How to turn public rate, availability and demand data into a pricing advantage your competitors do not have.
In short: the signals that tell you a market is about to compress are public, and they show up weeks before your own bookings do. They fall into four layers: forward demand, supply, price and reputation. Most teams only watch price. This article explains what each layer is worth commercially, how often to watch it, and how to put it to work without building a data team.
Who it is for: revenue managers, commercial directors and distribution leads at hotel groups, vacation rental managers, experience marketplaces and OTAs.
By the time you see it in your bookings, someone else already priced it
A market that is compressing rarely announces itself. It shows up first in availability starting to close across three properties comparable to yours, then in the minimum-stay restrictions someone loads for a weekend, then in a rate that moves up without any announcement, and only at the very end in a confirmed booking.
By the time it reaches your own numbers, the pricing decision that would have captured that booking has already been made, and someone else made it.
That lag is not a flaw in your revenue management system. It is a property of how the system is built: it measures confirmed bookings, and a confirmed booking is the last link in a chain that played out over weeks, in public sources your system never queries. The advantage is not in measuring that last link more precisely. It is in watching the ones before it.
Those earlier links can be organized into four layers.
The four signal layers, and what each one is worth
Not every source says the same thing, and they do not move at the same speed. Separating them is what lets you decide what is worth collecting, how often, and what commercial decision it feeds.
Layer 1. Forward demand: know which dates will sell before they sell. The first to appear and the one almost nobody collects. Public event calendars for the destination, conventions, sporting fixtures and school calendars: open sources that tell you which dates are going to behave differently. It tells you demand is coming, not how much or at what price. There are also flight-search and accommodation-search signals sold by specialized providers, with reported lead times of 80 to 145 days, worth knowing about even though they do not come from open sources.
Commercial value: you open rates, restrictions and campaigns for high-demand dates while there is still time to capture them.
Layer 2. Supply and availability: see compression before it hits your occupancy. Inventory published by your comp set, minimum-stay restrictions, date closeouts. This is the first hard indicator of compression: when three comparable properties close a date, that date has changed character before your own occupancy reflects it.
Commercial value: you raise rates on the right dates instead of selling your last rooms at yesterday's price.
Layer 3. Price: know where you stand, channel by channel. Rates published by channel, active promotions, cancellation policies, cleaning fees and mandatory charges. This is the layer everyone watches and the easiest one to misread, because the published number is not always comparable across channels.
Commercial value: positioning decisions based on real, comparable prices, and early detection of competitor promotions.
Layer 4. Reputation and product: protect your pricing power. Scores, rating distribution, recent text and the attributes guests mention. It is the slowest layer and the one with the most predictive value for medium-term pricing power: the Cornell study published in 2012 associated a one-point increase in the reputation index with 1.42% higher RevPAR.
Commercial value: you know when you can charge more than your comp set, and when a reputation gap is costing you rate.
All four are public. The difference between the teams that use them and the teams that do not is collection capacity, not access.
| Source | Update frequency | Signal layer | Primary use case |
|---|---|---|---|
| Public destination event calendars | Weekly | Forward demand | Early detection of dates that will compress |
| Booking.com, Expedia, Hotels.com | Daily or higher | Price + Supply | Published rate, restrictions, comp set positioning |
| Airbnb, Vrbo | Daily | Price + Supply | Nightly rate, minimum stay, seasonal pricing |
| Viator, GetYourGuide | Daily or weekly | Price + Supply | Experience catalogs, pricing tiers, availability windows |
| TripAdvisor, Booking, Google | Continuous | Reputation | Scores, rating distribution, recent text |
| Property direct websites | Daily | Price | Disparity between direct and intermediated channels |
How often each layer needs to be watched
The four do not move at the same rate, and that is what sets collection cadence, and cost.
Layer 1 can be swept weekly or monthly: an event calendar changes little once it is published. Layer 4 needs no more than a weekly pass either, because a reputation score moves slowly by construction.
Layers 2 and 3 are a different matter. They move daily and, on compressed dates, several times a day. And they have a property the other two do not: they do not archive themselves. An event calendar is still published next week. The rate a competitor posted yesterday is not stored anywhere unless someone captured it. Every day you do not collect is history you will never get back.
One reference point helps size the minimum coverage: the average booking window was 32.15 days in 2025, according to SiteMinder's report built on more than 130 million hotel bookings. It is a global figure, it is not broken out by country, and it comes from one provider's customer base, so it is worth comparing against the number for your own market. That month is the period during which the competitive set forms the price of every date.
What it takes to turn public data into rate intelligence
Collecting is not the hard part. The hard part is making what you collect comparable, consistent over time, and usable without a manual cleanup that cancels out the benefit of having automated it. A working rate intelligence pipeline has five stages:
1. Capture, from the traveler's position. The rate that gets displayed depends on the country you query from, the currency, the device, and whether you are logged in. A capture taken from the office with the corporate account open is not the rate your source market sees. Every channel is queried with the same parameters and at the same moment: capturing Booking on Monday and Expedia on Thursday produces false disparities that burn hours of investigation and do not exist.
2. Normalization, into a comparable unit. In vacation rentals that means the total cost of a defined stay pattern, not the nightly rate, because cleaning fees, minimum stays, occupancy-based pricing and length-of-stay discounts reorder the ranking. In hotels it means storing the components separately: base rate, mandatory charges, taxes and optional extras.
3. Entity resolution, so you compare the same property. The same property appears on four channels with four names, four differently written addresses and four internal identifiers. Without a reliable map of what is what, the dataset produces comparisons between properties that are not the same property, which is the most expensive error because it is silent.
4. Aggregation, as a series that remembers. The series is appended, not overwritten. A spreadsheet that gets rewritten every week cannot answer what this market did last year on this same date, and that is exactly the question that decides a seasonal rate. Append-only storage with timestamps, list price and effective price in separate columns.
5. Distribution, where decisions get made. Delivery in the format where the decision actually gets made: CSV, Excel, JSON, by API, or directly into the data warehouse or the revenue management platform. An excellent dataset living in a system nobody opens does not change a single rate.
You do not need to build this pipeline yourself. AUTOScraping runs capture, normalization and delivery as a managed service. See how Travel Data Extraction works.
Who wins with this: hotels, vacation rentals, marketplaces and OTAs
Hotel revenue management teams
The most common case is also the most manual: an analyst opens an incognito window, checks their own property and five competitors across two or three dates, writes it down, and repeats the following week.
The cost of that method is not the hours. It is three things the method cannot produce by definition, and all three are needed to read pickup, meaning the rate at which reservations come in for each arrival date, against what the market is actually doing:
- Coverage: five properties and three dates is what fits in a morning, while your real market has thirty properties and ninety relevant dates.
- Reproducibility: two analysts capture different things because the rate depends on the session.
- History: a spreadsheet that gets overwritten has no memory.
The 2026 context makes this more expensive than it used to be. The CoStar and Tourism Economics forecast published on January 28, 2026 projects ADR up 1%, occupancy at 62.1% and RevPAR up 0.6% for US hotels. With RevPAR growing less than a point, one operator's growth is arithmetically share that another one loses. There is no rising tide to cover a pricing mistake, and the operator with better market visibility takes the share.
Vacation rental managers
Here the problem is not the price, it is the unit. A property listed at $180 a night with a five-night minimum and a $150 cleaning fee is more expensive, for a five-night stay, than one listed at $210 with a two-night minimum and a $60 cleaning fee. Both show up in the same search, sorted by the number that decides nothing.
Normalizing to the total cost of a defined stay pattern, applied identically across the whole set, is what turns a list of prices into a pricing dataset, and what lets you price to win the stays you actually want.
There is a regulatory layer worth knowing about. Since May 12, 2025, the Federal Trade Commission's total-price rule, 16 CFR Part 464, has applied in the United States to short-term lodging, including vacation rentals and home shares. It requires displaying the total price with all mandatory fees more prominently than any other pricing information, but it allows three categories to be excluded: taxes and government charges, shipping, and optional goods or services. Those exclusions are why two totals from two platforms can still fail to be comparable, and why capturing components beats capturing the final number.
Experience marketplaces and operators
This is the fastest-growing segment and the worst-observed one. According to Phocuswright, the experiences market went from $253 billion in 2024 to a projected $342 billion in 2029, growing 17% in 2024 against 6% for travel overall.
But only 33% of tours, activities and attractions bookings go through online channels, against 64% for the rest of the industry, and more than 70% of operators are small or micro businesses. There is no consolidated feed because there is no consolidation: price visibility gets built source by source, or it does not get built. For a marketplace, that makes catalog and pricing data a competitive moat, not a commodity.
OTA platforms and distribution teams
The structural problem is that the channel manager is an output system, not an observation system: it records what was sent, not what the channel published.
The size of the problem has been measured. Research published by Expedia Group on October 1, 2025, conducted with 2,000 hotel revenue managers across eight markets including the United States, found that 98% lose an average of 6% of annual revenue to rate leakage, that more than half detect it weekly or more often, and that 52% manage between four and six distribution partners, at an average administrative cost of $40,100 per property per year. It is worth noting that Expedia is an interested party, since the research accompanies the promotion of its own centralized distribution platform, though the operational figures are consistent with what the rest of the industry reports.
Observing what gets published from the outside is the only way to close that gap, and it is external collection work before it is a feature any distribution system could have.
Three mistakes that destroy the return on travel data
1. Storing the final number instead of the components. This is the error that goes unnoticed for six months, until someone asks why one channel's total differs from another's and the answer lives in a field that was never stored. Separate components from day one: they are cheap to store and impossible to reconstruct.
2. Overwriting the series instead of appending to it. A dataset that reflects today's state is a dashboard. A dataset that preserves every prior state is an asset. The cost difference is marginal and the value difference is total, because seasonality, trend detection and any year-over-year comparison live exclusively in the history.
3. Building alerts without a threshold defined in advance. Without an explicit criterion for which change warrants a review, the system produces notifications nobody processes. It is the most common way a project like this gets abandoned at the six-month mark, and it does not fail on engineering, it fails on design.
Where to start
You do not need all four layers on day one. A practical sequence:
- Start with price and supply (Layers 3 and 2) for your real competitive set and the dates inside your booking window. That is where decisions happen every day, and where history is lost if you do not capture it.
- Add forward demand (Layer 1) once the price series is running, to know which dates deserve closer watching.
- Add reputation (Layer 4) to support medium-term positioning and rate strategy.
Each step adds a layer your competitors are probably not watching.
Frequently asked questions
What is travel data extraction?
It is the automated collection of public travel data, such as rates, availability, restrictions, catalogs and reviews, from OTAs, rental platforms, experience marketplaces and direct websites, delivered as structured data for pricing and commercial decisions.
Is it different from a rate shopping tool?
Rate shopping tools usually focus on price (Layer 3). A travel data extraction project can cover all four layers, with the parameters, sources and delivery format defined for your business.
How much history do I need before it is useful?
The data is useful from the first day for current positioning. Seasonal and year-over-year analysis needs the series to keep growing, which is why starting the capture early matters.
Do I need a data team to use this?
No. With a managed service, the data arrives in the tools your team already uses, such as Excel, your data warehouse or your revenue management platform.
Almost nobody collects all four. That is your opportunity.
Most teams work on Layer 3, because price is what feeds the rate decision and it is what the tools on the market surface first. Some also watch Layer 2, almost always manually and across a handful of properties. Layers 1 and 4 get left out: the first because it lives in sources nobody associates with pricing, the second because it belongs to another team and gets reviewed once a month.
All four layers are public and none of them requires privileged access. The difference between two operators running the same revenue management tool is rarely in the tool. It is in how many of the four layers feed it, and in whether they arrive as a continuous series or as a snapshot someone took on a Tuesday.
Get all four layers without building the pipeline
AUTOScraping runs the collection layer for revenue management teams, OTA platforms, experience marketplaces and vacation rental managers:
- Daily monitoring of rates, availability, catalogs and reviews.
- Captured from the traveler's position, with consistent parameters across channels.
- Normalized and delivered in CSV, Excel, JSON, by API or straight into your data warehouse.
Tell us your market, your competitive set and the decisions you want to support, and we'll show you which layers to start with. Learn more about Travel Data Extraction or get in touch to review your case.
Sources
- SiteMinder, Hotel Booking Trends, report published February 3, 2026, based on more than 130 million hotel bookings from 2025
- Hospitality Today, Forward-looking data for hotel forecasting: What's available in 2026
- Hotel Dive, US hotel RevPAR, ADR set for growth in 2026, CoStar and Tourism Economics forecast published January 28, 2026
- Phocuswright, Tours and activities are outgrowing travel overall. Distribution hasn't caught up.
- Hospitality Net, The Untold Story of Rate Misuse: Expedia Group Research, published October 1, 2025
- Hospitality Net, Cornell Hospitality Research: Online Reputation Directly Affects Pricing Power, Occupancy & RevPAR, study by Chris Anderson published November 16, 2012
- FTC, Rule on Unfair or Deceptive Fees: Frequently Asked Questions
