A sole-source award is defended by three records built during the transaction, not after it: the necessity record (why competition was unavailable or inadequate), the value record (independent evidence of what the deal should cost), and the integrity record (who decided what, and what separated execution from oversight).
Reviewers rarely argue that a direct award was inherently wrong. They ask whether the government could demonstrate, at the moment of decision, that it knew what it was doing. All three records must exist contemporaneously; a file assembled in response to an audit query answers a different question from the one being asked.
What you will learn
- The question reviewers actually ask, and why a good commercial outcome does not answer it.
- How to build the necessity, value and integrity records while the transaction is live.
- Why advice that was considered and rejected strengthens a file rather than weakening it.
Written by the G2G Deal Design practicepowerabode DMCC's neutral buy-side practice for government-to-government procurement. We hold no position in the transactions we advise on — no cargo, no supplier, no equity. People, standards and certifications.
What reviewers actually test
Public accounts committees and supreme audit institutions work from a narrow question: was the decision defensible on the information available when it was made? That framing matters. It means a sole-source award that produced a good commercial outcome can still fail review, and a deal that later suffered from market movement can pass it.
Three lines of enquiry follow from that question, and each corresponds to a record the buying government either has or does not have.
Record one: necessity
The necessity record explains why an open process was not run. It is the weakest part of most files, because the reasoning is often obvious to the officials involved and therefore never written down.
A sufficient record identifies the specific ground relied upon under the applicable procurement framework, and evidences it:
- Sole realistic source. What market testing established that no alternative supplier could meet the requirement — and how recently.
- Urgency or security of supply. What the timeline required, what created it, and whether the government's own delay contributed to it. Self-inflicted urgency is a finding, not a justification.
- Failure or compromise of a prior process. What went wrong, what was concluded, and what changed to prevent recurrence.
- State-to-state framework. The instrument under which the arrangement sits and the authority relied upon to conclude it.
The record should also show what was considered and rejected. A file that presents a single route as inevitable, with no evidence that alternatives were examined, invites the conclusion that they never were.
Record two: value
Necessity explains the absence of competition. It says nothing about price. The value record supplies what the tender would otherwise have produced: an independent view of what the contract should cost, formed before the price was settled.
This is the benchmarking file — published index references, netback or freight parity, comparable transactions and cost build-up, reconciled into a range with divergence explained. Two attributes determine whether it holds up under examination:
- It predates the price. A dated method note, approved before substantive negotiation, is the single most useful document in the file.
- It is reproducible. A reviewer working only from the named sources should reach the same range.
Value also extends past the unit price. Indexation, quality adjustment, delivery tolerance, penalty mechanics, termination and change provisions all move realised value, and reviewers increasingly test them.
Record three: integrity
The integrity record answers a question officials often find uncomfortable: who could have changed the answer, and would anyone have known?
In most direct-award structures the same team scopes the requirement, selects the counterparty, negotiates the price and assures the outcome. That is a structural conflict, and it is visible from an organisation chart without any inference about individual conduct. The remedy is separation: an assurance function that reports independently of the negotiating team, with the authority to record a dissenting view that travels to the decision-maker intact.
The record should establish:
- The delegation under which the award was made, and its limits.
- The decision points, dated, with the material available at each.
- The separation between execution and oversight, and how dissent was handled.
- The conflict declarations obtained, and from whom.
- The advice received, including advice not followed and the reason.
Advice not followed is not a weakness in the file. A record showing that a recommendation was considered and consciously rejected for stated reasons demonstrates a functioning decision process. A record showing only advice that was accepted suggests either an unusually smooth transaction or an incomplete file — and reviewers rarely assume the former.
Sequence, not assembly
All three records must be built while the transaction is live. This is the point at which most files fail, and it is not a documentation problem — it is a sequencing problem. A necessity assessment written after the counterparty is selected, or a benchmark produced after the price is agreed, cannot demonstrate that either constrained the decision.
Practically, this means the assurance workstream starts when the requirement is defined, not when the contract goes to signature. It also means the file is closed and retained at award, in a form that a reviewer with no institutional memory can follow years later, after the officials involved have moved on.
What causes findings
- Necessity asserted, not evidenced. The ground is named; nothing establishes it.
- Value tested by the counterparty. The supplier proposes both the price and the reference against which it is measured.
- Execution and oversight in the same reporting line. No mechanism existed by which a problem would have surfaced.
- Retrospective file assembly. Documents dated after the decisions they purport to support.
- Mechanics left unexamined. Price reviewed, but indexation, penalties and change provisions never assessed.
Key takeaways
- Reviewers ask whether the decision was defensible on the information available at the time — not whether the outcome was good.
- Build three contemporaneous records: necessity, value and integrity.
- Necessity must be evidenced, not asserted; show what alternatives were examined and rejected.
- Separate execution from oversight, and preserve dissent in the record.
- A file assembled after an audit query answers a different question from the one being asked.