The Complete Restaurant Setup Process in Dubai — Every Workstream, in Order

Opening a restaurant in Dubai is usually described as a licence checklist. From the owner’s side of the table it is something else entirely: a development project — capital committed early against a P&L that does not yet exist, across seven workstreams that each produce a deliverable, close a decision gate, and hand their answers downstream. We have written the step-by-step licence-path guide separately; this piece is the other view — the complete restaurant setup as a development programme, run the way a principal runs it. One caveat before the map: this is not legal, licensing or tax advice. Requirements change, and current fee schedules sit with the relevant authority — the Department of Economy and Tourism, Dubai Municipality, Civil Defence — so confirm your own case with them directly.

A setup is gates, not tasks

The workstreams below are not a to-do list; they are a chain of decision gates. A gate closes when a specific question is answered with evidence — can this model carry its costs? what exactly are we building? at what contracted price? — and once a gate closes properly, nothing downstream should reopen it. Nearly every distressed opening we are later asked to examine broke this rule somewhere: work started before the gate feeding it had closed, and the cost of the reopened decision compounded through everything built on top of it. Hold that lens and the rest of the process reads itself.

Workstream 1 — feasibility & the investment case

What it produces: a demand model for the actual catchment; the break-even covers per day the site must deliver; a rent test against realistic revenue (healthy operations typically hold rent inside a 6–12% share of sales, varying by format); a capital budget built by category with an owner for each line; and a cash plan that funds the ramp — assembled into an investment case a bank, a partner or a sceptical spouse could interrogate.

The gate it closes: invest, re-scope, or walk away — decided on paper, while walking away is still cheap.

Everything downstream inherits this workstream’s answers, which is why it opens the build-a-restaurant path and why it is the one piece of work we anchor publicly: a formal feasibility and investment case (From AED 45,000 — indicative, scoped per project) exists to kill weak models before they are built in tiles and steel. A setup that starts at workstream two has not skipped a step; it has skipped the verdict.

Bought separately, it goes wrong like this: it usually is not bought at all — it is back-filled to justify a lease already signed, which turns the verdict into decoration.

Workstream 2 — concept, brand & menu

What it produces: the concept defined tightly enough to build from — positioning, format, service model, brand identity — and a menu treated as a financial document: every dish recipe-costed to a target (food cost typically engineered to 32% of price or below, as a published teaching band), portion-specified, and executable by a kitchen the budget can actually afford.

The gate it closes: what exactly are we building? — frozen. This freeze is what workstreams four and five will price; every concept edit made after it reopens the design, the equipment schedule and the tender at once.

The menu belongs here, not at pre-opening, because the menu sizes the kitchen and the kitchen sizes the budget. A concept, brand and menu package built in sequence is cheap insurance against the most expensive edit in F&B: changing your mind after the contractor starts.

Bought separately: a branding studio delivers a beautiful identity, a consultant chef writes a menu the format cannot afford, and the kitchen is quoted against neither.

Workstream 3 — site, lease & the licensing route

What it produces: a site shortlist tested against the feasibility demand model rather than a viewing-day feeling; a premises survey that documents handover condition — shell-and-core versus previously fitted is a different project, not a different price; a negotiated lease covering the fit-out period, any rent ramp, and absolute clarity on landlord scope versus tenant scope; and the licensing-route decision — mainland through the Department of Economy and Tourism, or a free-zone structure — chosen to follow the concept and the customer, never the other way round.

The gate it closes: where, and under which legal form — with the rent ratio and the approval pathway now fixed for the term of the lease.

This is the gate you cannot cheaply reopen, which is why site and lease due diligence is a named discipline rather than a viewing, and why the licensing and compliance route is scoped before signature: the premises must be capable of carrying the approvals the concept needs, and discovering otherwise after signing is how fitted-out units sit paying rent while their file waits.

Bought separately: the broker optimises for the signature, the licensing agent starts after it, and nobody ever checked whether this shell can carry this concept’s approvals.

Workstream 4 — the HACCP-ready kitchen & design package

What it produces: a kitchen designed from the menu’s production reality — stations, flow from receiving to pass, storage and extraction sized to what will actually be cooked — with Municipality food-safety requirements and Civil Defence expectations folded inside the drawings, and the MEP loads verified against what the shell can supply. The output is an approval-ready design package: buildable, inspectable, tenderable.

The gate it closes: a design the authorities can approve and a contractor can price — before anyone builds anything.

HACCP-ready kitchen design earns its name at this gate: compliance drawn in from the start costs a fraction of compliance retrofitted after an inspector’s comments, because retrofitting is paying for the same wall twice. The wider design and fit-out discipline treats the design package as the foundation of the construction contract — which is exactly what workstream five needs it to be.

Bought separately: an interior designer draws a dining room, a kitchen vendor drops a standard block plan into the back of it, and the first authority review sends both back to the start.

Workstream 5 — BOQ, tender & fit-out oversight

What it produces: a measured bill of quantities — every work and material itemised, which is the only thing that makes contractor quotes comparable at all; a tender run against it; a contracted price with staged payments tied to inspection-ready milestones; and oversight through the build to handover — progress certified against the BOQ, variations priced before they are built rather than after, and a snag list closed against retention rather than goodwill.

The gate it closes: the contracted cost, and a compliant handover — the difference between a budget and a hope.

This workstream is where restaurant project management stops being a phrase and becomes a daily job: the owner’s interests represented in every site meeting, between designer, contractor, kitchen vendor and authorities, none of whom is paid to protect the investment case. The anonymised development dossiers on our work page show what that oversight discipline looks like on real files.

Bought separately: without a BOQ, the lowest quote is simply the vaguest one — and every ambiguity converts into a variation order once the hoarding is up.

Workstream 6 — recruitment, training & SOPs

What it produces: the written operating system — kitchen, service, cash and stock procedures that exist on paper rather than in someone’s head — and a team hired against the organisation the feasibility priced, sequenced backwards from opening so that sponsorship, visas, onboarding and training all complete on the finished premises, not in a meeting room while the site is still dusty.

The gate it closes: a team that can run the written system, proven before the first paying guest.

In the GCC this sequencing is unforgiving: hiring runs through sponsorship and visa processing, so the recruitment clock has to start while the build is still under way. Recruitment and training and the SOP and manuals discipline are two deliverables with one deadline — because a trained team executing a written system is really a single deliverable pretending to be two.

Bought separately: an agency fills seats against a date the contractor has already missed, the SOPs arrive as an unlocalised template, and the team’s first real service becomes the public’s first impression.

Workstream 7 — pre-opening & launch control

What it produces: the countdown run as a checklist — supplier contracts and opening stock, dry runs and invited services that test the kitchen at load, the launch marketing plan, and the control set armed for day one: recipe costs live, cash procedures rehearsed, daily reads defined and owned.

The gate it closes: open on controls, not on hope — opening night as an execution, not an experiment.

A structured pre-opening and launch programme is what stands between a fitted-out restaurant and an operating one. Its quiet, decisive output is the first honest month: covers, food cost and labour read against the feasibility model — labour typically holding around 25–30% of sales and prime cost inside a typical 62% ceiling — with the gaps ranked and worked rather than explained away.

Bought separately: it usually is not bought — it is improvised in the last days before opening, which is why so many launches spend their most fragile period discovering their own basics on paying guests.

Why buying the workstreams separately is how openings go wrong

Read back through the seven “bought separately” failure lines and notice what they have in common: not one of them is a bad supplier. They are broken handoffs — work priced before its inputs existed, gates skipped because no single party owned them. The market sells the setup in pieces because pieces are easy to invoice: a feasibility PDF, a brand deck, a licence application, a drawing set, a build, a hiring drive, an opening party. What no individual piece contains is the thing the owner actually bought: one investment case, protected end to end.

That is the honest meaning of turnkey restaurant setup — not a package price and not a brochure, but a single party accountable for the sequence: the same discipline that ran the feasibility sitting in the tender meeting; the same owner of the menu’s food-cost target checking the kitchen drawings; the same hand on every gate. Whether that party is a development principal or a formidably organised owner, someone must hold the whole braid — because the suppliers, individually excellent, are structurally unable to. After 28+ years, 45+ concepts developed and launched and 300+ project engagements (how these figures are counted), our view on who should hold it is unsurprising; but the argument stands on its own arithmetic, because the margin between a controlled setup and a fragmented one is usually several times the fee that controlled it.

The GGB read

We run setups the way the order above implies: gates first, work second, and no workstream started before the gate that feeds it has closed. Feasibility is the verdict, not the paperwork. The concept freeze is what everything downstream prices. The lease and the licensing route are one decision, not two. Compliance is drawn, never retrofitted. The BOQ is what makes every dirham of the build traceable. The team finishes its training on the finished floor, and the launch is executed against controls that already exist. Run in sequence, the seven workstreams protect each other; bought in pieces, they quietly bill each other — and the owner pays the difference. If you are at the beginning, begin where the sequence begins: with the verdict.