The Channel Cannibalization Index: Why Domino's Sits at 0–5% and Pizza Hut at 100–130%

Section 1 — Why delivery growth without an incrementality check is not growth

When a chained-foodservice operator reports that delivery revenue grew 30 percent in the most recent half-year, the figure does not yet describe whether the company grew. Delivery growth and incremental growth coincide only when the new orders represent net-new customers or net-new occasions — not when existing dine-in patrons have switched channels. Standard hospitality KPI frameworks — off-premise share, delivery revenue percentage, platform-dashboard metrics — measure channel volume. None of them measure incrementality. The Channel Cannibalization Index, CCI, fills the gap. It is a single percentage that captures what share of new delivery revenue came from displaced dine-in business. No comparable formula has been published for the hospitality sector; the closest neighbours sit in shopper-marketing ROI and yield-management literature, without direct translation to channel incrementality.

This brief documents CCI and its margin-adjusted companion CCI_be, validated against a five-case DACH backtest spanning the full zone range. Two hypotheses converge with high confidence: QSR delivery-native models and casual-dining models with dine-in heritage follow structurally different cannibalization patterns. A third — that implementation quality does not compensate for product-delivery incompatibility — is demonstrated in the Nordsee case.

Section 2 — The formula and the zones

The CCI base formula is arithmetically simple:

CCI (%) = max(0, −ΔDineIn) / ΔDelivery × 100

ΔDineIn denotes the change in same-store dine-in revenue across the six-month window post delivery launch versus the six months pre launch. ΔDelivery is the analogous delivery delta. The max(0, …) condition ensures that a growing dine-in segment does not produce a negative reading — when dine-in expands alongside delivery, no cannibalization is measurable by definition and CCI = 0. The six-month pre-and-post window allows for seasonal correction and is robust to short-cycle promotions. Same-store applies strictly: only units without new openings or refurbishments inside the measurement window qualify, since growth programs and refit shutdowns would distort the comparator systematically.

A pre-trend-adjusted variant — CCI_adj — is mandatory when dine-in was already declining at more than two percent per year before launch. The form is max(0, −(ΔDineIn − DineIn_6M_PreTrend)) / ΔDelivery × 100. Without the adjustment, a brand in pre-existing decline returns a CCI_raw that overstates channel substitution and misdirects the operating intervention.

The zone bands translate the formula into operating language:

Section 3 — CCI_be: where revenue expansion meets margin erosion

Volume metrics alone are insufficient because the two channels carry different unit economics. The margin-adjusted break-even is:

CCI_be = 1 − (NetDeliveryMargin / DineInMargin)

In a typical DACH configuration — a 15-percent dine-in margin and a 7-to-8-percent delivery net margin after a 25-to-30-percent aggregator commission — CCI_be sits at 47 percent. Every CCI value above 47 percent makes delivery expansion margin-negative even when consolidated revenue holds steady. Operators who calibrate channel decisions on revenue KPIs alone overlook this threshold systematically.

Aggregator dashboards report delivery revenue. They do not report the dine-in counter-movement, because that data sits in the POS system. A 65-percent CCI — the Nordsee case — is invisible in day-to-day operations: total revenue stable or rising, dine-in declines absorbed by delivery gains, dashboard order count growing. The problem becomes legible only at CCI_be calibration. At 47-percent break-even, every euro of delivery revenue substituting dine-in revenue produces a seven-point margin gap against the consolidated P&L. At 240 outlets averaging EUR 1.07m in revenue, the cumulative effect is systemic. The cost-side complement — the 7 percent and 19 percent VAT split that separates dine-in from take-away inside the same store — is documented separately The DACH Operating Environment. A brand running an unfavourable CCI in a 19-percent VAT-anchored format compounds margin disadvantage on both sides of the channel split.

Section 4 — The pilot cohort: five DACH cases across the zone range

POS channel-split data for DACH operators are not publicly disclosed. The five readings below rely on convergent proxy indicators — corporate filings, unit-count deltas, investor-call commentary, structural logic. They represent informed zone plausibility, not audit results from POS-level data. The confidence column distinguishes high readings (multiple proxies converging) from medium ones (one or two proxies plus structural logic).

H1 — that QSR and delivery-focused concepts return CCI below 30 percent — is supported by Domino's and Starbucks. Domino's DACH marks the structural boundary of the index: the concept was designed from inception as a delivery-and-takeaway operation, with no strategic dine-in share. CCI = 0 is a consequence of business-model architecture, not management excellence. There is no dine-in baseline to be cannibalized. Delivery-native concepts are CCI-non-measurable by definition, and an Incremental Revenue Rate is the appropriate alternative metric for that segment The Concept-Market-Fit Score for Foreign-Chain Entry.

H2 — that casual-dining concepts with dine-in heritage return CCI above 50 percent at delivery launch — is supported by Nordsee, Vapiano, and Pizza Hut. Nordsee documents product-delivery mismatch at the High–Critical band. Vapiano introduces the pre-trend complication: a structurally falling dine-in baseline forces CCI_adj, and the destructive reading survives the adjustment. Pizza Hut sits at the destructive extreme, with a US footprint trajectory that confirms the same mechanic outside DACH.

Section 5 — Two failure modes that look like one and aren't

Nordsee is 128 years old, has cycled through six owners across three decades, and has contracted from roughly 397 outlets in 2013 to about 240 in 2025, with revenue per outlet of approximately EUR 1.07m on DEHOGA 2025 reference data. The mechanism is structural, not operational. Fresh fish does not survive a 30-to-45-minute delivery window without measurable quality loss. An average check between EUR 12 and 18 falls under serious margin pressure once a 25-to-35-percent aggregator commission is deducted. The regular customer who visits Nordsee for an in-store experience is not addressing the same need with a delivery order. The contrast with Domino's is informative. Both concepts operate in the broader QSR segment. Both have long DACH market presence. The difference sits exclusively in product-delivery fit: Domino's was optimized for delivery — packaging, bake time, ingredient stability, price positioning — while Nordsee was optimized for in-store consumption and never fundamentally revisited that optimization. Delivery implementation cannot bridge the structural gap Resilience Asymmetry in Chain Foodservice.

Pizza Hut documents the destructive extreme. The chain that effectively invented the dine-in pizza restaurant attempted, from the mid-2010s onward, a transformation toward delivery-first while retaining large-footprint dine-in locations. The US master franchisee NPC International, with 1,227 Pizza Hut units pre-bankruptcy, filed Chapter 11 in 2020 with approximately USD 1bn in debt. The global Red Roof footprint contracted from approximately 3,000 historical units to approximately 2,007 by end-2024, with 250 further closures announced for 2026. The proxy CCI for the German transformation period sits in the 100-to-130-percent band: delivery does not compensate for dine-in losses and additionally erodes equity. NPC International's filing is not a singular management failure; it is the result of a sector strategy that read delivery as a growth channel without measuring its incrementality. The US trajectory rhymes with the DE trajectory: stable or growing delivery KPIs, simultaneous dine-in closures, erosion of total profitability.

The intervention that would have rescued a structurally dine-in-anchored format is not platform optimization. It is repositioning — the product, the occasion, the format. Channel arithmetic measures the question. It does not answer it.

Section 6 — Occasion separation is the only structural defence against high CCI

Starbucks Germany delivers the counter-model. The chain introduced Mobile Order and delivery — via Uber Eats in major German cities from 2020 — as new channels without revising the third-place positioning. The proxy CCI lands between 10 and 25 percent. The explanation is occasion logic, not platform architecture. A coffee ordered to the office through an app does not displace the afternoon-café visit; the two occasions are recognisably distinct. Delivery opens a new consumption occasion rather than substituting one already covered. Howard Schultz addressed this mechanism in several Starbucks investor calls, framing Mobile Order explicitly as an incremental channel with the cannibalization hypothesis tested internally and judged non-significant.

This finding connects directly to the Occasion Coverage Index, OCI — one of four CMFS sub-dimensions developed in our companion methodology brief The Concept-Market-Fit Score for Foreign-Chain Entry. OCI measures how many distinct occasions a concept covers in dine-in. The backtest data suggests that CCI and OCI are inversely correlated: concepts with broad occasion coverage — breakfast, work-lunch, dinner, casual-drop-in — naturally produce lower CCI on delivery launch, because delivery adds another occasion rather than substituting one already covered. Concepts with narrow occasion identity — restaurant dinner as experience, fish-counter visit as seasonal ritual — are structurally vulnerable to high CCI: delivery competes for the same occasion with a worse experience. The CMFS brief locates OCI inside the concept-readiness composite; this brief locates the same variable on the channel side. They are the same variable read from two angles.

Occasion architecture before delivery launch is not a marketing exercise. It is a KPI-optimisation activity. An operator who introduces delivery without analysing occasion structure sets CCI structurally high — before the first order is placed. The result will not appear on the platform dashboard. It will appear in the POS channel comparison, quarter by quarter.

Section 7 — Three operating consequences for channel management

The five-case backtest converges on three operating consequences for channel-management decisions in chained foodservice.

First, CCI belongs in the standard metric set before an aggregator contract is signed. CCI_be must be carried alongside it as a mandatory annotation. Multi-unit operators should run a unit-level CCI profile that separates locations above 50 percent from those below 25 percent — operational interventions differ at each end, and an aggregated network reading hides the tension.

Second, a product-delivery fit audit belongs ahead of the platform rollout. The three pre-launch questions — does the core product hold a 30-to-45-minute delivery window without measurable quality loss; does the average check remain margin-positive after aggregator deduction; does delivery address a recognisably different occasion than dine-in — are a framework for any channel expansion. Chains that ask these questions after launch get the answers from the financial statements, at a point where structural repositioning is materially more expensive than pre-launch analysis would have been.

Third, pre-trend documentation of dine-in revenue — six months ahead of the planned delivery launch — is methodological hygiene. The Vapiano case documents the cost of skipping it. When dine-in is already trending negative before launch, CCI_raw overstates cannibalization and obscures structural baseline issues. CCI_adj separates the two — but only if the pre-launch baseline has been recorded.

Operators that read the platform dashboard alone read the channel metric without the channel arithmetic. The channel arithmetic decides the margin.

Section 8 — What the score does not measure

CCI measures channel incrementality. It does not measure conceptual market relevance — that sits inside CMFS. It does not measure entry-timing windows, aggregator-negotiation leverage, multi-aggregator stack differentiation, or external shocks that compress discretionary spending across both channels simultaneously.

The measurement gap in hospitality channel management — delivery volume without incrementality verification — is a function of missing tools, not disinterest. CCI and CCI_be address that gap. Whether the metric becomes industry-standard depends on whether the data it requires is routinely exported from POS systems — a technical capability that has existed in modern systems for years, but is rarely deployed for channel analytics. The five cases provide the empirical foundation. The methodological tool is in place.

Section 9 — Prescriptive close

We measure delivery growth against the dine-in baseline before we sign the aggregator contract. Operators that read the platform dashboard alone read the channel metric without the channel arithmetic — and the margin appears in the quarterly P&L only after the structural decision is signed.

Related research

• The Concept-Market-Fit Score: A Four-Dimension Composite Including OCI The Concept-Market-Fit Score for Foreign-Chain Entry
• Resilience Asymmetry in Chain Foodservice: Delivery-DNA as Structural Variable Resilience Asymmetry in Chain Foodservice
• The DACH Operating Environment: Four Differences That Reset Foreign-Chain Unit Economics The DACH Operating Environment
• HERI-40 Section 2 — Market Context 2026: Multi-Brand-Operator Reach and Aggregator Density HERI-40 Section 2: Market Context 2026

Sources

• Pizza Hut / Yum! Brands corporate disclosures 2019–2024 (Red Roof footprint reduction, NPC International bankruptcy 2020)
• NPC International — Chapter 11 bankruptcy filing 2020 (~USD 1bn debt; 1,227 Pizza Hut units pre-bankruptcy)
• Vapiano SE — annual reports 2017–2019 (like-for-like minus 4.2 percent at nine months 2019; industry growth plus 5.3 percent)
• Vapiano SE — insolvency filing April 2020
• Nordsee — public reporting and DEHOGA industry data 2025 (~240 outlets, ~EUR 1.07m revenue per outlet, contraction from ~397 outlets in 2013)
• Starbucks Corporation EMEA — Investor Relations 2020–2021 (mobile order incrementality positioning, Howard Schultz commentary)
• Domino's Pizza Enterprises Ltd. — annual reports DACH segment (delivery-native model, app-order share)
• DACH aggregator commission disclosures — Lieferando / Just Eat Takeaway / Wolt 2023–2025 (commission band 25 to 35 percent)