I Signed 23 Developer Contracts in Dubai on the Wrong Pricing Model — the 7-Step Way I Now Choose Between Fixed Price and Time and Materials

I have run delivery for Dubai companies since 2019, and in that time I have signed or countersigned somewhere north of 90 developer and vendor contracts. Twenty-three of them were on the wrong pricing model. I know the number because I went back through them last year, after a fixed-price mobile build that finished 7 months late and 61% over its original figure, and asked a simple question of each one: knowing what we knew at signature, was fixed price or time and materials the right call? Sixteen were fixed prices on scopes that were not fixed. Seven were open-ended time and materials on work we could have priced to the dirham. The pattern was not bad vendors. It was that we chose the model by who was in the room, finance or engineering, instead of by the shape of the work. This is the seven-step method I use now, with the UAE-specific clauses that the generic guides leave out.

The Two Models, and the Question Neither Side Asks

A fixed-price contract commits the vendor to deliver a defined scope for a defined sum. You buy certainty of cost; the vendor takes the risk that the work is bigger than estimated, and charges a premium for that risk. A time-and-materials contract (T&M) commits the vendor to supply people at an agreed rate; you pay for the hours worked. You buy flexibility and transparency; you take the risk that the work is bigger than you thought.

Finance directors in Dubai like fixed price because it fits a budget line and a board pack. Engineering leads like T&M because they know the scope will move. Both are right about their own concerns, and both skip the only question that decides the matter: how certain is the scope, really? A fixed price on an uncertain scope does not remove risk. It converts it into change requests, delays and a strained relationship, which is more expensive than the premium would have been. A T&M contract on a certain scope pays for flexibility you will not use. The seven steps below exist to answer that question with a number instead of an opinion.

Step 1 — Score the Scope for Certainty, Out of 20

Before any commercial conversation, score four factors from 1 (no) to 5 (fully) and add them up.

1. Requirements. Are they written down precisely enough that a tester could pass or fail each one without asking anybody? A 40-slide deck scores 1. A backlog of testable user stories with acceptance criteria scores 5.
2. Integrations. Are the systems the software must talk to known, documented and accessible today? If the payment gateway, the ERP or the government API still needs a sandbox request, score 2 at most.
3. Acceptance criteria. Has the person who will sign off written down what “done” means, and has the vendor agreed it? If acceptance is “when the CEO is happy”, score 1.
4. Decision-maker availability. Can your side answer a vendor question within two working days, every week, for the length of the project? Be honest. In Dubai, with travel, Ramadan hours and summer leave, most teams score 3.

14 or above: fixed price can work. 10 to 13: hybrid, per package. Below 10: T&M with a cap, and spend the first four weeks getting the score up. Of my 16 wrong fixed prices, 13 would have scored under 10 on the day we signed. The scoring takes twenty minutes and would have saved us, by my estimate, about AED 2.1 million across those contracts.

Step 2 — Price the Risk Premium Honestly

Ask every vendor for both numbers: a T&M estimate (rate card, hours by role, total) and a fixed price for the same scope. Do not ask for one or the other; ask for both, on the same page. The gap between them is the price of the risk you are transferring, and it tells you more about the vendor than either number alone.

In the engagements we have priced or reviewed in the UAE this year, a credible fixed price sits 25% to 40% above the same vendor’s T&M estimate. That is a reasonable premium for absorbing scope risk on a three-to-six-month build. Two other cases should worry you. A fixed price at or below the T&M estimate means the vendor either has not understood the scope or plans to make the margin back through change requests; in my 16 wrong fixed prices, 9 had a premium under 10%, and every one of those turned into a change-request fight. A premium above 50% means the vendor scored your scope under 10 in step 1 and is telling you so politely; listen to them and go hybrid.

Step 3 — Choose the Model Per Work Package, Not Per Project

The single change that removed most of my errors was to stop signing one model for a whole project. A typical Dubai build has packages with very different certainty scores. The customer-facing web front end with approved designs scores 17: fix it. The integration with a bank’s payment API that has not yet issued sandbox credentials scores 7: T&M with a cap. The admin dashboard whose requirements exist only as a conversation with the operations manager scores 9: four weeks of T&M discovery, then fix the rest.

Structure the contract as a master services agreement with one statement of work per package, each with its own model, price or cap, acceptance criteria and review point. The MSA carries the UAE clauses from step 5 once; each SOW carries only what changes. This is what lets you switch model at a package boundary later (step 7) without renegotiating the whole relationship, and it is the structure the better Dubai vendors will propose themselves if you let them.

Step 4 — Write the Change-Request Clause Before the Commercial Terms

Every fixed price lives or dies on one paragraph: the change-request clause. I now write it first, before the price, because if we cannot agree this paragraph the price is fiction. It needs four things, each answered with a specific rule rather than a good intention.

• What counts as a change. Anything not in the SOW’s acceptance criteria is a change; anything in them is not. This sounds obvious and is the sentence most contracts omit. It is also why step 1 insists on testable criteria: they are the boundary the clause refers to.
• Who approves it. One named person on each side, with a named deputy. Not “the client”.
• How it is priced. At the T&M rate card from step 2, which you already have on the page. Fixed-price contracts should carry the rate card as a schedule precisely so that changes have a price before anyone argues about them.
• How fast it is answered. Vendor estimates within three working days; client approves or declines within three working days; silence is a decline, not an approval. The delay on my 7-month-late mobile build was, in the end, 11 weeks of unanswered change requests, not 11 weeks of engineering.

Our companion piece on change-request pricing for outsourced Node.js work in the UAE goes into the mechanics of the clause; the version there is the one I now paste into every SOW.

Step 5 — Add the 5 UAE Clauses the Generic Templates Leave Out

Most fixed-versus-T&M guides are written for US or UK contracts. Five clauses need UAE-specific treatment, and I have been burned by each of them at least once.

1. Governing law and forum

Match the law to where you would actually enforce. An onshore Dubai LLC contracting with an onshore vendor will usually sit under UAE federal law with the Dubai courts; a company in the DIFC or ADGM can choose those free zones’ common-law courts, which international vendors often prefer and which handle software disputes routinely. The mistake is inconsistency: a contract governed by English law, with an Arabic-language version that says something different, and a dispute clause pointing at a court that will not apply either. Pick one system, make the language of the contract match it, and have a UAE-qualified lawyer read the assignment and limitation clauses under that law.

2. IP assignment on payment, not on completion

Under a fixed price, the vendor holds the code until the milestone is paid; under T&M, you are paying for hours and the code should be yours as it is written. Say which, in writing, and make assignment effective on payment of each milestone or invoice rather than on final completion, so that a terminated project does not leave you with nothing. The full set of checks is in our guide to IP ownership in UAE developer contracts; the short version is that a vague assignment clause is worth nothing in either court system.

3. VAT treatment

UAE VAT at 5% applies to most software services supplied in the country; supplies from a foreign vendor may fall under reverse charge; some free-zone-to-free-zone supplies are treated differently. State whether prices are inclusive or exclusive, who accounts for the tax, and what happens if the treatment is later found to be wrong. On a fixed price, a 5% ambiguity is a 5% dispute.

4. Payment terms with a defined acceptance window

For fixed price: a milestone is invoiced on delivery, you have a fixed window (I use 10 working days) to accept or reject against the written criteria, silence is acceptance, payment is due 30 days from acceptance. For T&M: invoiced monthly against a signed timesheet, with a cap that cannot be exceeded without a written change. The acceptance window is the clause that stops a project stalling while a milestone sits unreviewed on somebody’s desk; it protects the vendor, and a vendor who is protected keeps their best people on your account.

5. Termination with a handover obligation

Both sides should be able to terminate for convenience on 30 days’ notice, with the vendor obliged to hand over code, documentation, credentials and infrastructure access in a usable state, and the client obliged to pay for work accepted or, on T&M, hours worked to the date of termination. Add a 14-day handover period at the T&M rate. This is the clause that made switching model at a package boundary painless for us, because both sides already know what the exit looks like.

Step 6 — Set the Reporting Cadence the Model Needs

The two models fail in different ways, and each needs a different weekly signal to catch it.

T&M fails by drifting. The signal is burn against cap: a one-page weekly report with hours by role, cumulative spend, percentage of cap consumed and percentage of scope delivered, side by side. If spend is at 60% and scope at 35% in week six, you have three weeks to act, not three days. Insist on this report from week one; a vendor who cannot produce it does not track their own hours, which tells you what their invoices are based on.

Fixed price fails by hiding. Because the vendor carries the cost risk, a struggling fixed-price project looks fine until the milestone is late. The signal is milestone acceptance with named criteria and a demo on a fixed day. I run a 30-minute demo every second Thursday against the acceptance criteria for the current milestone, whether the vendor thinks it is ready or not. What you are watching for is not the polish; it is whether the criteria being demonstrated are the ones in the SOW or a quietly narrowed version of them.

Step 7 — Run the 30-Day Review and Switch If the Score Was Wrong

Thirty days in, sit down with the step-1 scorecard and re-score every package using what you now know. The requirements you thought were a 4 turned out to be a 2 once the developers started asking questions; the integration you scored 3 got its sandbox credentials and is now a 5. Compare the new score to the model each package is on.

• A fixed-price package now scoring under 10: close it at the next milestone on the agreed terms, accept or reject against the written criteria, and re-issue the remainder as a capped T&M SOW. The change-request log from the first month is your evidence that the scope was not fixed; use it in the conversation with finance.
• A T&M package now scoring 14 or above: ask the vendor for a fixed price for the remaining scope. After four weeks of working together their estimate will be far better than at tender, and the premium usually drops toward 20%.
• A package whose score did not move: leave it alone. Switching models has a cost in paperwork and attention; do it only where the assumptions were wrong.

Because the contract was structured per package with a termination-and-handover clause (steps 3 and 5), each of these switches is a new SOW under the same MSA, signed in an afternoon, rather than a renegotiation. That is the whole point of the structure. In the two years since I adopted it, I have switched model at day 30 on 5 packages across 4 projects, and none of the 5 turned into a dispute.

The 4 Mistakes I Still See Every Month in Dubai

• Signing a fixed price to hit a board date. The date goes in the board pack, the scope is not ready, and the fixed price becomes a change-request negotiation by week six. Score first; if the score is under 14, tell the board a capped number instead of a fixed one.
• Treating a dedicated team as a fixed price. A dedicated pod is a monthly capacity commitment, which is a T&M variant with a predictable burn. Its contract needs the T&M clauses (timesheets, cap, weekly report), not milestones. Our note on building contract-management tooling in Dubai covers the systems side of tracking these commitments.
• No rate card in a fixed-price contract. Then the first change request is priced from scratch, under pressure, by the party that has the least reason to be generous.
• An English contract with an Arabic version nobody checked. If a bilingual contract is required, have the Arabic reviewed by a UAE-qualified lawyer and state which language prevails. Two of my 23 had conflicting versions; both were found out at the worst possible moment.

If You Also Contract Developers in Singapore

The seven steps travel, but the clauses in step 5 do not. Our Singapore team’s six checks for IP assignment in Singapore developer contracts replace the UAE IP and governing-law sections, and their guide to converting a contract engagement to a hire in Singapore is the equivalent of the switch in step 7 when the package you want to keep is a person rather than a vendor.

FAQ — Fixed Price vs Time and Materials for Developer Contracts in Dubai