T&M, Fixed Price, Staff Augmentation: the CIO Guide

Somewhere between the business case and the signature, someone asks: Time and Material, fixed price, or do we simply bring in extra engineers? The answer usually gets settled by procurement habit, by whatever the last programme used, or by whichever line looks cheapest on the comparison sheet. All three are bad reasons.

This guide is for the person who will have to defend that choice a year later. Three models, what genuinely separates them, when each is right, and the specific way each one fails. On the narrower question of paying for output rather than for time, we went deeper in Output-based vs Time & Material (/en/insights/output-based-vs-time-material).

The three models, defined by who carries what

The commercial labels are close to interchangeable. The contracts are not. Each model answers the same four questions: who absorbs the cost when the scope moves, who decides what gets built next, who manages the people day to day, and what success looks like contractually.

Time & Material

You buy engineering capacity, billed on time spent, against a backlog you control. The scope stays yours and is allowed to move as you learn. The provider commits to the quality of the profiles, their availability and a way of working, not to a fixed content at a fixed date. Drift risk sits on your side, and success is measured by what actually shipped.

Fixed price, often sold as turnkey delivery

You buy a defined result for a price agreed upfront. The provider commits to content, quality level and date, carries the drift risk inside the signed scope, prices that risk into the number, and protects itself with change control. You keep the acceptance criteria and the arbitration. Turnkey delivery is the strongest form, where the provider owns the method, the team composition and the delivery chain end to end.

Staff augmentation

You bring in named individuals who join your teams and report into your management line. The provider commits to the person, to replacement, and to the administrative side. You direct the work and own the backlog, the technical decisions and the quality bar. Drift risk is entirely yours, and success is measured the way you would measure an internal hire.

One confusion is worth clearing early. Time & Material is not staff augmentation. Both bill by time, and that is all they share. A well run T&M engagement gives you a team with its own lead, its own practice and collective accountability for delivery. Staff augmentation gives you individuals, and you supply the practice. Some providers sell the second while pricing it like the first.

Price is the wrong first criterion

Comparing a daily rate against a fixed-price envelope compares two things that are not comparable. A fixed price contains a risk premium, because the provider is selling certainty and has to price the uncertainty it absorbs. On a stable scope that premium is small and buys real predictability. On an uncertain scope it is large, and either the provider priced the risk honestly, and you pay heavily for a certainty your first change of mind will break, or it underpriced the risk and you pay in another currency: who actually gets staffed, what quality ships, a renegotiation halfway through.

Staff augmentation looks cheapest per unit, and usually is. What never appears on the sheet is the management load it transfers to you: onboarding, review, technical direction, performance management, ownership of the result. If your management line has that capacity, the saving is real. If not, you have bought a second job for a manager who was already full. So the real question is not price, but these two.

Criterion one: how uncertain is the scope, honestly

The useful test is not whether you have a specification. Everyone has a specification. The test is whether you could write the acceptance criteria today and still recognise them as correct in six months. Scope is genuinely uncertain when any of these is true.

• What to build next depends on what you learn from real users after a first release.
• A key upstream system is undocumented, or owned by a third party who has committed to nothing.
• Business or regulatory rules are still being decided while the work starts.
• Feasibility itself is part of what you are buying.
• Stakeholders have not had the argument they need to have, and the specification papers over it.

Uncertain scope plus fixed price produces an amendment machine, and the provider raising change requests is not being difficult, it is doing the only rational thing the contract allows. Stable scope plus Time & Material is the mirror image: you pay for flexibility you do not need and give up predictability you could have had almost for free.

Criterion two: your own capacity to steer

The second variable sits on your side of the table, and buyers skip it. Each model demands something from the client organisation, and a model whose demands you cannot meet will underperform however good the provider is.

• Time & Material needs a product owner with real arbitration authority available weekly, a definition of done that someone enforces, and the willingness to say stop.
• Fixed price needs the ability to specify precisely and then hold still, someone who runs acceptance seriously, the discipline to route informal requests through change control rather than around it, and clean ownership of your dependencies.
• Staff augmentation needs management capacity to onboard, review and coach people who are not yours, a technical standard that exists in practice rather than on a wiki page, and someone internal who owns the outcome.

Cross the two criteria and most decisions make themselves. Stable scope with thin steering capacity points to fixed price. Moving scope with strong steering capacity points to Time & Material. A bounded gap inside a team that already runs well points to staff augmentation. Moving scope with thin steering capacity is the dangerous quadrant: buy a team that comes with its own delivery leadership, or fix the steering problem before signing anything.

Fixed price: right, then a trap

Fixed price is at its best on work with clean edges. A migration where the target state is known, an integration against a documented API, a regulatory deliverable whose content is dictated from outside. The provider can price the risk accurately and everyone plans around the date.

It becomes a trap when the scope was fuzzy at signature and both sides pretended otherwise. The failure is quiet: reporting stays green while the conversation shifts from what the product needs to what a sentence written months ago actually meant. Watch for these.

• Change requests growing faster than delivered scope.
• Discussions moving from "is this right for users" to "is this in scope".
• The provider's strongest people quietly rotating off your account.
• Acceptance turning into a negotiation rather than a test.
• Your teams routing around the contract informally, because the formal path is too slow.

Time & Material: right, then a trap

Time & Material is at its best on a living product: continuous discovery, priorities that reorder, a team that stays long enough to accumulate context. It follows reality instead of fighting it, and lets you change your mind without paying a contractual toll every time.

It becomes a trap without steering discipline. Nobody is lying, work is genuinely being done, and no single week looks wrong. What is missing is the force that makes someone ask whether this is the highest value thing available. The engagement drifts into an annuity, comfortable for both sides, and the drift is only visible in aggregate. Watch for these.

• You cannot state in one sentence what the team is trying to achieve this quarter.
• The roadmap is a description of what the team is already doing.
• Headcount on the engagement only ever moves in one direction.
• No one can remember the last time unused capacity was handed back.
• Steering meetings review activity instead of taking decisions.
• The provider has never told you a task is not worth doing.

Staff augmentation: right, then a trap

Staff augmentation is at its best when a team that already works well has a bounded gap: a surge in load, a scarce skill for one phase, a hiring pipeline running behind a real need. The practice, the standards and the ownership already exist, and you are adding hands to a machine that runs.

The first trap is that the temporary becomes structural. A peak becomes the baseline, renewals stop being argued, and you carry a permanent workforce on a temporary contract, paying a flexibility premium for flexibility you no longer exercise.

The second trap gets discussed less and does more damage: knowledge transfer running backwards. External engineers accumulate the deep operational knowledge of your system, because they do the work, while your permanent staff drift toward coordination and ticket shaping. When the contract ends the knowledge leaves with it, which is exactly why the contract never ends. Watch for these.

• One external contributor is the only person who understands a critical component.
• Your permanent engineers spend more time coordinating than building.
• Renewals happen automatically and are never re-argued on merit.
• The honest answer to "what if this person left next month" makes people uncomfortable.
• External contributors carry on-call duty but sit outside architecture decisions.

None of this argues against the model. It argues for deciding upfront which capabilities you own permanently and which you are content to rent, then defending that line.

What to ask before you sign

Generic due diligence produces generic answers. These questions discriminate between providers.

Fixed price or turnkey delivery

• What does the price cover, and name three things it explicitly does not.
• Who prices a change, on what basis, and how fast.
• What happens to your commitment if one of our dependencies slips.
• Will the people who wrote this proposal be on the delivery team.

Time & Material

• Who leads this team, and can that person push back on us.
• How do you report on work done, and how would we detect padding.
• What do you do when the team finishes early.
• What would make you tell us to stop doing something.

Staff augmentation

• Who manages performance, and who has the difficult conversation.
• What does your replacement commitment actually cover.
• What is your handover protocol when someone rolls off.
• If we want to hire this person permanently, how does that work.

Hybrids and mid-programme switches

Real programmes are rarely pure. A fixed-price framing phase followed by a Time & Material build is the most reliable answer to the moving-scope problem, because nobody has to pretend the scope was known before the work started. A Time & Material core team with fixed-price satellites also works, until the satellites depend on decisions the core team has not made yet. A capped Time & Material arrangement gives finance a number without freezing the content, but only if someone is genuinely prepared to cut scope as the ceiling approaches.

Switch when the underlying variable changes, not when the relationship gets tense. A scope that stabilised after discovery can move to fixed price for the next block. A fixed-price engagement that turned out to be exploratory should be stopped and reframed rather than ground through change requests. And a model change is a governance event, not an invoicing change: rebaseline, restate who decides what, revisit what gets reported. Changing the invoice format while keeping the same operating model changes nothing.

A decision frame for your next steering meeting

Four questions, in this order, because each one removes options.

1. Could we write the acceptance criteria today, and would they still be right in six months? If yes, fixed price is on the table. If no, take it off, whatever finance prefers.
2. Do we have a product owner with genuine authority and weekly availability? If no, fix that first: unsteered Time & Material is the most expensive of the three.
3. Does our engineering management have spare capacity to onboard, review and coach people who are not ours? If no, we would be buying individuals we cannot direct.
4. Is this a capability we must own long term, or one we are content to rent? Never rent the first.

If more than one model survives, the choice is genuinely open: decide on price, on speed of mobilisation, or on which provider you trust more. If none survives, the problem is not the contract model. The programme is not ready to be bought yet, and saying so before signature is worth far more than saying it after.

We sell all three, which is why this article does not end with one of them being the right answer. Each works extremely well and each fails badly, and the failures rarely come from the model itself. They come from a model chosen against the reality of the scope, or against the client's real capacity to steer it.