brand-architecture question deserves a considered decision, not a quick fix, and rate yielding comes after parity is fixed, not before. No paid acquisition of any kind: nothing above requires it.

Estimated financial impact (illustrative bands)

Bands, not a promise, and redacted from a group property rather than a client's accounts. Several domains are honestly marked "not quantifiable from public data": that is the point at which private data, and the recurring control layer, take over.

On the group property, opening the books later moved the channel-margin figure above this public band, not below it. The proxy model runs conservative by design.

Data limitations, stated openly

A Standard Read is a public-data outside view. Its honesty is part of the product, so its limits are named rather than hidden.

• Public data only. No credentials, no system access, no internal figures. Where a claim needs private data to size, the register says so and points to the exact data that would prove it.
• Estimates are ranges, with the arithmetic shown. Commission and margin figures are built from public proxies (review-rate assumptions, shelf-price ADR, standard commission bands) and marked as estimates, never presented as fact.
• Prices are point-in-time. Each price is recorded as shown to an anonymous shopper on a stated date and may differ by device, login or currency.
• AI-visibility results are method-and-date specific. They are run in a neutral environment with a contamination check, reported per assistant, and never compared across method changes.
• Cash clarity is deliberately partial here. The true margin cost, payout timing and reconciliation gaps sit behind private data. Sizing them is the Pro Read and the recurring control layer, not the Standard.

90-day action sequence

The close of Read

Every delivered Read ends with one question, which is the bridge to what comes next:

The standard behind it, stated as a plain commitment and not a claim of results: if the readout does not leave the owner with at least three specific, costed actions they did not already have, we do one further round or refund the fee.

The ladder this sits on

Free AI Visibility Check, then the paid Owner's Read below, then the recurring control layer if the owner wants the findings watched and fixed over time.

The Owner's Read fee is credited in full against protected work started within 30 days of the readout. Credited once, non-transferable.