Fixed Price, Time and Materials, or a Cap? A Risk-Transfer Matrix and the 20-35% Uncertainty Premium Hidden Inside Fixed Quotes

A 60-person instrument distributor needed a dealer ordering platform. The vendor was chosen; the sticking point was the last page of the contract, where that vendor offered three ways to price the same work: a fixed NT$1.68 million, NT$2,200 an hour (estimated 600–850 hours), or a NT$1.85 million cap billed against actual hours. Same team, same requirements, three numbers. The procurement lead asked the most common and most wrong question: which is cheapest? The right question is different. Which of these three structures puts the risk that requirements will change on whom — and how much are you paying for that insurance?

Myths Worth Breaking

• Myth 1: fixed price is safer. Reality: fixed price transfers risk to the vendor, so the vendor prices uncertainty in and hides it in the total. In our estimating experience, on projects with moderate requirement clarity that premium runs 20%–35%. You did not avoid the risk. You prepaid it.
• Myth 2: time and materials is a vendor's ATM. Reality: in most runaway T&M projects the vendor is not stalling, the client has no decision rhythm. An eleven-day wait for a design sign-off still bills as engineer idle time.
• Myth 3: a capped contract is just fixed price with a floor. Reality: a cap limits money, not scope. Without a change process attached, you get "budget spent, features unfinished," which is harder to resolve than a fixed price.
• Myth 4: the lowest quote is the most sincere. Reality: a conspicuously low fixed price usually means the vendor plans to recover margin through change orders. The thing to compare is the change-order rate and the rules that classify a change, not the headline total.
• Myth 5: the pricing model must be chosen up front. Reality: the effective approach is staged — T&M for the planning phase, then fixed price or cap for build depending on how clear the spec turned out. One model for the whole engagement is usually laziness, not rigour.

The Framework: Uncertainty × Change-Frequency Matrix

Score the project twice, 1 to 5 each.

Requirement uncertainty (U): complete spec and wireframes = 1; a verbal description and a competitor link = 5.
Expected change frequency (C): one decision-maker, already settled = 1; cross-departmental with an owner who adds requests = 5.

• U ≤ 2 and C ≤ 2 → fixed price. The spec is clear enough that having the vendor carry risk pays. Require a specification baseline document as a contract annex.
• U ≥ 4 or C ≥ 4 → time and materials. Any fixed price here is a high-multiple guess, and you pay a large premium for a contract that will still be amended.
• Everything in between → capped contract. Set the ceiling, bill actual hours, and define what happens to unused hours.

Formula for reverse-engineering the uncertainty premium:

Reasonable fixed price ≈ baseline hours × rate × [1 + 0.07 × (U + C − 2)]

At 700 baseline hours, NT$2,200 per hour, U=3 and C=3: 700 × 2,200 × (1 + 0.07 × 4) ≈ NT$1.97 million. A NT$1.68 million fixed quote means either the hours are underestimated or change orders will make up the difference — at which point the question to ask is the change-order rate, not the discount.

Three Scenarios Compared

Hidden Costs, Fully Listed

• Uncertainty premium: 20%–35% inside a fixed price, appearing on no line item
• Change-order administration: 2–4 combined hours per change (writing, estimating, confirming, approving), about NT$4,400–8,800
• Client-side decision latency: under T&M, three to six hours a week of waiting and rework can land on the invoice
• Scope dispute negotiation: fixed-price projects average two to five dispute meetings, 1.5 hours with two people per side, roughly NT$26,000–66,000
• Sunk unused hours under a cap: where the contract is silent, 8%–15% of budget is typically wasted
• Delayed launch opportunity cost: for a store doing NT$5 million a month, each month of delay defers roughly NT$150,000–400,000 of benefit
• Internal management hours: without a PM, the client-side contact spends four to six hours a week, over NT$250,000 annualised

Vendor Scorecard for Pricing Discipline

Score each 0–3, total 30. Below 18, do not hand this vendor a large fixed-price budget.

• Do they proactively assess requirement uncertainty rather than quoting immediately?
• Is the quote broken into comparable blocks rather than one lump sum?
• Do they provide a written specification baseline as a contract annex?
• Is the change-order rate disclosed in advance, and equal to the base rate rather than 1.5×?
• Are the rules defining a change written down (a useful model is how Taiwan's Government Procurement Act handles contract amendment)?
• Does T&M come with weekly task-level time detail?
• Does the capped contract state what happens to unused hours?
• Do they offer staged renewal rather than signing the whole engagement at once?
• Do they disclose actual headcount and roles, avoiding one person wearing four hats?
• Are acceptance criteria quantified — for example an LCP under 2.5 seconds instead of "it should feel fast"?

How ScriptWalker Maps, and Who We Are Wrong For

Four engagement models:

• Project (fixed price): for U ≤ 2 and C ≤ 2, from NT$350,000, including a specification baseline and a stated change-order rate
• Time and materials: NT$2,000–2,600 per hour with weekly task-level detail, suited to rollouts where the process is not settled
• Capped contract: a ceiling billed against actual hours, unused hours convertible 1:1 into maintenance hours within 12 months
• Advisory: NT$30,000–60,000 a month, no code, covering technology selection, vendor assessment and quote review

Clients we will send elsewhere:

• Those demanding a fixed price while refusing to provide any written spec or join a planning phase
• Those who want to "start building and discuss money later" — the most expensive opening available to both sides
• Those who treat change orders as after-sales service the vendor should absorb
• Those with a fixed budget more than 40% below the reasonable band, expecting negotiation to close the gap
• Those demanding lowest price, shortest timeline and maximum flexibility simultaneously

Kick-Off Playbook

• Week 1: a half-day scoring workshop on uncertainty and change frequency, both sides scoring separately then reconciling. Deliverable: U/C scorecard and recommended pricing model.
• Weeks 2–3: a T&M planning phase of roughly 40–60 hours, producing the specification baseline, wireframes and an hours range. No total price is discussed here.
• Week 4: rescore on spec clarity, formally choose fixed price or cap for the build, and sign the annex containing the change-order rate.
• Weeks 5–12: build. A 30-minute scope review every two weeks looking at three numbers only: hours consumed, scope completed, open decisions outstanding.
• Day 90: retrospective comparing actual hours against the original estimate range (healthy is within ±15%), and deciding whether to move to monthly maintenance.

Decision Checklist

• ☐ Do I have a written spec or wireframes?
• ☐ Is the decision-maker one person or three departments?
• ☐ Will the owner add requirements mid-project?
• ☐ Can I accept a total varying by ±20%?
• ☐ Is my budget a hard ceiling or an approximation?
• ☐ Have I asked for the change-order rate?
• ☐ Does the contract define what counts as a change?
• ☐ Can someone on my side decide within three working days?
• ☐ Am I willing to pay for a planning phase (8%–12% of total budget)?
• ☐ Does the contract say what happens to unused hours under a cap?
• ☐ Are acceptance criteria quantified rather than "as expected"?
• ☐ Can I accept staged renewal instead of signing everything at once?
• ☐ Have I counted my own contact's four to six hours a week as cost?

FAQ

Is fixed price always more expensive?

No. When requirement uncertainty is low — a complete spec and wireframes — the premium may be only 10%–15%, and budget certainty is usually worth that. What genuinely does not pay is forcing a fixed price onto vague requirements: the premium climbs past 35% and you still end up signing change orders.

How do I tell whether a fixed quote is padded?

Reverse-engineer the reasonable band with the formula above, then check three things: whether a specification baseline is provided, whether the change-order rate equals the base rate, and whether hours are estimated at feature level. A quote well below the band usually plans to recover through change orders.

How do I keep time and materials under control?

Three mechanisms: a monthly hours ceiling requiring approval to exceed, weekly task-level time detail, and a 30-minute scope review every two weeks. Also write a client-side decision deadline into the contract — the most overlooked and most impactful clause.

Do I get money back on an unused cap?

It depends entirely on the wording, which is exactly what gets omitted. Two reasonable forms: unused hours convert 1:1 into maintenance hours within 12 months, or are refunded at cost. Left unwritten, 8%–15% of budget is typically wasted.

Can the pricing model change mid-engagement?

Yes, and you should deliberately reassess at the end of the planning phase. The most effective pattern is T&M for planning, then fixed price or cap for build depending on spec clarity, with that switch point written into the original contract.

Next Step

If you are holding two or three quotes priced on different models, we offer a one-off quote and pricing model review at NT$12,000, delivering a U/C score, a reverse-engineered reasonable band and risk notes on the change-order clauses within two working days. Or take 30 minutes free to score your project before deciding anything.

• Email: [email protected]
• Phone: 0916-224-047
• LINE: @ufv9089p