How the travel industry's plumbing actually works — and why Accomy is building the infrastructure layer that sits underneath it all.
Booking a hotel room or flight seems simple. Underneath, your request passes through up to 5 layers of middlemen — each taking a cut. Understanding this plumbing is essential to understanding what Accomy is doing.
Hotels and airlines have rooms and seats to sell. But they can't reach every buyer directly, so a chain of middlemen connects supply to demand. Each middleman adds cost. Accomy's play: collapse those layers into one modern platform.
| Layer | Who | What They Do | How They Get Paid |
|---|---|---|---|
| 1. Suppliers | Airlines, Hotels | Own the actual inventory (seats, rooms) | Set base prices |
| 2. GDS / Switch | Amadeus, Sabre, Travelport | Route inventory data to sellers (like SWIFT for travel) | $2-6 per booking segment |
| 3. Aggregators | Hotelbeds, Duffel, Kiwi | Bundle supply from many sources into one feed | Markup on wholesale rates |
| 4. Distributors | Booking.com, Expedia, Navan | Sell to consumers or companies | 15-25% commission or markup |
| 5. End Buyers | You and me | Book and travel | Pay the final retail price |
Airlines publish fares through four channels, each with different costs and tradeoffs:
Built in the 1960s-70s. Amadeus (~43% share), Sabre (~30%), Travelport (~20%). Airlines pay $2-6 per segment — a round trip with connection = $8-24 in fees. That's 4-12% of a $200 ticket just for distribution.
New IATA standard letting airlines distribute fares directly via API. Richer content, dynamic pricing, lower costs. Lufthansa, AA, and AF-KLM are pushing hardest. But inconsistent implementation and immature post-booking support.
LCCs like Ryanair and EasyJet have proprietary REST APIs. Total control for the airline, but limited distribution reach. Only large partners integrate directly.
Google Flights, Skyscanner, Kayak compare prices across all sources and send users to the booking site. They don't hold inventory — they're comparison engines monetizing via clicks.
Hotel distribution is far more fragmented than flights. A single room might be sold through 15+ channels simultaneously — with 5 different prices on 5 different platforms.
| System | Plain English | Key Players |
|---|---|---|
| PMS | Hotel's operating system — check-in/out, housekeeping, room status. The source of truth. | Opera (Oracle), Mews, Cloudbeds |
| CRS | Manages rates and inventory across all sales channels. The commercial brain. | SynXis (Sabre), TravelClick (Amadeus) |
| Channel Manager | Pushes rates to Booking.com, Expedia, bedbanks all at once. Prevents double-booking. | SiteMinder (dominant), RateGain, D-EDGE |
| RMS | Sets dynamic pricing based on demand, competition, and occupancy goals. | IDeaS (dominant in upscale), Duetto |
| Channel | How It Works | What Hotels Pay |
|---|---|---|
| Direct (Brand.com) | Hotel's own website. Best margin. | Just tech costs — no middleman. |
| OTAs | Booking.com (~28% global share), Expedia, Agoda | 15-25% commission per stay |
| Bedbanks (B2B) | Buy rooms at wholesale "net rates," resell to travel agents and other distributors | Hotel gets paid upfront at net rate. Bedbank marks up 15-30%. |
| GDS | For corporate travel bookings via TMCs | 10-20% commission |
| Meta | Google Hotels, Trivago, Kayak send users to booking source | Pay per click |
~180K properties. Supplies 60K+ travel intermediaries worldwide. The wholesale plumbing that most travelers never see.
ASX-listed. Growing fast in Asia-Pacific and Middle East. Main challenger to Hotelbeds.
900K+ properties. Key player for China inbound/outbound and the Trip.com ecosystem.
Israeli startup enabling direct hotel-to-distributor connections — cutting out the bedbank. "Stripe for hotel distribution."
Accomy isn't an OTA or just a corporate travel tool. It's a vertically integrated travel infrastructure stack — four distinct layers, each making money independently.
Layer 1 connects to all the travel supply sources. Layer 2 distributes that supply. Layer 3 sells it to companies as a corporate travel tool. Layer 4 (the endgame) lets anyone build travel products on top of Accomy's APIs — like Stripe did for payments.
Navan ($9.4B valuation) and TravelPerk ($1.4B) dominate corporate travel — but they focus on mid-market and enterprise. Companies with 5-200 employees are stuck between consumer tools (Booking.com + manual expense reports) and enterprise TMCs they can't afford. Accomy is the "easy button" for that gap.
Just as Stripe let any developer accept payments without building payment infrastructure, Accomy lets any company offer travel without building supply connections, booking engines, or servicing. Potential customers:
Revolut, Brex, Ramp embed travel booking into their corporate cards.
Deel, Remote, Oyster add business travel for distributed teams.
Plug in Accomy's supply instead of integrating 20 bedbanks individually.
Dynamically package using Accomy's real-time inventory and pricing.
Five layers of technology turn raw, messy supply data into a clean, bookable product.
The same hotel appears differently in every system. Accomy's core IP is recognizing they're the same property and merging the data:
| Hotelbeds | ID: 12345 · "Ritz London The" · 150 Piccadilly W1J 9BR |
| WebBeds | ID: RIT-LON-001 · "The Ritz, London" · 150 Piccadilly London |
| Amadeus | Chain: RC · Prop: LONRT · "THE RITZ LONDON" |
| Direct | ID: ACC-4421 · "Ritz Hotel - London Piccadilly" |
Solving this requires geo-matching, fuzzy name matching, address normalization, chain code mapping, and manual curation. It's built from millions of edge cases over years — extremely hard to replicate.
| Engine | What It Does | Why It's Hard |
|---|---|---|
| Search | Sends your query to 15+ suppliers at once, collects results, normalizes, ranks | 15 suppliers × 200ms each = 3 seconds if sequential. Must run in parallel with timeout management. |
| Pricing | Compares rates from all sources, applies markup, discounts, corporate rules | Real-time comparison across currencies, dynamic markup, rate parity compliance. |
| Booking | Takes payment, confirms with supplier, creates reservation | Two-phase problem: payment collected but supplier might reject. Needs robust retry/refund logic. |
| Cache | Stores results for repeated searches | Hotel availability changes every second. Stale cache = overbooking = disaster. |
| Availability | Final real-time check before confirming | Some suppliers have >5% failure rate at booking time. This pre-check prevents bad bookings. |
Infrastructure businesses trade at 2-5x higher multiples than application businesses. Here's why — and why Accomy is positioning as infrastructure rather than "just another travel app."
Apps compete for customers. Infrastructure gets paid by everyone — including the apps. Stripe doesn't care which shopping site wins. AWS doesn't care which startup scales. Accomy doesn't care which OTA or corporate tool succeeds — they all need travel supply.
| Company | Type | Valuation | Revenue Multiple | Why |
|---|---|---|---|---|
| Stripe | Payments infra | ~$50B+ | 15-20x | Every app that takes payments is a potential customer |
| PayPal | Payments app | ~$60B | 3-4x | Competes with every new fintech. Substitutable. |
| AWS | Cloud infra | ~$800B | 8x | Every software company is a potential customer |
| Amadeus | Travel infra (legacy) | ~$30B | 6-8x | Every TMC, OTA, airline uses it. Deeply embedded. |
| Booking.com | Travel app | ~$140B | High | Anomaly — but spends $6B+/yr on Google Ads to keep it |
More suppliers makes the platform better for distributors. More distributors makes it better for suppliers. This two-sided flywheel is the hardest competitive advantage to replicate.
Once a company builds their booking flow on Accomy's API, ripping it out means re-engineering everything. Technical lock-in through API dependencies and data formats.
Every search and booking generates pricing intelligence that compounds over time. "Which supplier has the best rate for this hotel on this date?" — knowledge no new entrant can replicate from day one.
Once the infrastructure is built, serving the next customer costs almost nothing. Fixed platform cost, pure marginal revenue per API call. Operating leverage apps can't match.
An OTA grows by getting more travelers. Infrastructure grows every time anyone in travel builds a product — OTA, corporate tool, fintech, HR platform. The ecosystem is the TAM.
Multiple revenue streams across all four layers — from hotel margins to SaaS subscriptions to API transaction fees.
| Revenue Stream | How It Works | Range |
|---|---|---|
| Net rate margin | Buy wholesale, sell at markup — primary revenue | 8-25% margin |
| Commission | Hotel pays Accomy a % after the guest stays | 10-20% |
| Revenue Stream | How It Works | Range |
|---|---|---|
| GDS incentives | GDS companies pay Accomy for booking volume | $1-4 per segment |
| NDC incentives | Airlines pay for NDC adoption (cheaper than GDS for them) | $1-3 per booking |
| Service fees | Per-ticket fee charged to the booker | $5-25 per ticket |
| Ancillary commissions | Commission on upgrades, baggage, seat selection | 5-15% of ancillary |
| Revenue Stream | How It Works | Range |
|---|---|---|
| SaaS subscription | Monthly platform fee for corporate features | $5-15 per user/month |
| Booking fees | Per-booking charge on top of travel cost | $5-15 per booking |
| API search fees | Per-search or monthly tiers for API partners | $0.01-0.10 per search |
| API transaction fees | Per-booking when partner books via API | $2-10 or % of GMV |
How Accomy compares to the major players — and why none of them are doing exactly what Accomy is doing.
| Company | What They Are | Where They Overlap | Why Accomy Is Different |
|---|---|---|---|
| Navan ($9.4B) | Corporate TMC | Layer 3 competitor | App company, not infrastructure. No APIs. Targets large companies. Accomy targets SMEs + builds for others. |
| TravelPerk ($1.4B) | SME TMC | Most direct TMC competitor | Also app-layer. No infrastructure play. Their supply comes from third parties. |
| Booking.com ($140B+) | Consumer OTA | Hotel supply, B2B expanding | Demand engine spending $6B/yr on ads. Accomy is infrastructure — Booking.com could be a customer. |
| Amadeus ($30B) | Legacy GDS | Layer 4 incumbent | Airline-centric, 1960s architecture, expensive. Accomy is hotel-native, API-first, modern. |
| Hotelbeds | Bedbank | Both a supplier and competitor | Hotelbeds is a wholesaler. Accomy aggregates FROM Hotelbeds and builds tech layers on top. |
If Accomy executes over 5-10 years, it becomes the default infrastructure for travel commerce — the way Stripe became default for payments.
| Moat | Strength | What It Means | Years to Build |
|---|---|---|---|
| Supply Aggregation | Medium | 15+ sources with normalization. Replicable but takes years of engineering. | 2-3 |
| Normalization IP | Strong | Hotel mapping built from millions of edge cases. Tacit knowledge, hard to copy. | 3-5 |
| Two-Sided Network | Strong | Suppliers + distributors. Classic marketplace lock-in. | 3-5 |
| API Lock-in | Strong | Once partners build on the API, switching means re-engineering everything. | Immediate |
| Pricing Intelligence | Very Strong | Historical data compounds with every search. The "Waze effect" — gets better with use. | 3-7 |
APIs without partners are infrastructure for nobody. Must reach minimum viable ecosystem before the flywheel turns. This is the existential risk.
Infrastructure is capital-intensive before it's profitable. Stripe took 14 years. Can Accomy fund the build long enough?
If Amadeus goes API-first and hotel-native, they have 40 years of relationships and $5B in revenue to outspend any startup.
Already doing B2B supply. A proper infrastructure API backed by 28M properties would be instantly formidable.
Building normalization + pricing + booking at scale is genuinely hard engineering. The required talent is expensive and scarce.
Hotels increasingly monitor where their rates end up. Too much redistribution and hotels may restrict access.
Step-by-step: what actually happens when someone books a hotel, a flight, or uses the API.
POST /api/v1/hotels/search with destination, dates, and guests. Accomy authenticates, rate limits, logs for billing.POST /api/v1/bookings. Accomy validates, books with underlying supplier, returns confirmation.Accomy helps companies book travel for their employees — flights and hotels. Think Booking.com, but built for businesses. It handles approvals, budgets, and invoicing so companies don't have to manage travel manually.
Accomy connects to dozens of travel supply sources (hotel wholesalers, airline systems, direct contracts), normalizes it all into one system, and distributes it through its own corporate platform plus APIs that other companies can plug into. Revenue comes from hotel margins, flight fees, and SaaS subscriptions.
Accomy is a vertically integrated travel infrastructure stack with four monetizable layers: supply aggregation, distribution infrastructure, a corporate TMC platform, and platform-as-a-service APIs. The TMC generates cash; the infrastructure layer creates long-term moats through two-sided network effects, API lock-in, and compounding pricing intelligence. The strategic endgame: become the "Stripe of travel."
The bet is on travel distribution unbundling. Thesis: Amadeus/Sabre are overpriced and airline-centric. Hotels are underserved by GDS. NDC creates chaos that benefits aggregators. SME corporate travel is a $200B+ TAM with no dominant player. If Accomy reaches critical mass (100K+ properties, 300+ airlines, 50+ API partners), the two-sided network becomes self-reinforcing and unit economics improve with volume. The question: can they reach escape velocity before capital runs out, and before Amadeus modernizes or Booking.com B2B scales? High execution risk — but if they execute, this is a $1B+ outcome at 8-15x revenue multiples.