What Counts as Proof

There is one part of your track record that Defence will not take your word for.

On Defence estate tenders where the Australian Skills Guarantee applies, the current tender templates state that the Commonwealth will obtain your past performance and reporting under that policy independently, from the Skills Guarantee Online Reporting System. Whatever you write, Defence checks the government's own record.

For almost everything else, including your previous projects, your people and their workloads, Defence starts with the evidence you supply.

Edition 18 showed that the evaluator only scores what is on the page. This edition asks a harder question: what makes that page believable?

Defence has written down what "substantiated" means

On Defence's capital works tenders, evaluators score each weighted criterion out of ten using a guide published in the CFI Tender Evaluation and Probity Protocol (October 2025). Each band describes how convincing a response is, how well it understands the requirement, the risk to Defence and how well its claims are supported. The language on claims steps down band by band:

9, Outstanding: all claims are substantiated

7, Good: most claims are well substantiated

6, Fair: most claims are substantiated

5, Acceptable: only some claims are substantiated

4, Marginal: claims are generally not substantiated

Edition 15 looked at the two ends of this scale. The commercial action is in the middle. Part of what separates a 6 from a 5 is the difference between "most" and "only some" claims being supported. On a criterion weighted at, say, 30 per cent, that single point is worth three percentage points of the total technical score.

Substantiation is not the only factor. The protocol tells evaluators to choose the band that best fits overall, and not every element must apply. But it is the factor a bid team controls most directly. As Edition 18 noted, the protocol belongs to Defence's Capital Facilities and Infrastructure branch, and we have not found a published equivalent for Estate Works Program evaluations.

The schedules turn claims into facts

For past projects and people, the estate templates do not ask for a capability narrative. They ask for fields.

For each project in Tender Schedule C (Previous Performance), Defence asks for the client, value and duration, current status, your involvement, the key risks and how you handled them, innovations, how the project shows you can do this work, whole-of-life outcomes, any material non-compliance on time, cost or quality, and a client referee. The Estate Works Program Medium Works template sets the window at the last two years, caps it at five projects, and adds a table on proactive work health and safety management.

Tender Schedule B does the same for people. For each key person it asks for hours per week (on a 37.5-hour week), current and future workload on other projects, and experience linked to a CV of no more than three pages that names a client referee.

Each field attaches something checkable to a claim. Compare two answers for the same role, written here as an illustration:

• "Our project manager has extensive Defence experience and will be dedicated to this project."

• "Project manager: 30 hours a week on this project. Currently completing a $4.2 million amenities upgrade, finishing in March. Referee: the client's contract administrator."

The second can be verified. On our reading of the scale, that is what moves a response from "some" claims supported towards "most".

Paper first, then verification

On capital works tenders, the written evidence earns the first score. Board members score each submission independently against the guide. Referee reports can be requested early, but the board cannot have regard to them until preliminary scores are agreed.

Only then does verification begin. Referees are asked a substantially consistent set of questions. Where used, key personnel interviews can test previous experience, including validating CVs, and current workload. Any score change that follows must be justified in writing.

Two consequences follow. Your schedule sets the starting score, and verification can confirm it or wear it down. And your schedule must match what your people say in the room. If Schedule B gives your project manager 30 hours a week and the interview reveals three other live projects, the gap is on the record.

The national auditor expects this kind of testing. In a 2018 audit of a Defence fuels procurement, the ANAO found that claimed past performance had been assessed on what tenderers submitted, with no evidence that incumbents' claims were tested against performance information that should have been in Defence's own systems. That was fuel, not estate works, but the expectation is plain: claims should be checked.

Defence does keep score, at the top end

Defence runs its own Supplier Rating System, which rates contract performance every six months. Its third roundcovered 316 contracts across 139 suppliers, worth about $92 billion, and Defence lists informing tender evaluation among its aims.

The published threshold in May 2023 was contracts managed by Defence's Capability Acquisition and Sustainment Group valued above $20 million, with a pilot extending coverage. We found no published detail on whether it reaches estate works of the size most SMEs bid for. For those firms, the main record Defence sees is the one they write and their referees confirm.

Research shows why buyers care. When an Italian utility began scoring suppliers' past performance in contract awards, average performance rose from 25 per cent to 90 per cent of audited parameters, with no significant overall change in price. It is a working paper on utility maintenance, not Defence construction, but the lesson travels: recorded performance changes behaviour.

How to write evidence that scores

1. Choose projects for relevance, not size. Each one must answer how it shows you can do this work.

2. Put a number on every claim. Values, durations, dates and hours give an evaluator something to confirm.

3. Confirm every referee. Check the phone number is current and that the person knows the project well.

4. Disclose non-compliance with the fix. Edition 14 covered why Schedule C asks for it.

5. Make Schedule B match reality. The workload you list may be tested in an interview.

The record you write

On the page, Defence is not scoring your record. It is scoring the evidence of it. Two firms with similar histories can land a band apart, because one wrote facts a referee can confirm and the other wrote adjectives.

Before you lodge, read each claim and ask one question: if an evaluator rang the referee about this line, what would they hear? If you are not sure, the line is not finished.

Next week, Edition 20 looks at the requirements that sink capable bidders.

This is Edition 19 of the Australian Defence Tender Framework series by Visionex Solutions. The series publishes every Wednesday. Series 2: Inside the Tender runs from 16 September to 2 December 2026.

How does your team decide which projects make it into Schedule C?

Between editions, follow us on LinkedIn for Australian defence industry news, facts, and market insights.