How Government Tenders Are Actually Scored

Your team spent weeks on the submission. You lodged it before the deadline, and then — silence. Six weeks, sometimes twelve, and eventually an email saying you were unsuccessful on this occasion.

Somewhere in between, a small group of people read your tender, formed a view about your business, and put a number on it.

That process is not a mystery to the people inside it. It is documented, repeatable, and has a defined sequence, defined participants and a defined output. It is simply never shown to you.

This article describes it from the buy side: circa 15 years as a Government Procurement Manager, more than 1,000 tender responses personally received and reviewed, and over 500 evaluation meetings chaired.

Step 1: Your submission is checked for conformance before anyone reads it

Before a single word of your methodology is assessed, your submission goes through an administrative check. Was it lodged on time and through the correct portal? Are all required schedules and forms present and completed? Are the mandatory declarations signed? Have you responded to the addenda?

This is not an assessment of quality. It is a checklist, usually run by the procurement officer rather than the panel, and it is unforgiving by design — because it has to be. A procurement that lets one tenderer skip a mandatory schedule while holding another to it has a probity problem, and probity problems are the thing agencies fear most.

Alongside it sit any mandatory pass/fail requirements the tender sets: prequalification categories, accreditations, minimum experience thresholds, required certifications. There is no partial credit at that gate.

Submissions can be set aside here, before their content is ever considered. A strong response with a missing schedule and a weak response with a complete one are not in the same competition.

Step 2: Who actually reads it

An evaluation panel, usually an odd number of members — most commonly three to five — so that a position cannot deadlock. Typically a technical or project lead, someone with contract or asset knowledge, sometimes an operational end user, and a procurement representative who chairs the process and keeps it probity-compliant.

Two things about this group matter more than anything else in this article.

They score independently before they discuss your tender. Each member reads every submission alone and forms their own scores against the published criteria. Only then does the panel meet.

They can only score what is on the page. Panels are not permitted to assume capability, fill gaps from reputation, or ring you to ask what you meant. This is not a theoretical constraint. I watched very capable contractors — organisations the panel knew could deliver — omit required documents, leave out CVs for nominated personnel, or submit a program making no allowance for the agency's review periods when the tender expressly required it. That last one is quietly fatal: a program that ignores mandated review periods is inherently incorrect, and it tells the panel the contractor has not properly read the documents they are bidding under.

The committee knew those organisations were capable. We could only assess what was presented. It was never the agency's job to get a contractor across the line, and we could not show preference without rightly being accused of treating one tenderer better than another.

Step 3: Non-price is scored before anyone sees your price

Nearly every structured government evaluation runs in two stages.

Stage one is the non-price assessment. The panel scores your submission against the qualitative criteria — typically organisational capability, personnel, technical capability and methodology, management systems, and social procurement — without seeing your price.

Stage two is the commercial assessment. Only once the technical scores are settled is pricing evaluated and combined into a final ranking.

That sequencing is the most underappreciated fact in government tendering. Your technical response is judged purely on merit, uncontaminated by what you charged. Which means a genuinely superior technical submission can justify a higher price — the model is explicitly designed to allow it. How the two are combined varies by procurement, and some tenders publish weightings while others do not, but non-price criteria commonly carry substantial weight. Agencies award on value for money, not lowest cost.

Step 4: The moderation meeting is where scores actually get set

This is the part almost nobody outside the process has seen, and it is where submissions are won and lost.

The panel comes together with their independent scores. The chair works through each submission, criterion by criterion, and each member gives their score and their reasoning. Where scores diverge — and they always do — the panel discusses until it reaches a consensus score with a documented rationale.

In my experience those disagreements were rarely settled by force of argument. They were settled by evidence, most often one evaluator pointing to something in the submission the others had missed. There is an enormous amount for a panel to digest in a competitive evaluation, and even diligent evaluators miss things. The organisations that had tendered and won before understood this: they made the information that mattered impossible to miss — readable, and exactly where an evaluator would look for it.

Three things happen in that room that should change how you write.

Somebody has to say your score out loud and defend it. The member who scored you highest is asked why. If the answer is "they seem capable", the score comes down. If it is "section 4.3 sets out their staging approach for the occupied areas, and appendix B has the traffic management drawing for it", the score holds. Your submission's job is to hand that person their argument.

Disagreement gets resolved by returning to your document. When two members are three points apart, the chair asks them both to point at the page. The submission that is easy to navigate wins those exchanges. The one requiring assembly does not — and the member arguing for you tires of hunting faster than you would like.

The comparison is relative and immediate. Panel members have just read four other submissions on the same criterion. Generic content that reads acceptably on its own reads poorly next to a competitor who named the actual site constraints.

What a score actually measures

Most agencies score qualitative criteria on a defined scale — commonly zero to ten — with written descriptions of what each level means. The detail varies. The underlying logic does not: every score is a measurement of evaluator confidence.

A middling score does not mean your submission was bad. It usually means it was fine — compliant, competent and unconvincing. Most competitive submissions cluster in the middle of the scale, and what separates them from the small group at the top is not writing quality or page count. The top scorers removed the panel's uncertainty: every claim backed by specific verifiable evidence, every risk the panel was privately worried about already identified and addressed, and everything connected explicitly to this project rather than projects in general.

Having sat through more moderation discussions than I can count, evaluators are consistently asking five questions of every submission:

1. Will this organisation actually deliver?
2. Have they understood and addressed the real risks?
3. Is this value for money across the life of the project, not just the cheapest number?
4. Does it comply with everything we asked for?
5. Is any of this substantiated, or are we being asked to take it on faith?

Every paragraph either helps answer one of those questions or is costing you attention.

Where good tenders quietly lose

Losses at evaluation are rarely dramatic. In moderation, doubt shows up in quiet phrases — "I had some concerns here", "this felt generic", "I couldn't find where they addressed X". Each shaves points, and one point on a heavily weighted criterion can decide the outcome before price is opened.

The patterns are remarkably consistent: experience listed but never connected to the project at hand; methodologies that could have been written for any project anywhere; risks downplayed instead of confronted with credible mitigations; accreditations claimed without evidence; key information buried where a time-poor evaluator will not find it.

None of these make a submission non-conforming. All of them erode confidence — and confidence is the currency the entire process trades in.

Step 5: The recommendation goes up, not the scores

Once consensus scores are set and combined with the commercial assessment, an evaluation report is written recommending an outcome. It goes to a delegate or approving authority with power to accept, question or send it back.

Then comes the fact that surprises tenderers most: the panel's consensus scores are never divulged. Not at award, not at debrief, not ever. The reasoning behind your score has been written down, reviewed and filed, and the feedback you receive is typically as vague as "competitive" or "not competitive".

What you can do about the parts you control

You cannot change the panel, the weightings or the competition. You can change three things, and all three are decided before you lodge.

Answer the question that was asked. Map your response to the published criteria and their stated assessment requirements, in the order the tender sets them out. Do not reorganise the panel's structure into yours; you are asking busy people to do translation work while holding a scoresheet.

Replace claims with evidence. "We have extensive experience in occupied school sites" is a claim, and a panel cannot score a claim much above the middle of the band. Two comparable projects, named, with the specific constraint they shared with this job, is evidence.

Be specific to this project. Reusable content is efficient and it is exactly what reads flat in a moderation meeting. The project-specific paragraph is what a panel member quotes when defending your score.

The gap between knowing and scoring

Understanding this process is necessary and not remotely sufficient. The panel does not score what you know about evaluation. It scores what your submission demonstrates, criterion by criterion, through the eyes of people trained to distrust unevidenced claims.

That is the gap most tenderers cannot see from the outside: the distance between a submission that feels strong to the team that wrote it and one that scores strong with the panel that reads it.

ProcureHQ exists to close it. The Digital Evaluation Committee assesses your response against the evaluation criteria the platform extracts from your own tender documents, using an Evaluation Engine built on the methodology described above — including circa 150 recurring Decision Signals drawn from how panels consistently assess submissions. You get a current score, a potential score, criterion-by-criterion findings, the evidence the assessment relied on, and a recovery plan for the gaps.

Two honest limits. We assess the non-price criteria — price, commercial value and value for money are deliberately excluded, because those are the agency's judgement. And we do not run a line-by-line compliance check against your specifications; the assessment is scoped to the published criteria and assessment requirements, read in the context of the tender's scope and technical documents.

Your first tender is free. One full evaluation and one complete Smart Submission Blueprint, per organisation, at no cost — so you can see your score before a panel does.

---

Frequently asked questions

How many people sit on a government tender evaluation panel? Usually an odd number, most commonly three to five: a technical or project lead, someone with contract or asset knowledge, sometimes an operational end user, and a procurement representative who chairs the process. Panel composition is recorded as part of the probity trail.

Do evaluators score my submission together or separately? Separately first, then together. Independent scoring comes first so no panel member anchors the others, then a moderation meeting sets consensus scores with a documented rationale.

Does the panel see my price while scoring my technical response? In a structured two-stage evaluation, no. Non-price criteria are scored before the commercial assessment, which is why a technically stronger submission can justify a higher price.

Can an evaluator take my past performance into account? Only where the tender makes past performance a criterion and asks you to evidence it. Panel members bringing in personal knowledge that is not in your submission creates a probity risk, and well-run panels guard against it.

Why won't the agency tell me my score? Consensus scores are not released to tenderers. Many agencies offer a debrief and it is worth requesting one, but a debrief is a summary delivered after the decision, not the scoresheet.

---

Recipient of the RICS High Achievement Award (Royal Institution of Chartered Surveyors).