Why The Reservation Wars Are Now Three-Front: DoorDash vs Amex vs Booking

It is the first Friday of the year that feels like a working Friday, and I am in the home office reading the December fund letters that landed in the inbox between Christmas and New Year. The restaurant-tech-adjacent books are unanimous on one thing — none of them have a clean framework for what just happened in the reservations category. Two frame it as a DoorDash story. One frames it as an Amex story. One — and this one I read twice — does not mention Booking Holdings at all, which is wild given Booking still owns the largest reservations platform in the United States.

The framing those letters are missing, and the framing I want on the table heading into Q4 print season, is this: the reservations category is now a three-front war between DoorDash, Amex, and Booking. None of the three fronts moves in isolation. None is priced into the parent equities the way it should be. In my read — interpretation, not consensus — this triangle is the most under-priced battleground in restaurant equities right now. Q4 prints will be the first quarter you can read it in the numbers.

Three fronts in order, because the order matters.

Front one: DoorDash is buying the in-venue data layer

The DoorDash piece is the one most of the desk has at least skimmed. Headline: DoorDash agreed to acquire SevenRooms for approximately $1.2 billion in cash in May 2025 and closed the deal on June 13, 2025. (DoorDash IR)

The buy-side memos that circulated in the back half of 2025 mostly framed this as DoorDash buying a reservations platform to compete with OpenTable. That framing is shallow. SevenRooms is, primarily, a guest-CRM and marketing platform — reservations is a feature of a larger suite. What DoorDash bought is the in-venue data layer to sit alongside its delivery data layer. The reservations platform is the cheapest way to acquire that data at scale.

The numbers: roughly 67% US food delivery share for DoorDash, per the company’s own investor materials cited in the Deliverect 2025 State of Food Delivery report. SevenRooms sits at roughly 13,000 venues globally, weighted toward enterprise hotel F&B, multi-unit fine dining, and global hospitality groups — Marriott, MGM, Mandarin Oriental, Wolfgang Puck, USHG, Nobu, Dishoom. That customer mix justifies the price.

The integration moved faster than the M&A press cycle expected. By September DoorDash had launched “Going Out” in Miami and New York — in-app reservations, in-store rewards, DashPass-tied benefits including “exclusive tables,” and zero cover fees on reservations sourced through the DoorDash marketplace. The zero-cover-fee line is the operator-facing weapon. OpenTable charges per-cover fees on top of monthly platform fees. DoorDash is undercutting OpenTable’s per-cover economics at the marketplace layer and recouping the spread through DashPass attachment and cross-sell into delivery and ads.

I covered the deal mechanics and the operator playbook in the May Bottom Line (post 15). Short version: DoorDash is no longer a delivery company that owns a reservations platform. It is a horizontal commerce platform that touches every revenue line in a restaurant, with SevenRooms as the keystone of the in-venue data layer.

What this means for the equity. DoorDash trades on delivery-economics multiples — EV/Sales in the high single digits, with the ads business and the international footprint as the two pieces the sell-side models. The reservations piece is mostly hand-waved. If SevenRooms-attached venues lift attach rates on the rest of the commerce stack — ads, ordering, logistics — the acquisition is meaningfully accretive to unit economics on every other line. That second-order math is what the buy-side has not yet priced.

Front two: Amex is building a dining operating system, not a booking app

The Amex front is the one the desk has the least conviction on, partly because the public information has come in chunks rather than a single deal announcement, and partly because Amex doesn’t break out Resy in segment reporting the way DoorDash has to break out commerce platform investments.

Here is what is on the record. Amex acquired Resy in 2019. Squarespace acquired Tock in 2021 and sold it to Amex in 2024 for approximately $400 million. The current plan, as reported in 2025, is to merge Resy and Tock into a combined platform of approximately 25,000 venues, with the merger scheduled for summer 2026. (Entrepreneur)

The framing that matters: Amex is not building a reservations business. Amex is building a dining operating system that uses reservations as the front door and the card network as the monetization engine. Tock specifically gives Amex the prepaid-ticketing and tasting-menu inventory layer that Resy never had — the high-end, reservation-required, prepay-or-deposit segment where average ticket is meaningfully higher than the broader Resy book. Juliet covered the product roadmap in post 3 — Toast handheld integration, Tock fold-in, the guest record as the integration substrate. Worth a re-read.

The strategic logic is the network effect. Amex cardholders skew higher-spend, the Resy book skews higher-end urban, the Tock book skews destination-restaurant and reservation-required-prepay. Cross-pollinate those three pools and you get a closed-loop dining stack where the card network monetizes the reservation, the reservation drives venue acquisition, and venue acquisition drives card-network spend at the restaurant. The flywheel is real if it spins.

The Q4 read on whether it is spinning. Amex reports Q4 2025 earnings on January 30. That print will be the first quarter where you can plausibly see the early Resy/Tock effect in the dining vertical of card spend. I will note that the dining vertical has been the fastest-growing card-spend category at Amex for several quarters running, and that the Q4 print is the natural place for Squeri to put a marker down on the Resy strategy if he has one to plant. Worth listening live.

What this means for the equity. Amex trades on credit-card-network economics. The Resy/Tock investment is essentially free option value at the current valuation — $400 million for Tock plus an undisclosed but presumably similar number for the original Resy acquisition, sub-$1B of cost basis inside a $200B-plus market cap. If the dining-OS thesis works, the optionality is asymmetric to the upside. If it doesn’t, the write-down is rounding error. That asymmetry is what the desk has not yet priced.

Front three: Booking is the largest player and the most under-discussed

The Booking front is the one I want to spend the most time on, because it is the one the December fund letters mostly skipped.

OpenTable, owned by Booking Holdings, manages reservations for approximately 60,000 restaurants. That is the largest US reservations book by a meaningful margin — roughly 2.4x the planned post-merger Resy/Tock book and 4.6x the SevenRooms book DoorDash now owns. OpenTable has been the incumbent platform in the US since 1998. (Entrepreneur)

The framing I want to push back on. Most of the trade coverage treats OpenTable as an aging incumbent being eaten from both sides by DoorDash and Amex. That framing is half right. OpenTable is the incumbent and it is being attacked from both sides. But OpenTable is also part of a parent with its own platform strategy, its own AI roadmap, and a balance sheet that can fund a defensive response at a scale neither DoorDash nor Amex would casually match.

Marco walked through the Booking Holdings AI strategy in detail in post 2. Short version: OpenTable is increasingly being treated by the Booking parent as a restaurant-shaped lab for the same AI-agent and trip-planning playbook running across Booking.com, Priceline, and Agoda. The reservations stack is the substrate. The AI layer is the strategic asset.

The trade press through the back half of 2025 was openly calling the category a “reservation war.” The signal that most underlines incumbent concern: OpenTable has been moving on contract language that would limit operators’ ability to run a parallel platform. That is an incumbent-defending-the-moat move you only make when you have seen meaningful churn risk. Watch this through Q1 — the form the exclusivity language takes will tell you how much pressure OpenTable is actually under.

What this means for the equity. Booking Holdings trades on travel-platform economics, and OpenTable inside Booking is a small fraction of the consolidated revenue line. The market is pricing OpenTable as a stable contributor inside a larger travel thesis. If the AI-layer strategy works, OpenTable becomes a meaningful product-experiment line inside Booking’s agent-economy roadmap. If the incumbent moat erodes faster than the AI-layer strategy compounds, OpenTable becomes a slow-growth drag. Either outcome is asymmetric to the current valuation in ways the sell-side is not modeling.

Plain English on the multiples

A brief detour on the boring numbers.

EV/Sales is enterprise value divided by sales — the multiple sell-siders use to compare scaling businesses across categories. DoorDash, Amex, and Booking trade on three different regimes — commerce-platform, card-network, travel-platform — and the reservations business gets folded into a different parent every time. Standalone, a reservations book would probably trade at a SaaS multiple in the high single digits to low double digits. Inside DoorDash it gets the commerce-platform multiple; inside Amex the card-network multiple; inside Booking the travel-platform multiple. None is correct for the underlying reservations business. That mispricing is the opportunity.

Take rates are the percentage of GMV the platform keeps as revenue. On reservations the take rate is small per cover — single-digit dollars in OpenTable’s traditional model, less in Resy’s monthly-fee model, less still in DoorDash’s zero-cover-fee marketplace. The take rate looks unattractive on its own. It becomes attractive when it is the acquisition cost for the data layer that monetizes everywhere else — cards, ads, delivery, hotels.

Network effects on the reservations side are weaker than the trade press claims — a restaurant on OpenTable does not become more valuable when the next restaurant joins, because diners book one restaurant at a time. The real network effects sit on the consumer side and the data side. Amex cardholders, DashPass subscribers, Booking.com travelers — those are the demand pools the parents are routing through their reservation books. The reservations platform is the conversion layer. The consumer pool is the moat.

What to watch on Q4 prints

Three things on the calendar through earnings season. None are headline-event predictions — they are the second-order tells that will tell you which front is winning.

One. Amex Q4 print, January 30. Listen for any segment commentary on Resy-restaurant card spend or the Tock integration timeline. If Squeri puts a number on Resy specifically — Q4 spend at Resy restaurants versus overall restaurant spend, year-on-year growth on the Resy book — that is the first quarter the dining-OS thesis becomes legible in the financials. If he doesn’t, that is also a signal — that the integration story is on track for the summer 2026 merger and the financial read comes later in the year.

Two. DoorDash Q4 print, February. Reservations won’t show up as a revenue line yet — SevenRooms is rolled into the broader Commerce Platform line. Listen for attach-rate commentary. If management talks about advertisers attaching at higher rates inside SevenRooms-served venues, or DashPass renewal rates being higher in Going Out launch markets, the second-order math is working. If they don’t, the integration is still pre-revenue from a monetization standpoint.

Three. Booking Holdings Q4 print, late February. OpenTable will get one or two mentions in the prepared remarks. Listen for any commentary on AI-agent traffic into OpenTable — surfacing inside Booking.com’s trip-planner, or external agents routing reservations into OpenTable APIs at scale. That is the AI-layer strategy Marco flagged in post 2 and the leading indicator on whether OpenTable is becoming a Booking-wide AI product or staying a standalone reservations incumbent.

The interpretation

Now the part where I take a position.

My read — and this is interpretation, not consensus — is that the three-front structure is the most under-priced battleground in restaurant equities right now. Not because any single front is dramatically mispriced on its own, but because the cross-front dynamic is not modeled by the sell-side and is not in the buy-side memos. The dynamic: each parent is pulling the reservations book into a different monetization model, and the parent that monetizes most efficiently per cover ends up with the durable category position. DoorDash’s model is delivery and ads. Amex’s is card spend. Booking’s is travel adjacency and AI agents. Three different cash registers, one shared customer.

The under-priced part is that all three parents have the balance sheet, the demand pool, and the strategic conviction to run their playbooks for years before any of them concedes. This is not a winner-take-all market. It is a multi-incumbent market where each incumbent has a different reason to win. That structure is unusual in software. The sell-side has not figured out how to model unusual structures.

Risk to the interpretation. If one of the three parents pulls back materially in 2026 — strategic review at Amex, capital-return pivot at Booking, delivery-margin shock at DoorDash — the three-front structure collapses to two-front and the thesis weakens. None look likely on the public information today. All three are worth watching.

Operator-side read, short version: multi-home while it is cheap, document your dependencies on each platform’s demand graph, and keep a recent export of your guest CRM on your own systems. The platforms are competing for the data layer. Your job is to keep yours.

Three fronts. One battle. Q4 prints start in three weeks and will be the first quarter you can read the structure in the numbers rather than the press releases. I will be reading them line by line.

— Marcus runs The Bottom Line and gets the deal flow before the brokers. Tips: [email protected].