A cloud kitchen has none of the forgiveness of a dine-in restaurant. There is no server to smooth over a slow order, no ambience to compensate for lukewarm food, and the customer’s entire experience is the bag that arrives thirty minutes after they tapped a button. Run three or four virtual brands out of the same kitchen for two or three aggregators and the number of ways to get that bag wrong multiplies.
This makes delivery kitchens unusually dependent on standard operating procedures. The good news is that the environment is controlled: no front of house, a fixed set of stations, a known order flow. Here are the SOPs that matter, what good looks like for each, and what proof they should leave. For the writing method see how to write an SOP; for definitions, our guide to standard operating procedures.
1. Station setup and mise en place by brand
Covers: which stations serve which brands, par levels of prepped items per brand per daypart, labelling and dating of prepped containers, allergen separation between brands sharing a station.
Standard: every prepped container labelled with item, brand if brand-specific, prep time, use-by, and initials; pars hit before the daypart opens; allergen-sensitive items in dedicated, colour-coded containers.
Proof: photo of each station at setup against the standard photo; par checklist with counts.
Multi-brand kitchens live or die on labelling. A mislabelled container is a wrong order at best and an allergen incident at worst.
2. Receiving and storage
Covers: the standard receiving SOP (vehicle, quantities, probe temperatures, dates, accept or reject, storage within 15 minutes), plus brand-level stock separation where brands have distinct ingredient specs.
Standard: chilled 0–5°C, frozen solid, dates within policy. Storage locations fixed and labelled by brand or shared.
Proof: readings per category, photos of rejections and signed delivery notes. The full receiving SOP is written out in our SOP template article.
3. Order flow and ticket handling
Covers: how orders from each aggregator tablet or the integrated screen are accepted, prioritised, called and fired; how multi-brand orders in one basket are handled; how modifiers and allergen notes are communicated to the line; what happens when an item is unavailable.
Standard: accept within a defined time of receipt; fire in a defined priority order (usually by promised time, not by aggregator); allergen notes read back aloud; unavailable items switched off on every platform within minutes, not left to generate cancellations.
Proof: aggregator dashboards give acceptance and prep-time data; the SOP’s own proof is the menu-availability check at each daypart and the cancellation log in the logbook.
4. Cooking, holding and food safety
Covers: core temperatures, cooling and hot-holding, fryer oil quality, hold times for cooked components awaiting assembly, cold-holding logs for all units.
Standard: core 75°C or your regime’s equivalent; hot-holding at or above 60°C; cooked components discarded after a defined hold; fryer oil tested and changed on polar-compound reading or schedule; cold units logged four times a day.
Proof: temperature readings with timestamps, hold labels, oil test log. Delivery kitchens in the Gulf are inspected like any other food establishment; see our guides to food safety SOPs in the UAE and in Saudi Arabia.
5. Assembly, packaging and labelling
Covers: the pack standard per item per brand (container, portion, garnish, sauce placement), sealing, tamper-evident labels, order label with brand, items, allergens and time, bag standard, hot and cold separation, cutlery and napkins per brand policy.
Standard: a photo standard for every item as packed; every bag sealed and labelled before it reaches the handover shelf; hot and cold in separate bags or compartments.
Proof: periodic pack audits (a shift lead photographs three random orders per daypart against the standard); label stock check at opening.
Packaging is the moment the brand becomes real to the customer. It deserves the same photo-standard discipline as a restaurant’s plating.
6. Handover and courier management
Covers: the handover shelf layout by aggregator and brand, how couriers are greeted and verified, order ID check before release, maximum shelf time before escalation, remakes when shelf time is exceeded, courier conduct in the kitchen area.
Standard: order ID read back before release; nothing on the shelf beyond a defined time (often 10 minutes for hot food) without a decision to remake or escalate; couriers do not enter food-prep areas.
Proof: shelf-time exceptions logged; handover area photo at each daypart.
7. Cleaning and sanitation
Covers: the standard cleaning schedule (surfaces, equipment, floors, drains, hoods, fryers, ice machine) with concentrations and frequencies, plus between-brand sanitation on shared stations where allergens differ.
Standard: defined per surface; sanitizer tested; shared stations sanitized at brand changeover.
Proof: completion per item, test strip photo, end-state photos of fryers, hoods and drains.
8. Menu availability and 86 list
Covers: how out-of-stock items are identified, who switches them off on each aggregator, how they are switched back on, communication to the line.
Standard: 86’d items off every platform within 5 minutes; checked at every daypart change.
Proof: a daypart menu check task with per-platform confirmation; cancellation reasons logged and reviewed weekly.
Cancellations for unavailable items are the most avoidable loss in delivery and the most damaging to aggregator rankings.
9. Complaints, refunds and remakes
Covers: how aggregator complaints and ratings are reviewed, who authorises refunds or remakes, how each is recorded with cause (missing item, cold, wrong item, late), weekly root-cause review.
Standard: every complaint categorised the same day; remake and refund authority defined; causes reviewed weekly by brand.
Proof: logbook entries with category, brand, platform and remedy.
10. End-of-shift reset and closing
Covers: stations broken down and cleaned, prepped items labelled and stored or discarded, waste counted by brand, equipment off, temperature logs completed, bins out, security.
Standard: nothing unlabelled in cold storage overnight; waste recorded; photo end states.
Proof: cold-holding readings, waste counts, photos of stations and bin area, back door.
What is different about cloud kitchens
- Labelling is the master SOP. Containers, bags and orders. Most errors trace back to it.
- Brand changeover on shared stations is a procedure of its own, especially where allergens differ.
- Aggregator metrics are your inspection. Acceptance time, prep time, cancellations and ratings per platform per brand belong on the same weekly review as food-safety completion.
- Speed pressure is constant. Keep each SOP short and require proof on few, high-value steps: labels, temperatures, pack audits, shelf-time exceptions.
Running it across sites
Groups running several kitchens, or several brands across several cities, need the SOPs centralised and the execution local. In Opscale, the pack standards, labelling rules and SOP videos live in the operations library; daypart setup, cold-holding, menu-availability checks and closing run as role-based tasks in tasks and checklists; shift leads run pack and hygiene audits in inspections; complaints and cancellations go to the logbook; and the executive dashboard shows completion and exceptions by kitchen and brand. The AI checklist creator will draft any of these from a description of your brands and stations.
More on how delivery operators in the region use it on our cloud kitchens page. The bag is the brand; the SOP is how the bag comes out right every time.