Choosing Your Restaurant Architect, Kitchen Consultant and Contractor

A restaurant build is decided long before the hoarding comes down. It is decided in the appointments — the architect, the kitchen consultant, the MEP engineer and the contractor you choose, and the contracts you choose them on. Each of these parties is essential. Each controls a different piece of the outcome. And none of them — this is the part owners learn expensively — is appointed to hold the whole.

What follows is the owner’s selection system: what each party actually controls, the tests that separate a strong appointment from a plausible one, the red flags that predict trouble seat by seat, and the coordination gap that sits between all of them. It is not legal or licensing advice — statutory and regulatory questions belong with licensed professionals and the authorities — and it pairs with the lease and fit-out traps, which covers the deal you sign before this team ever mobilises.

What each party actually controls

Most opening problems are really scope problems: work everyone assumed someone else owned. So start with an honest job description for each seat at the table.

The architect owns the spatial and statutory design. The plan and section of the venue, the guest experience, the materials and finishes — and the drawings the authorities approve. In Dubai and across the GCC that statutory layer is real, layered work: food-safety authority, civil defence, the landlord’s technical team, sometimes a freezone or master developer on top. The architect’s natural centre of gravity is the dining room, because that is where architecture is visible. That is exactly why the kitchen must not be left to arrive as whatever space remains once the dining room is drawn.

The kitchen consultant owns the production side: the flow from receiving to pass, the equipment schedule, and the food-safety zoning that keeps raw and ready-to-eat paths from crossing — the discipline that HACCP kitchen design exists to enforce. A commercial kitchen layout is a food-safety diagram before it is a room, and the equipment schedule that comes out of it is not a shopping list: done properly, it is a priceable, testable document that carries the services demand of every item — power, water, drainage, gas, extract — which is what the engineers design from.

The MEP engineer owns the services: electrical load, ventilation and extract, gas, water supply and drainage — MEP, the invisible half of the build. It is the least forgiving scope on the project, because an undersized service is discovered under full load, after opening. The engineer can only be as good as the demand information they are given, which is why this seat depends so heavily on the kitchen consultant’s schedule — a dependency covered in depth in the kitchen MEP requirements piece.

The contractor owns execution: pricing the drawings, mobilising and sequencing the trades, managing subcontractors, passing inspections and closing snags. The contractor builds what the documents describe — no better than the documents, and no more coordinated than the design that reached tender.

And who owns the owner’s interest when these four disagree — when the architect’s wall sits where the kitchen consultant’s chiller must go, or the contractor’s variation prices a gap between two drawings? Read the four appointments again. By default: nobody.

The boundary that keeps the project honest

Before selection even begins, fix one boundary, because it defines what you are buying from every seat: statutory design, licensed engineering and authority approvals belong to licensed professionals. The architect and engineers appointed to carry that work, and the authorities who approve it, are the only parties who can lawfully perform it. No adviser outside those appointments should be doing it, and anyone who blurs that line has told you something important about how they treat rules in general.

GGB’s own role in a build sits deliberately on the other side of that boundary: owner-side coordination, specification and programme control. We define what the licensed parties must design to, hold their work to one coordinated programme, and keep the owner’s interest in the room — we do not perform the licensed work itself. Any owner assembling a team should demand that same clarity from every consultant they appoint.

Four tests for every seat

The selection criteria are the same four tests, applied party by party. What changes is what a pass looks like.

Portfolio class. Not “have they done restaurants?” but “have they done your class of restaurant?” A fine-dining specialist is the wrong architect for a delivery-first rollout, and the reverse is just as true. Look for projects of your service model, your intensity of production, your kind of site — and ask what role they actually played on each, because portfolios inflate.

F&B specificity. Restaurants are a specialist domain wearing a general costume. An architect whose portfolio is villas and offices will draw a beautiful room that fights service. A contractor who has never built a commercial kitchen will underprice the services packages and recover the difference in variations. An engineer without F&B work will size from floor-area habit instead of the cooking line. Generic excellence does not transfer; ask specifically for the food-and-beverage evidence.

Documentation depth. Ask to see a real, past-project document set — redacted is fine: a drawing package, an equipment schedule, a programme, a tender comparison. The depth of the documents predicts the depth of the work, because thin documentation means decisions deferred to site, and decisions made on site are priced on site, by the party with the least incentive to price them kindly.

Variation behaviour. The most revealing test, and the least asked. Every build changes; the seat’s behaviour under change is the real product you are buying. Ask each candidate how variations were handled on their last project, then ask their references the same question. You are listening for process — notified, priced against tendered rates, approved before executed — versus improvisation invoiced afterwards.

Red flags, seat by seat

The architect. A portfolio of dining rooms photographed empty, with no kitchens shown. Resistance to early kitchen-consultant involvement — “we’ll leave space for the kitchen” is the sentence that precedes crossing flows and under-sized extract. A vague statutory scope in the appointment: who, precisely, carries which approvals. And any sign the kitchen is being treated as leftover space rather than the machine the venue exists to house.

The kitchen consultant. Leading with equipment brands before understanding your menu — sometimes the consultant is a dealer in costume, and an equipment-sales incentive quietly shapes the schedule. A schedule with no services data per item, which makes it useless to the engineers. No zoning logic — if raw, cooked, wash and waste paths are not drawn as separated flows, the food-safety thinking has not been done. Ask how they are remunerated, and prefer the answer that is independent of what you buy.

The MEP engineer. Sizing from rules of thumb before the equipment schedule exists. No F&B projects in the portfolio. Treating extract as ducting to be routed rather than a system sized from what the menu burns. Silence on makeup air, grease management or suppression interfaces — the places where kitchen services interlock and generic buildings do not.

The contractor. A lump-sum price against concept drawings — pricing what is not yet drawn is either padding or a variation machine, and both are your cost. The lowest bid with the thinnest measurement behind it. A long exclusions list buried in the tender return. No named MEP subcontractor for the packages that decide the kitchen. References only from fast, simple, dry builds. Award behaviour under a proper tender tells you most of this before it can hurt you.

The coordination gap nobody owns

Now the failure pattern that wrecks openings — and it is rarely any single party’s failure. The architect’s partitions are agreed while the equipment schedule is still moving. The engineer sizes services from a superseded schedule revision. The contractor tenders on a drawing set in which those three positions have never been reconciled. Every party did their contracted job. The seams did the damage — and on site the seams surface as cores drilled through finished walls, services rerouted around each other, re-approvals, variations priced under time pressure, and an opening date moved by paperwork rather than construction.

The disconnection is structural, not moral. Each appointment ends at its own scope boundary; reading the other parties’ documents deeply is unpriced work; and the only person with a whole-project interest is the owner — usually the least technical person at the table, and the one writing every cheque.

Owner-side restaurant project management exists to close exactly this gap. One seat, on the owner’s side of the table, that holds a single coordinated design freeze before tender, chairs the coordination between architect, kitchen consultant and engineers so no discipline designs against another, runs the tender so bids land comparable, governs variations against tendered rates, and reports in the operator’s language: cost, date, risk. That is the shape of GGB’s fit-out project management — and if you want to see what it looks like as paper rather than promise, the waterfront-café fit-out dossier on /work holds the contemporaneous records of one such delivery: a work programme spanning plumbing, electrical, gas, fire suppression, kitchen exhaust and grease management, an itemised equipment BOQ, site-status records, and a landlord-stamped final design approval. Coordination, when it is real, is boring documents — that is the point.

The BOQ: what makes bids comparable

One instrument does more than any other to keep the contractor selection honest: the bill of quantities. A restaurant BOQ breaks the fit-out into measured line items — quantities, specifications and rates for every trade — so that when three contractors price the project, they price the same scope, and their numbers can be compared line by line instead of as three incomparable lump sums. Without one, the lowest quote is usually the one that measured least, and the missing measurement returns later, priced as variations at the moment you can least refuse them.

The BOQ keeps working after award, too: it is the reference that keeps variations honest, because a change is priced against the tendered rates rather than invented under schedule pressure. The discipline around it is simple to state and demanding to hold — design freeze before tender, every bid returned against the same bill, award on comparable bids only.

The GGB read

We are unsentimental about this: the team is the project. Choose an architect for spatial and statutory command, a kitchen consultant for flow and a schedule the engineers can actually design from, an MEP engineer who starts from the cooking line, and a contractor whose measurement is as thick as their promises — then accept that none of those appointments, however strong, covers the seams between them. Someone must sit owner-side and hold the whole: freeze, tender, variations, programme, truth. In a full build mandate that seat is built into the system; appointed standalone, it is still the difference between the venue you priced and the venue you get. The licensed professionals design and build the restaurant. The owner’s side of the table is where its interests are defended — staff that side as deliberately as you staff the kitchen.