Who at head office should own delivery app uptime?
In most restaurant chains the honest answer is nobody, and the reason is structural rather than negligent. Delivery availability is assembled from three departments that each own a fragment: operations holds the branch and the tablet, technology holds the integration, marketing holds the listing and the promotional eligibility that depends on it. Every one of them can point at a working part. The outcome, which is whether a customer could order, belongs to none of them. The fix is to name one owner, give them the access the platforms actually require, and hold them to a narrow set of numbers.
Why does delivery uptime end up with no owner?
Because the failure crosses the seams. A branch that is dark because a schedule expired is an operations problem. The same branch dark because a catalogue push failed is a technology problem. The same branch dark because a marketing tool was withdrawn on a monthly assessment is a marketing problem. From the customer’s side these are one event, and inside the business they are three tickets in three queues.
The seam produces a predictable behaviour. Each function verifies its own layer, finds it healthy, and closes the item. Nobody checks the storefront, because checking the storefront is not in any of the three job descriptions.
Which permissions do the platforms already tie to specific people?
More than chains expect, and this is the practical constraint on who the owner can be. Deliveroo restricts its bulk controls by role, stating that “Hub users with the Admin, Admin (limited), Manager, or Manager (limited) role can access the Site Status feature”, while noting that “a store can still be manually opened or closed using the Deliveroo tablet”. Source: help.deliveroo.com. Careem attaches a condition to the marketing section that catches people out, “Access to marketing features is restricted unless you sign in via phone number”. Source: Careem Partner Portal.
Integration authority is narrower still. Foodics tells operators connecting The Chefz to “Log in to your Foodics console at console.foodics.com/login using the owner email”, warning that “If you logged in with a non-owner user, you will not be able to activate the integration.” Source: help.foodics.com. And on Talabat and HungerStation, the switch that decides whether a branch opens without a human is enabled by an account manager rather than by anyone inside the business at all.
An owner who does not hold these permissions cannot act, and an owner who cannot act is a reporting line rather than a role.
What does the owner have to hold before anything else?
An inventory, and most chains discover they do not have one. The unit is a listing, meaning one location on one platform, so a group with forty locations across four platforms owns a hundred and sixty of them rather than forty. Each carries its own state, its own schedule and its own identifier on the platform side.
The inventory needs four columns to be useful: the platform, the platform’s own identifier for that location, the internal branch code, and who holds the credentials. Chains routinely find listings nobody remembers creating, locations still catalogued after the site closed, and branches trading on a platform head office believed they had left. Until that list exists there is nothing to be accountable for, because the denominator is unknown.
Where should the role sit?
In operations, with the measurement deliberately placed outside it.
Operations is the right home because almost every remedy is an operational action taken quickly: clearing a pause, answering a check in, correcting hours, calling the platform. Technology owns the pipe and marketing owns eligibility, and both should be named as dependencies rather than as owners, since neither is set up to respond within the life of an incident.
The measurement is different. If the same function both restores the storefront and reports how long it was down, the number becomes a performance review and starts to shrink. Whatever the reporting line, the record has to come from somewhere that has no stake in the total.
What does the owner actually do day to day?
Triage, not vigilance, and the measured distribution is why. Across the UAE panel, 36.5 percent of listings had no downtime at all in the month, while 21 percent of the affected listings were dark on five or more separate days, which is 13.4 percent of everything measured. Source: UAE delivery downtime report.
That shape defines the job. Most of the estate needs nothing. A small group of listings produces most of the loss and produces it repeatedly, which means the work is investigation of a shortlist rather than surveillance of everything. An owner spending the week watching a wall of green is doing the wrong job, and the right one is asking why the same six listings appear every month.
What is the owner accountable for, and what are they not?
Three things, and they should be written down as such.
Detection time, meaning the gap between a listing becoming unorderable and somebody in the business knowing. Restoration time on the states the business can lift, which is most of them. And recurrence, meaning whether a branch that failed last month failed again.
What the role is not accountable for matters equally, because an unfair target is abandoned quietly. Platform side closures that only the platform can lift are outside it. So is courier supply, which Talabat and HungerStation both publish as a distinct closure reason, TOO_BUSY_NO_DRIVERS, separate from TOO_BUSY_KITCHEN. Source: developer.talabat.com. Splitting the total into hours the business could have prevented and hours it could not is the single change that makes this role manageable.
Two questions sit next to each other here and answering them together produces an unfair job description. This page is about who inside your business owns the outcome. Who created a particular closure, and who carries the trading time it cost, is a separate question with a separate answer, set out on who is responsible when a delivery app takes your restaurant offline. An owner appointed without that distinction inherits every hour, including the ones nobody in the building could have prevented, and the role does not survive its first bad month.
How does a franchise change the answer?
It moves the switches out of the building. A franchisee holds portal access for its own branches, sets its own pauses and answers its own tablet, and none of that is visible to a brand owner who lacks credentials. The brand still carries the consequence, since the customer sees the brand rather than the operating company.
The workable arrangement separates observation from control. The brand owner measures every listing from the customer side, which needs no credentials at all, and the franchisee retains the actions. The detail of that arrangement is set out in how franchise operators monitor all branches.
What number belongs on the owner’s page?
Lost trading hours per listing per month, split into liftable and not liftable, with a recurrence column beside it and the market rate for comparison. Not uptime as a percentage, because a percentage of a large denominator hides a lost evening.
Hours are the right unit for a second reason. They convert into money with one multiplication, they can be argued with a platform, and they are the same unit the branch manager thinks in. The conversion itself is set out in how to calculate what delivery downtime costs.
The owner needs that number to arrive without being requested, from a source none of the three departments controls, which is the position Kitchain (kitchain.co) occupies by reading every listing from the customer side and comparing it against the hours that listing itself published.