Recurring signal
The pattern
A virtual brand run out of an existing kitchen looks like free money: same lease, same equipment, same staff, a second storefront on the delivery apps. The recurring operator lesson was that the second brand is only free money if it captures demand the first brand was not already going to get.
Delivery-only concepts that duplicated a restaurant's own core menu under a different name multiplied packaging cost and kitchen-ticket volume without multiplying real revenue. The platforms themselves have started enforcing this distinction directly, which changed the operating math for anyone running more than one brand from one kitchen.
Pull-quotes are editorial reconstructions of recurring operator accounts. Identifying details are changed.
Most kitchens never measured whether their brands were competing with each other.
Industry analysis of virtual-brand portfolios in 2026 described brand-on-brand cannibalization as common and almost never priced into the decision to launch a new concept. When two brands from the same kitchen appear in the same delivery radius, some share of orders that would have gone to the higher-margin brand diverts to the lower-margin one — leaving the lower-margin brand's contribution look artificially strong while quietly eroding the portfolio's actual total.
The diagnostic operators used, once they started measuring it, was straightforward: for each pair of brands, ask what percentage of one brand's orders would migrate to the other if it were retired, versus leaving the portfolio entirely. A high migration rate toward a higher-margin sibling brand was a signal to retire the weaker one, not add a third.
"We had two brands quietly stealing from each other for over a year before we ever looked at the overlap. Once we did, one of them was clearly just a worse-margin version of the other."
The platforms started enforcing real differentiation.
Uber Eats has publicly described purging thousands of near-duplicate virtual brands and now requires more than half of a virtual restaurant's menu to differ from its parent brand or from other brands sharing the same kitchen. DoorDash applies similar scrutiny, requiring virtual brands to be meaningfully distinct, with minimum item counts and rating thresholds. Both platforms' stated motivation was the same: too many near-identical listings had begun eroding customer trust in search results.
Operators who had built brands as thin renames of an existing menu found themselves needing genuine reformulation to stay listed — a forced correction of exactly the cannibalization problem the platforms were, separately, also trying to solve for their own search quality.
"The fifty-percent menu-difference rule felt like a compliance headache until we realized it was solving the exact cannibalization problem we hadn't gotten around to fixing ourselves."
The lowest-risk brands captured a different daypart, not a different name.
The pattern that recurred most often among operators who reported a genuinely incremental virtual brand was narrow: a late-night or off-peak concept using the same equipment during hours the kitchen would otherwise sit idle, built around a different craving than the core menu rather than a repackaged version of it. Rent, core staff, and core equipment were already paid for; the marginal cost was closer to true incremental cost than a standalone concept ever achieves.
The riskiest pattern was the opposite: a standalone ghost-kitchen brand with no walk-in traffic at all, depending entirely on paid visibility inside a delivery app. Analysis of that model's economics in 2026 described commissions of 25% to 30%-plus, combined with the marketing spend required to be visible at all, permanently consuming a large share of revenue — a structural problem no amount of operating discipline fully offsets.
"The brand that worked was the one using our fryer at 11 p.m. when it was already hot and already paid for. The brand that didn't work was the one that existed only inside an app and had to buy its way into being seen."
Uncoordinated promotions funded the cannibalization directly.
Operators running multiple brands reported a specific, avoidable failure: two sibling brands running competing discounts on the same night, splitting demand that would otherwise have gone to one brand at full margin. Without a shared promotion calendar across brands, a restaurant group can end up discounting against itself as effectively as any competitor could.
The fix that recurred was administrative rather than technical: a single shared calendar, reviewed before any brand launched a promotion, checked against every other brand's active offers in the same delivery radius.
"Our wing brand and our taco brand ran competing deals the same Sunday night for months before anyone noticed we were bidding against ourselves."
Standalone ghost-kitchen margin math
The math
Industry analysis of standalone, no-walk-in ghost kitchens describes a well-run kitchen netting roughly 5% to 7% on a delivery order at a 25% marketplace commission — falling to 0% to 2% at 30% commission, before promotion and marketing spend required for app visibility. Once processing, promotions, and refunds are counted, the effective take from a standalone kitchen's revenue often runs 30% to 40%.
An incremental virtual brand sharing an existing kitchen's rent, staff, and equipment does not carry that fixed-cost burden — its economics are structurally different, and considerably more forgiving, than a standalone ghost kitchen built for delivery alone.
Source notes
Public documents behind the field notes
- The Wall Street Journal reporting on Uber Eats' virtual-brand policy tightening, as summarized in trade coverage of the change.
- DoorDash and Uber Eats public merchant guidance on virtual-brand menu-differentiation and rating requirements.
- Industry analysis of standalone ghost-kitchen unit economics and multi-brand portfolio cannibalization measurement, 2026.
Source review: July 17, 2026. Platform policy on virtual brands has changed multiple times since 2022 and may change again.
Verified operators only
Add your field notes.
Membership is free and always will be. Applications are reviewed weekly. Vendors are declined.
Request to join →