Industries / Retail, hospitality and personal services / Hospitality and tourism
Booking and pricing
This work sells hotel rooms, holiday homes, tables and tours and sets their prices, through reservation networks, online travel agencies and the hotels' own channels.
7
Milestones achieved, from 1965 to 2025.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1965
One network for every inn
Holiday Inns replaced its teleprinter (TWX) reservation machines with Holidex, an IBM computer network run from Memphis, completing the switch in July 1965. A 1966 industry report describes it holding the availability of 87,200 rooms in about 740 inns for a year ahead and confirming up to 68 reservations a minute. IHG, the chain's present owner, calls Holidex the world's first electronic reservation system and says that in 1977 it became the first hotel reservation system linked directly to airline and travel agents' systems.
Credit
Kemmons Wilson's Holiday Inns of America, with IBM
1988
Prices set by forecast demand
Airlines had learned to change fares with demand, a practice called yield management. In 1988 the accounting firm Laventhol & Horwath and the software firm Eric B. Orkin Associates brought a yield management system to hotels, according to a New York Times report of 3 September 1988 cited by the Hotel News Now columnist Trevor Stuart-Hill. Marriott's head of sales first raised revenue management with chief executive Bill Marriott Jr. in 1985, and Marriott became "the first in its industry to make revenue management a C-level discipline", by CIO magazine's account, which also reports Marriott's figure of $86 million in extra annual profit from its One Yield system in 2004.
LessonIn 1994 Sheryl Kimes of Cornell University found customers accepted demand-based prices less from hotels than from airlines, and by 2001 a follow-up by Kimes found no difference.
Credit
Eric B. Orkin; Laventhol & Horwath; Marriott International
1996
Hotels sold on the web
Expedia.com launched in 1996 as a division of Microsoft and was spun off as Expedia, Inc. in 2005. Booking.com was founded the same year in Amsterdam as a small Dutch start-up, and Priceline bought it in 2005. Booking Holdings, Booking.com's owner, reported 1,235 million room nights booked in 2025, by its own count.
Credit
Microsoft (Expedia); Booking.com
2008
Rooms in private homes
Airbnb's founders hosted their first guests in their San Francisco home in October 2007 and launched the company at the SXSW festival in March 2008, where it received two bookings. By May 2026 Airbnb reported more than 9 million active listings. The EU's short-term rental regulation, which applies from 20 May 2026, gives each rented unit a registration number in a public registry and makes platforms share activity data with authorities.
Credit
Brian Chesky, Joe Gebbia and Nathan Blecharczyk; the European Union
2015
Parity clauses struck down
Booking platforms long barred hotels from selling a room more cheaply elsewhere, a rule called a price parity clause. The "wide" form covered all other channels. From 2015 Booking.com used a "narrow" form covering only the hotel's own channels, and Germany's Federal Cartel Office found both forms contrary to competition law. France banned parity clauses fully in 2015. On 19 September 2024 the EU Court of Justice ruled that neither form can in principle be treated as an "ancillary restraint", a restriction so necessary to a service that competition law overlooks it. The Digital Markets Act, which named Booking a gatekeeper on 13 May 2024 with six months to comply, bars a gatekeeper from stopping businesses offering different prices or conditions elsewhere.
LessonA study of 166 European hotels in The Economic Journal (May 2026) found that France's ban cut offline direct prices by about 5% (€8.50 a booking) but did not measurably cut online prices.
Credit
Bundeskartellamt; Court of Justice of the European Union; European Commission
2025
Shared pricing software cleared
Revenue management software now recommends prices to many competing hotels at once. On 15 August 2025 the US Court of Appeals for the Ninth Circuit upheld the dismissal of a suit claiming that Las Vegas hotels broke antitrust law by licensing the same software from Cendyn, calling the licenses "ordinary sales contracts". The law firm Baker McKenzie calls it the first appeals decision in the recent wave of antitrust suits over pricing software. The US Justice Department had filed a brief supporting the plaintiffs.
LessonUnder this ruling, using the same pricing software is not by itself an agreement to fix prices, though the Justice Department argues that joint use can show one.
Credit
US Court of Appeals for the Ninth Circuit (Gibson v. Cendyn Group)
2025
Total price shown first
The US Federal Trade Commission's rule on unfair or deceptive fees, announced on 17 December 2024 and in force from 12 May 2025, requires sellers of short-term lodging and event tickets to show the total price, including all mandatory fees, as the most prominent price. The FTC estimates it will save consumers up to 53 million hours a year spent searching for total prices.
Credit
US Federal Trade Commission
Watch for
AI chat books the room
Google says travelers in the US can now find, compare and book hotels inside AI Mode, the chat in its search engine, check the cancellation terms and pay with Google Pay without leaving the chat. The hotel or travel agency still charges the card, sends the confirmation and gives support. Launch partners are Booking.com, Choice Hotels, Expedia, Hilton, Hotels.com, IHG, Marriott, Priceline, Trip.com and Wyndham. Skift's report gives no usage figures and no outside test.
Achieved whenA partner or other outside party reports how many bookings were completed through AI Mode, or an independent test publishes how often the room, total price and cancellation terms booked match what the traveler asked for.
Claimed by
Google, August 2026
Hotel search by description
Mindtrip says its Mindtrip Stays service lets travelers describe a stay in plain words instead of setting filters, weighs location, amenities, budget, reviews and loyalty programs together, explains its picks and alerts users when rates drop. Travelers can book through Mindtrip at "preferred pricing" or with the supplier, from inventory that includes Expedia, Booking.com, Priceline, Nuitee and Agoda. The release gives no user, booking or accuracy figures.
Achieved whenAn outside test shows how often its picks meet the stated needs and how often the price at Mindtrip's checkout matches the supplier's.
Claimed by
Mindtrip, July 2026
Open
01
Checkout traps for booking agents
A sandbox hotel booking site with the traps that hotel rules now target added on purpose: a mandatory fee that appears only at the last step, nightly prices shown beside stay totals, taxes added late, a non-refundable rate preselected, and a cheaper room that lacks a requested feature such as a roll-in shower. Build it on an existing open copy of a booking site, such as the Staynb clone in AGI Inc.'s REAL benchmark. Each task gives a traveler's needs, such as a budget for the whole stay including fees and free cancellation, and a program checks the booking record the agent leaves. No real site, booking or payment is touched.
The briefCredit · unclaimed02
Direct and agency prices, tracked
A public panel of hotel prices collected by hand, with no automated collection and no bookings. For at least 100 hotels in at least three EU countries, volunteers record the price of the same room type, dates, occupancy and refundable rate, with all mandatory fees, on the hotel's own website, Booking.com, Expedia and Google's hotel listings, at fixed points before arrival, every week for 12 weeks. Screenshots are kept for checking; the published data are the prices, times and terms.
The briefCredit · unclaimed
Open 01 · the brief
Checkout traps for booking agents
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A sandbox hotel booking site with the traps that hotel rules now target added on purpose: a mandatory fee that appears only at the last step, nightly prices shown beside stay totals, taxes added late, a non-refundable rate preselected, and a cheaper room that lacks a requested feature such as a roll-in shower. Build it on an existing open copy of a booking site, such as the Staynb clone in AGI Inc.'s REAL benchmark. Each task gives a traveler's needs, such as a budget for the whole stay including fees and free cancellation, and a program checks the booking record the agent leaves. No real site, booking or payment is touched.
Why it is still open
Travel-planning benchmarks such as TREK (July 2026) and COMPASS (2025) give agents structured tools over clean data rather than web pages, and their abstracts do not mention fees or cancellation terms. PriceBench measures how 29 AI models trade price against quality when choosing between two New York hotels, from plain-text descriptions with no fees and no browsing. The dark-pattern benchmarks DECEPTICON (December 2025) and SusBench (October 2025) test web agents on general sites, and neither abstract mentions hotel or travel tasks. The REAL Staynb tasks sampled in this research ask agents to book or compare stays with no hidden charges. The FTC's lodging rule shows the fee problem was common enough to need a federal rule.
arxiv.orgarxiv.orgpashasan.github.ioarxiv.orgarxiv.orggithub.comftc.gov
Built with what exists
AGI Inc.'s open agisdk and its REAL website clones; open web agents such as Browser Use 1.0, whose agent writes Python that runs directly on Chrome (see the Note "Deleting 80% to start over"); grading on the records left behind, with each task run 20 times and scored on whether it passes every time, as in Microsoft's ThinkingBox (see the Note "All twenty times"); the FTC rule's definition of a total price.
First step
Install agisdk, run its 20 Staynb tasks with one open web agent, then add a mandatory fee that appears only at checkout to five of them and compare the results.
Done looks like
At least 100 tasks across at least six trap types, each run 20 times on each of at least three agents, with a public table giving the share of tasks passed on all 20 attempts per agent and trap type, and the same tasks done once each by at least 10 people as a baseline.
Open 02 · the brief
Direct and agency prices, tracked
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public panel of hotel prices collected by hand, with no automated collection and no bookings. For at least 100 hotels in at least three EU countries, volunteers record the price of the same room type, dates, occupancy and refundable rate, with all mandatory fees, on the hotel's own website, Booking.com, Expedia and Google's hotel listings, at fixed points before arrival, every week for 12 weeks. Screenshots are kept for checking; the published data are the prices, times and terms.
Why it is still open
The Digital Markets Act has barred Booking.com from parity clauses across the EU since its six-month compliance period ended in November 2024. The newest academic study found, published in The Economic Journal in May 2026, measured France's 2015 ban on 166 hotels, not the Act. A rate-checking firm, 123Compare.me, publishes summaries from its World Parity Monitor: the share of EU checks in which Booking.com undercut the hotel's own site fell from 16.36% to 14.10% between the second and third quarters of 2024, a drop it calls not significant. It gives no sample size and releases its reports only through a sign-up form. This research found no open dataset measuring prices across channels since the Act took effect.
Built with what exists
An ordinary web browser; a written collection protocol; a shared form or spreadsheet; Python for the analysis. No new AI model is needed.
First step
Write the protocol, pick 20 hotels in one city that sell on all four channels, and record one week of prices by hand.
Done looks like
At least 4,800 observations (100 hotels, four channels, 12 weekly snapshots) released under an open license, a published table of how often each channel is cheapest and by how much, and a second person re-checking 5% of observations against the screenshots with at least 98% agreement.