The Reservations Endgame: A Three-Phase Framework for Operators Picking Between DoorDash/SevenRooms, OpenTable, and AmEx Resy Through 2026

It is December 26, and my desk is the only thing in the apartment that does not smell like a roast. Christmas crockery is still drying in the rack. The tree is shedding. My partner is watching something with subtitles. And I am here, with three browser tabs open, trying to draw the diagram I have been threatening to draw since June.

Tab one is the DoorDash press release announcing the SevenRooms acquisition, now six months closed. Tab two is the Restaurant Dive piece on OpenTable blocking data sharing with competitors, the dispute that quietly redrew the contract terms of the entire category. Tab three is the TechCrunch write-up of DoorDash’s Zesty launch from ten days ago, which on the surface is a discovery app and on the inside is the front door of a new acquisition funnel that ends at a SevenRooms host stand.

The thing I have been wanting to write is a framework. Not a roundup, not a buyer’s guide, not another “five platforms compared” listicle. A framework. Specifically: how the three-platform reservation oligopoly that has solidified this year — DoorDash/SevenRooms, OpenTable (Booking Holdings), and AmEx Resy/Tock — is going to resolve for operators over the next twelve months.

The contrarian thesis, stated plainly, is this: the three-platform oligopoly resolves differently for three operator segments. It does not resolve into a single winner. It does not resolve into a fragmented long tail. It resolves into three correct answers, each one wrong for two-thirds of the market. Independents win on data ownership. Upscale wins on premium-card audience. Multi-unit chains win on distribution. The mistake almost every consultant, vendor sales deck, and trade-press explainer is making is treating “the reservation platform decision” as one decision. It is three. The right answer is segment-conditional and largely deterministic once you know which segment you are in.

I am calling this the Four Tables framework, even though we are only seating three of them today. The fourth table is Toast Tables and the POS-native fourth path, which I will come back to. For now, three tables, three platforms, three segment-winners. This is the first draft. We will iterate on it through 2026. It will be wrong in places. It is already wrong in places. But the operators I respect most are making this decision right now, in January planning cycles, and they deserve a framework better than “well, it depends.”

It depends, but it depends on something specific. Let me show you.

H2: What the oligopoly actually looks like at year-end

Before we get to the framework, the lay of the land, because half of the bad takes I have read this month rest on bad numbers.

DoorDash closed its acquisition of SevenRooms on June 13, 2025 — six months ago, not the abstraction it still feels like in trade press. SevenRooms is no longer an independent company. It is a wholly-owned unit inside DoorDash, with all the strategic implications that entails. The SevenRooms blog still publishes vendor-comparison content as if nothing has changed, and that is a clue about positioning more than product — DoorDash wants SevenRooms to read as a CRM-first platform with a reservation layer, not as a reservation platform with a CRM bolted on. The data-ownership story is the spine.

OpenTable, owned by Booking Holdings, sits at roughly 60,000 restaurants globally and processes around 1.9 billion seated diners per year. Those are the headline numbers and they are real. OpenTable is the distribution platform — the largest active diner audience, the deepest search-page real estate, the longest-running brand recognition. It is also the platform that, this fall, moved to block partner data-sharing arrangements in a dispute centered on SevenRooms-resident guest data. The headline read like a routine contract dispute. It is not. It is OpenTable codifying contract-lockup-as-policy — the recognition that, in an oligopoly, the asset is not the technology but the data tether to the diner.

AmEx Resy, now combined with Tock, is the third pillar. Resy was acquired by American Express in 2019. Tock was acquired by AmEx in 2024. Post-merger consolidation has put the combined platform at roughly 25,000 restaurants (a forthcoming figure I am treating as anticipated, since the precise post-integration count has not been published, but the trajectory is clear). The strategic positioning is unambiguous: this is the premium-card audience platform. The Resy diner skews higher-income, higher-spend, and is co-marketed against AmEx Platinum and Gold benefits. Tock brings the prepaid-deposit, tasting-menu, fine-dining tail.

Underneath the oligopoly is Yelp Guest Manager, which Bistrochat’s market-data piece pegged at roughly 14% market share in 2024. That number is meaningful but declining-by-relevance. Yelp’s reservations product is increasingly a cost-leader play, not a premium one. It is the floor of the market, not the ceiling, and operators choosing Yelp Guest Manager in 2026 are mostly choosing it because they already pay for Yelp ads and the bundle math works, not because of platform-specific strengths.

And then there is Toast Tables, the POS-native fourth option, which I will come back to in section seven because it changes the framework rather than fitting inside it.

So: three platforms, three positionings, one challenger from below. That is the board.

H2: The framework — data-ownership, premium-card, distribution

Here is the diagram I have been trying to draw.

Every reservation platform decision, at root, is a decision about which of three goods you are buying. You are buying guest data you can own and use (data-ownership). You are buying access to a curated premium audience (premium-card). You are buying breadth of inbound discovery (distribution). All three platforms claim to offer all three. None of them do. Each platform is genuinely excellent at one of the three and merely adequate at the other two. Pretending otherwise is what produces the bad procurement decisions I see in my inbox every week.

Data-ownership means: when a diner books with you, the guest profile, the dining history, the preferences, the spend pattern, and the marketing rights are yours, exportable, addressable, and usable to drive repeat visits and direct bookings. The platform is a host stand, not a landlord.

Premium-card means: when a diner books with you, they are statistically likely to be a high-spend customer routed to you via a co-marketed card-benefit program, with a willingness to pay deposits, sit for tasting menus, and convert on cross-sell. The platform is an audience filter.

Distribution means: when a hungry diner two miles away opens a discovery surface, your restaurant appears in front of them with one tap to book. The platform is a search engine for available tables.

The three goods are not orthogonal. They overlap. But they trade off, because they require different incentive structures from the platform. A platform optimized for data-ownership will lean toward operator-side power, low marketing-list lockup, and CRM tooling. A platform optimized for premium-card will lean toward consumer-side power, audience curation, and tight co-marketing. A platform optimized for distribution will lean toward platform-side power, search-result control, and consumer acquisition spend. You cannot maximize all three from the same balance sheet.

The framework follows: identify which of the three goods is the highest-marginal-revenue good for your segment. Pick the platform that is excellent at that good. Accept that you will be merely adequate on the other two. Do not pay for the merely-adequate good. That is the entire framework. The rest is mapping segments to goods.

This is a first-draft framework. It will get refined. A forthcoming May framework essay — call it the Four Margins piece — will introduce a margin-side complement to this audience-side cut, looking at where the platform-fee economics actually land across the same three segments. For now, audience-side cut, three segments, three answers.

H2: Independents — why SevenRooms data-ownership wins

The independent operator — single-unit, owner-operated, $1.5M-$8M annual revenue, often chef-driven — should pick DoorDash/SevenRooms. The reason is data-ownership, and the reason data-ownership is the highest-marginal-revenue good for this segment is that the independent’s only durable competitive advantage is repeat-visit yield from a small, loyal guest base.

Let me explain that more carefully, because it is the crux of why so many independents are picking wrong.

An independent restaurant doing $4M a year on 80 seats with two services six nights a week is, in any given year, serving on the order of 30,000-60,000 covers. Of those covers, the top decile — call it 3,000-6,000 covers — accounts for somewhere between 40% and 60% of repeat-visit revenue. The economics of an independent are almost entirely the economics of the top decile of guests. If you know who they are, what they order, what they pay, when they last came in, what wine they like, what they noted in the booking comments — and if you can market to them directly without intermediation — you can systematically increase the top-decile yield by 15-25% over twelve months. That is the entire profit and loss of an independent restaurant, won or lost on CRM.

SevenRooms, as a product, was built around that fact. Its core value proposition is the guest profile and the marketing layer that sits on top of it. The Mina Group case study that SevenRooms cites — over a million guest profiles across the Michael Mina portfolio — is a marquee number, but the principle scales down to single units. The architecture of the platform assumes the operator owns the data and the relationship; the booking is the wrapper, not the asset.

Now, DoorDash’s ownership of SevenRooms is the thing that has made independents hesitate, and I understand why. The fear is that DoorDash will weaponize SevenRooms guest data for marketplace marketing — that your loyal regular will get pinged with a DoorDash push notification for a competitor down the street. That fear is real but mostly wrong. The DoorDash strategy as I read it, and as the Zesty launch confirms, is not to extract data out of SevenRooms back into the marketplace. It is to push discovery and acquisition INTO the SevenRooms host stand. Zesty is the front door, SevenRooms is the back office, and the asset that DoorDash is monetizing is the increased acquisition flow into your existing CRM. The bet is that operators will accept the marketplace adjacency because the acquisition lift makes the data-ownership package worth more, not less.

I read the recent Zesty coverage the same way. Zesty is not a competing reservations product. Zesty is a discovery-funnel feeder for restaurants that already use SevenRooms for the booking infrastructure. The independents that pick SevenRooms in 2026 are essentially buying a CRM with a free acquisition channel attached.

The OpenTable counter-move on this point — the data-sharing block — is the tell. OpenTable does not block partner data-sharing arrangements when its own data-ownership story is strong. It blocks them when SevenRooms is winning on data-ownership and OpenTable needs contract-lockup-as-policy to plug the leak. The dispute itself is the evidence that SevenRooms has the better hand for the data-ownership-prioritizing operator.

The procurement implication for independents: buy SevenRooms, run the marketing layer hard, accept the DoorDash marketplace adjacency, and do not pay for OpenTable distribution unless you have specific evidence that your local OpenTable inbound is material. (For most independents, it is not. For a few — neighborhood-destination independents in OpenTable-heavy metros — it is. Run the test.)

H2: Upscale — why AmEx Resy + Tock premium-card wins

The upscale operator — fine dining, tasting menu, $200+ checks, $5M-$30M revenue often across a small portfolio — should pick AmEx Resy + Tock. The reason is premium-card audience, and the reason the premium-card audience is the highest-marginal-revenue good for this segment is that upscale operators have a deposit-and-yield problem, not a top-of-funnel problem.

The upscale segment does not need more reservation requests. It needs better reservation requests. A 14-seat tasting menu at $350 a head with a 7pm seating has roughly 4,900 in revenue at full house. The marginal cost of a no-show or a downsold party is enormous. The marginal value of pre-payment, deposit-collection, and a guest segment that converts on prix-fixe is the difference between profit and loss on the night.

This is exactly the goods bundle that Resy + Tock has been built around. Tock invented the prepaid-deposit fine-dining model and made it default. Resy plus AmEx Platinum and Gold co-marketing produces a diner audience that skews higher-income, higher-spend, and — critically — accustomed to and comfortable with the upscale booking experience. The post-merger combined platform at approximately 25,000 restaurants (a number I am treating as forthcoming since the precise integration count is not yet public) is small relative to OpenTable, but the small-relative is the entire point. You do not want OpenTable’s 60,000-restaurant audience in front of your tasting menu; the conversion rate on the wrong-audience inbound is brutal and you absorb the labor cost of disqualifying it.

I will write more about AmEx Resy strategy in an upcoming spring piece, but the short version is that the platform’s economic model is the inverse of OpenTable’s. OpenTable monetizes the breadth of the diner audience and charges the restaurant for inbound. Resy effectively monetizes the AmEx cardholder relationship and treats the restaurant placement as a benefit-augmentation channel. That asymmetry means Resy has the budget and the strategic motive to keep curating the audience tighter, not broader.

For upscale operators, the trade-off against SevenRooms is real and worth naming. SevenRooms data-ownership tooling is genuinely stronger than Resy’s. If you are a 30-unit luxury group with serious CRM ambitions, you might rationally pick SevenRooms and accept that you will need to build the deposit-collection and the premium-audience-curation layers yourself or with adjacent tools. But for the single-unit or small-group upscale operator without a dedicated marketing team, the Resy + Tock package solves the highest-marginal-revenue problem (audience quality, deposit collection, no-show reduction) better than SevenRooms does.

The OpenTable bet for upscale is the weakest of the three. OpenTable’s audience is too broad, the co-marketing too undifferentiated, and the contract terms — especially after the data-sharing dispute — too platform-favorable. An upscale operator on OpenTable in 2026 is, in most cases, paying for distribution they do not need and not paying for audience curation they do.

The procurement implication for upscale: buy Resy + Tock, run the deposit collection hard, run the AmEx co-marketing harder, and do not buy SevenRooms unless you have a portfolio-scale CRM mandate that justifies the build-out alongside.

H2: Multi-unit chains — why OpenTable distribution wins

The multi-unit chain — 20+ units, regional or national footprint, often franchised, $50M-$2B revenue — should pick OpenTable. The reason is distribution, and the reason distribution is the highest-marginal-revenue good for this segment is that chains have an inbound-fill problem, not a data-ownership problem and not an audience-curation problem.

Let me say more.

A multi-unit chain restaurant doing $4M-$8M per unit is not running on top-decile-guest economics the way an independent is. It is running on cover-count economics. The chain’s brand handles much of the audience-curation work; the chef and the seasonal menu handle little of it. The marginal revenue question is not “did the right diner book?” but “did a diner book at all, in the off-peak window, in the third-best unit, on a Tuesday in March?” That is a distribution question, full stop.

OpenTable’s 60,000-restaurant network and 1.9 billion annual seated diners is overwhelmingly built for exactly this use case. The OpenTable consumer app and OpenTable search-page real estate is where casual-discovery inbound aggregates. Chains have invested heavily in OpenTable integration, OpenTable promoted-placement, and OpenTable yield-management tooling, and that investment compounds. Switching costs are not nominal — they are operational and analytical and ten years deep.

The data-ownership argument that I made for independents inverts for chains. A chain has its own first-party data infrastructure, often built on Toast or Olo or a bespoke CDP, and does not need a reservation platform to be the data spine. The reservation platform is a single channel — important, but one of many. The chain does not want the reservation platform to lock the data; it wants the reservation platform to fill the seats. OpenTable, despite the data-sharing dispute, is structurally the better fill engine because that is what its consumer-side incentives optimize for.

The Resy bet for chains is wrong-segment. Resy’s premium-card curation is the value, and chains do not particularly need it. The AmEx co-marketing flows in a chain’s favor only in a handful of premium-positioned concepts within larger portfolios. For the body of the chain market — fast-casual and casual-dining and approachable-premium — Resy’s audience is overprecise and underdistributed.

The SevenRooms bet for chains is more interesting and more partial. Some chains will run SevenRooms for the data and OpenTable for the distribution, in a deliberate dual-platform setup. That is a defensible architecture, especially for chains with sophisticated CRM ambitions. But the data-sharing dispute makes it harder to run cleanly. OpenTable’s contract-lockup-as-policy is, in part, designed precisely to make this dual-platform architecture more painful to maintain. Chains will spend 2026 testing whether dual-platform is worth the friction. My read is that for most chains, it is not — pick OpenTable as the primary, accept that the data layer will be merely adequate, and concentrate CRM investment in the first-party stack instead.

The procurement implication for chains: buy OpenTable, negotiate the contract aggressively (the data-sharing dispute strengthens the operator’s hand in negotiation more than the platform’s, in my read), and do not pay for SevenRooms or Resy unless you have unit-level economics that specifically justify either.

H2: Toast Tables and the POS-native fourth path

I promised I would come back to Toast Tables, and here is where I do, because it changes the framework rather than fitting inside it.

Toast Tables is the reservation product inside the Toast POS platform. It is the fourth table. The reason it changes the framework is that Toast Tables does not compete on data-ownership, premium-card audience, or distribution. It competes on POS-native integration, which is a fourth good that the three-platform framework does not capture.

POS-native means: the reservation platform is the same platform that runs the front-of-house POS, the back-of-house operations, the inventory, and increasingly the payroll and tip distribution. The marginal advantage is operational, not audience-based. The right cover lands at the right table, the guest history matches the check history, the staff scheduling matches the booking pace, and the data flows are unified by virtue of being the same data.

For the operator already on Toast — which is a very large fraction of the independent and small-chain segments — Toast Tables is a credible fourth option specifically because it eliminates the integration tax that the three-platform options all impose. The data-ownership story is partial (Toast does own the booking and the dining history, but its CRM tooling is less mature than SevenRooms). The audience-curation story is nonexistent. The distribution story is weak. But the operational story is uniquely strong, and for the operator whose highest-marginal-revenue lever is operational efficiency rather than CRM or curation or distribution, Toast Tables can be the right answer.

I will note that the POS-native path also opens a different question — about voice agents and the operations-side AI layer that sits on top of reservations and POS — and I am drafting an upcoming May framework piece on voice-agent maturity that is going to bear on this. The short version is that POS-native platforms have a structural advantage in deploying voice-agent reservation taking, because the agent has unified data access. That advantage compounds the operational case for Toast Tables.

For now, Toast Tables is the fourth table at the framework dinner. It is not the right answer for most operators, but it is the right answer for some, and the framework needs to acknowledge it before the framework is complete. We will see in a year whether Toast Tables grows fast enough to warrant being a peer of the three platforms or remains a credible fourth option for the POS-native segment.

H2: The procurement scorecard

Here is the scorecard, which is what I have been trying to give the operators emailing me with January planning questions. Use it as a starting point, not a verdict.

For each platform, score the four goods on a 1-5 scale, where 5 is “excellent and category-leading” and 1 is “merely present in the product.” Then weight the goods by your segment-conditional priority. The platform with the highest weighted score is your answer.

Goods:

1. Data-ownership (guest profile, marketing rights, export, CRM tooling)
2. Premium-card audience (curation, co-marketing, deposit support)
3. Distribution (consumer app, search-page presence, inbound volume)
4. POS-native operations (integration depth, unified data flow)

Platforms:

	Data-own	Premium-card	Distribution	POS-native
DoorDash/SevenRooms	5	2	3	2
OpenTable	2	2	5	2
AmEx Resy + Tock	3	5	2	2
Toast Tables	3	1	1	5

Segment weightings:

• Independent: data-own 50%, premium-card 10%, distribution 30%, POS-native 10%
• Upscale: data-own 25%, premium-card 50%, distribution 15%, POS-native 10%
• Multi-unit chain: data-own 15%, premium-card 5%, distribution 65%, POS-native 15%

Run those weightings against the platform scores and you get:

• Independent: SevenRooms 3.9, OpenTable 2.7, Resy 2.7, Toast 2.3
• Upscale: SevenRooms 2.8, OpenTable 2.6, Resy 4.1, Toast 1.6
• Multi-unit: SevenRooms 3.5, OpenTable 4.4, Resy 2.3, Toast 1.8

Three segments, three answers. The framework is dispositive when you fill in the weighting honestly. The mistake operators make is starting with the platform and reverse-engineering the weighting. Start with the segment, weight the goods, then pick the platform.

The numbers above are first-draft. I will publish refined scoring in a forthcoming May Pass piece on OpenTable and Booking Holdings and in a forthcoming May vendor comparison on SevenRooms versus Tablecheck, both of which will tighten the platform scores with more granular sub-criteria. The shape of the conclusion will not change. The numbers will.

H2: What I am betting on through Q4 2026

The three-segment-three-answer framework has a forward-looking implication, which is that the segment-paths solidify through 2026 rather than dissolve.

The competing forecast — the one I see most often in trade press — is that the reservation oligopoly converges. That one platform pulls ahead. That the differences I have just spent four thousand words describing get arbitraged away as each platform clones the others’ strengths. I do not believe it, for three structural reasons.

First, the strategic logic of each platform’s parent company runs in opposite directions. DoorDash’s interest in SevenRooms is to compound the marketplace acquisition flywheel with high-quality first-party data. AmEx’s interest in Resy is to compound the card-benefit flywheel with curated dining inventory. Booking Holdings’ interest in OpenTable is to compound the global-travel-distribution flywheel with reservation inventory. The three parent companies have three different theories of the business, and the platforms inherit those theories. Convergence would require the parents to abandon their respective flywheels, which is not happening.

Second, the contract terms are already diverging. The OpenTable data-sharing dispute is the canary. OpenTable is doubling down on platform-favorable contract architecture. SevenRooms, under DoorDash, is doubling down on operator-favorable data terms. Resy under AmEx is doubling down on consumer-favorable curation. These are not surface positioning decisions; they are the terms of the underlying licenses, and they are getting harder to undo as they accrete.

Third, the operator segments are getting better at self-sorting. The independents I talk to are increasingly fluent in CRM-first thinking. The upscale operators are increasingly fluent in deposit-and-yield thinking. The multi-unit chains are increasingly fluent in distribution-and-fill thinking. The market is educating itself, and an educated market sorts faster.

By Q4 2026, I expect the segment-paths to be largely locked in. SevenRooms will own the independent CRM-first segment. Resy + Tock will own the upscale audience-curation segment. OpenTable will own the multi-unit distribution segment. Toast Tables will hold a meaningful POS-native niche. Yelp Guest Manager will continue as the cost-leader floor. The fights between the platforms in 2026 will be at the segment boundaries — independents flirting with Resy, upscale flirting with SevenRooms, chains flirting with dual-platform — and most of those flirtations will not consummate, because the segment-conditional economics will pull the operator back to the right answer.

I wrote three weeks ago — in the December 3 reservations Vibe Check — that the year-end story was going to look more like settlement than disruption. I stand by that. The framework I have drawn here is the settlement, formalized. We will iterate, refine, and republish through 2026. The first draft is on the desk.

It is now nearly 11pm. The crockery is dry. The tree has shed half a needle pile onto the rug. The subtitled show is over. The framework is on the page. I will look at it in the morning and decide whether I still believe it.

I think I will. The thing about a framework is that you only know it works when the operators it is meant to serve can use it to make a decision without needing to call you. If you read this, and you know your segment, and you know which good is your highest-marginal-revenue good, and you can run the scorecard yourself — then this draft has done its job. If not, write to me and we will iterate.

The endgame is not one platform. The endgame is three answers, each one right for the operator who needs it, each one wrong for the operators who do not. That is what an oligopoly looks like when it resolves cleanly. Whether 2026 actually resolves it cleanly is the bet.

I think it does.

— Eitan is editor-in-chief of TableTransfers. Tips: [email protected].