Staff augmentation, managed services, fixed-bid systems integration, time and materials, outcome-based engagements, and co-development are six labels for buying delivery work. They are not six exclusive alternatives. Each fixes one of four variables that vary independently, which is why a single contract can honestly be described by two of the labels at once.
Four variables
Direction. Who decides what gets worked on next, and who accepts the work.
Price basis. What the money is fixed against: a total, an hourly rate, a milestone, a ceiling, or a measured result.
Continuing operation. Who runs the thing after it exists, and who is answerable when it stops.
Transfer object. What changes hands at the end, and what stays with the supplier.
A label pins one variable. The other three remain open and have to be written down.
What each label actually fixes
Staff augmentation fixes direction. Individuals or teams work alongside buyer staff under buyer direction, and the buyer’s service owner stays answerable for the whole. It does not follow that the supplier carries nothing: UK framework guidance for buying specialists states that the supplier retains responsibility and liability for its specialists’ work. Direction and liability move separately. Added capacity is a plausible choice where the recipient can direct and integrate the work, and where that capability is absent, adding people does not supply it.
Managed services fixes continuing operation. The supplier takes agreed responsibility for named ongoing activities — provisioning, monitoring, patching, and backup, in the scope AWS publishes for its own managed service. The boundary is the whole content of the arrangement: supported components, service hours, change authority, customer dependencies, exclusions, and exit assistance. Retained external operation is a legitimate destination rather than a failure to transfer, and it is compatible with exit planning that contemplates renewal, a replacement supplier, or bringing the work in house.
Fixed-bid fixes price. Under US federal rules, a firm-fixed-price contract excludes adjustment based on the contractor’s cost experience, and the wider fixed-price family admits adjustment mechanisms. It suits work where performance uncertainty is identifiable and its cost impact reasonably estimable, and where the contractor accepts that risk knowingly.
Time and materials fixes the rate and leaves the quantity open. It uses specified labour rates and material costs, and it requires surveillance and a ceiling. Neither of those converts it into a fixed price. A ceiling is a control that forces a decision before more money is committed, not a commitment that the work completes beneath it.
Outcome-based fixes the acceptance measure. Performance-based acquisition calls for required results, measurable standards, and a method of assessment, with incentives where appropriate. Three things get collapsed under this label and are distinct: specifying a result, accepting work against a measure, and making payment contingent on a downstream business benefit. Only the third moves value risk to the supplier, and it is the one performance-based rules do not require.
Co-development and venture studios fix the transfer object, and fix it furthest from the software. Collaboration guidance sets out several allocations of result ownership and use rights, and the standard bilateral models do not address jointly owned intellectual property at all. A studio can retain equity in exchange for funding and build services, as High Alpha states of its own arrangement. A collaborator holding a continuing stake is compatible with some duties transferring; equity is not free labour, and code access is not ownership.
The rule
The four variables are independent, so the label settles at most a quarter of the arrangement. Staff augmentation says nothing about price. Fixed-bid says nothing about who operates the result. Outcome-based says nothing about what transfers. A procurement that selects a model and expects the rest to follow has chosen one answer and inherited three defaults it never saw.
Where the labels fail
Two failures, from different causes.
Accountability read off the label. Neither “we’re using staff augmentation, so the risk is ours” nor “it’s outcome-based, so the risk is theirs” survives the contract. Supplier liability persists under buyer direction, and an outcomes label does not remove buyer duties to specify, to supply inputs, and to accept. The allocation is whatever the terms say, and the label is not a term.
A measure gamed by routing. Where payment attaches to a measured result, the measurement window becomes the target. A supplier rewarded for average handling time falling from ten minutes to eight can reach eight by diverting difficult cases elsewhere, and the before-and-after comparison then describes two different workloads. Contingent payment therefore requires the population, the baseline, the eligibility rules, the exception counts, and the treatment of changed conditions to be fixed in advance. Without those, the number is descriptive and the payment is arbitrary.
Not to be confused with
Insourcing. Continuing external operation and a plan to transfer later are compatible. Generic contract-expiry language is not a transfer plan, and an exit clause is not evidence that anyone can run the thing.
A ranking. None of the six is more mature or more advanced than the others. They answer different questions, and the right one is determined by which variable the buyer most needs pinned.