You've found a tender that looks winnable. The service is within your capability, the methodology is familiar, and your price is competitive. Then you read the evaluation table properly and realise the buyer isn't scoring “quality” as one broad idea. They're scoring several weighted questions, applying a formula, and possibly normalising the result before combining it with price.
That architecture shapes the bid before anyone writes the executive summary. Tender evaluation criteria scoring rewards relevant evidence placed against the right criterion, not polished prose. The UK government's guidance on assessing competitive tenders makes the underlying principle clear. Criteria must relate to the contract, remain clear and measurable, and be applied using the published methodology.
Why Scoring Architecture Beats Writing Skill
A panel can receive two technically credible bids and still produce a clear winner. One response may spread excellent evidence evenly across every question. The other may identify the heavily weighted sub-criteria, put its strongest proof there, and answer lower-value points efficiently. The second bid can win without being more entertaining to read.
That's familiar territory for anyone who's sat through evaluator debriefs. Assessors rarely award marks because a paragraph sounds elegant. They award marks when they can locate a complete, credible answer to the published requirement, then connect it to the scoring rubric without making assumptions.
Read the grid before the question
A local authority or CCS panel assessing a major contract may decide that the second-cheapest bidder offers the stronger overall proposition because its quality score creates enough separation. The exact contract value and result will depend on the procurement, but the lesson is consistent: price doesn't operate in isolation.
Start by extracting four things from the tender pack:
- Headline weighting: How much quality, price, social value, or other published element contributes to the total.
- Sub-criteria: The separate areas that make up the quality score.
- Question-level weight: The marks attached to each response, rather than the apparent importance of the subject.
- Evidence instruction: The documents, examples, methodology, credentials, or commitments the evaluator expects to see.
A bid writer who starts drafting immediately often gives every topic similar attention. That feels thorough, but it can be strategically weak. Your tender monitoring process should identify opportunities whose scoring structure fits your strengths, while your knowledge base should help locate evidence for the criteria that matter most.
Practical rule: Treat the evaluation table as the bid's operating system. The narrative is only the interface.
The architecture also limits what buyers can do during evaluation. That matters because suppliers can rely on the published framework, and because every strategic choice should begin with the rules governing it.
The Rules Every UK Evaluator Must Follow
UK buyers don't have unlimited discretion once a tender is published. The current government guidance states that award criteria must be connected to the contract subject-matter, sufficiently clear, measurable and specific, and proportionate to the contract's nature, complexity and cost. It also requires bids to be assessed against the published criteria and methodology, as set out in the official procurement guidance.
In practical terms, an evaluator can't decide halfway through assessment that one criterion now matters more than the published weighting. They can't award marks for a factor that wasn't included in the tender documents, and they shouldn't rely on an impression that isn't supported by the bid.
Five rules that lock down UK tender scoring
| Rule | What Evaluators Cannot Do | What Bidders Can Rely On |
|---|---|---|
| Published criteria | Introduce a new scoring factor after submission | The stated criteria define the assessment |
| Published methodology | Change the calculation informally | The disclosed formula and process should govern |
| Equal treatment | Apply a different test to one supplier | Comparable evidence should face comparable assessment |
| Proportionality | Demand irrelevant or excessive evidence | Responses should relate to the contract's needs |
| Auditability | Remove the reasoning behind score changes | Scores and moderation should leave a defensible record |
The legal labels matter less to a bid team than their operational effect. If the buyer asks for a mobilisation plan, answer mobilisation. Don't substitute a general corporate policy and expect the panel to infer delivery capability.
Clarifications also have limits. A buyer may clarify an ambiguity or confirm an existing point where the procedure permits it, but a clarification shouldn't materially rewrite a bid or give one supplier an opportunity to submit a better answer after the deadline.
Consensus scoring adds another control. Evaluators may score independently before a moderator or chair leads discussion, but the final reasoning should explain how the panel reached its agreed mark. That record protects the buyer and gives bidders a more useful debrief.
For bid teams, the implication is straightforward. Mirror the buyer's terminology, preserve the criterion order, and make every claim traceable to evidence. Bidwell's guides for public sector bidding can sit alongside the tender pack as a working reference, but the published procurement documents always control the bid.
Common Weighting Models and What They Really Mean
A bid team can spend days refining prose and still lose ground if it misreads the scoring architecture. The headline split determines how much a strong technical answer can offset a commercial disadvantage, while the sub-criteria decide where review time and evidence should go.
UK procurement documents use several recurring models. A Contracts Finder evaluation template sets 70% quality and 30% price, another public-sector example uses 60% quality and 40% price, and UK commercial guidance has used an 80% technical and 20% price structure for certain competitions. These examples are documented in the published evaluation template.
Read the split as a strategic signal
| Quality/Price Split | Typical Use Case | Example Sub-Criteria Weightings |
|---|---|---|
| 80/20 | Technically demanding services where delivery quality carries substantial importance | Methodology, team capability, mobilisation, risk management, and technical assurance |
| 70/30 | A common balance for professional services and other quality-sensitive procurements | Methodology, team and CVs, mobilisation, social value, and risk management |
| 60/40 | Procurements where buyers place greater emphasis on commercial competitiveness | Technical approach, delivery model, people, social value, and price |
| 100/0 | Quality-only evaluation where the published procurement structure permits it | Technical response, delivery approach, capability, and contract assurance |
The headline ratio is only the first layer. A 70/30 model may divide quality across several questions, each carrying a separate weighting. One published 60% quality model allocates 23% to the professional team, technical expertise and project management, 15% to design concept, 12% to consultation approach, and 10% to social value. A weak response in one category reduces the total, even if the wider narrative is persuasive.
That allocation should shape the bid plan before drafting starts. High-weight questions need stronger evidence, earlier review, and clearer links between commitments, delivery controls, and outcomes. Lower-weight sections still require compliance, but they should not consume the same writing effort.
Price is often scored separately and added to the quality result. The Contracts Finder example gives the lowest price the full available price marks and uses Price Score = (Lowest Price / Tenderer Price) × 30 for a 70/30 model. The formula rewards a lower compliant price, not an artificially low promise. Commercial teams must test whether reduced pricing leaves enough capacity to deliver the proposed service and protect margin.
A quality point has no universal cash value. Its effect depends on the published scale, formula, number of bidders, and contract structure. Run scenarios against the buyer's method rather than applying a generic rule.
A 70/30 split therefore directs scarce review time toward weighted quality questions while requiring a price strategy that can survive the published calculation. Weighting, normalisation, and consensus scoring should influence the response plan before a single paragraph is written.
Building a Weighted Scorecard That Holds Up
Build the scorecard before drafting. Start with the buyer's award criteria, not with the solution your team already wants to sell. The scorecard should translate the published structure into observable evidence that a reviewer can find and assess consistently.
A workable build sequence
- List the published criteria. Record the exact wording, weighting, response limit, mandatory requirements, and requested attachments.
- Define observable evidence. For a delivery criterion, that might include roles, milestones, governance, dependencies, risk controls, and continuity arrangements.
- Choose the permitted scale. A published NHS-style template uses a 0 to 5 quality scale, then multiplies the score by the relevant criterion weighting and overall quality weighting. The NHS-style scoring attachment also describes consensus scoring for the final technical result.
- Write score anchors. A score of zero should mean the response hasn't addressed the requirement. A middle score should represent an acceptable answer. The highest mark should require complete, evidenced, low-risk coverage, not confident language.
- Test for overlap. If “governance” appears in methodology, implementation, and risk management, decide whether the same evidence can legitimately earn marks in each place.
- Assign ownership. Give each evidence item an owner, source location, approval status, and final response destination.
Use this example for a criterion worth 15 points:
| Criterion | Sub-criterion | Weight | Evidence to provide |
|---|---|---|---|
| Implementation approach | Mobilisation | 6 points | Plan, responsibilities, dependencies, readiness checks |
| Implementation approach | Governance | 5 points | Meeting structure, escalation route, reporting and decision rights |
| Implementation approach | Business continuity | 4 points | Continuity plan, resilience controls, recovery responsibilities |
The raw score still needs to follow the buyer's method. If evaluators use a 0 to 5 scale, your internal scorecard can show the raw mark, the evidence supporting it, and the weighted contribution. Don't replace the published method with an internal formula and then mistake the output for the buyer's final result.
Make disagreement useful
Test the rubric with deliberately different answers. Give one response that is incomplete, one that is acceptable but generic, and one that is detailed and evidenced. If reviewers can't distinguish them using the anchors, the scorecard is ambiguous.
Keep pass or fail requirements separate from weighted scoring unless the tender documents expressly combine them. A mandatory certificate shouldn't become an informal quality preference. It should be checked according to the stated process.
The finished scorecard should record individual marks, consensus discussion, conflicts of interest, reasons for changes, and the evidence location. Bidwell's frameworks resource can help teams organise repeatable bid knowledge, but every criterion still needs a named internal owner who can confirm that the evidence is accurate and approved.
How Scores Are Calculated Behind the Scenes
A bid can read well and still lose points because the scoring architecture was misunderstood before drafting began. Evaluators usually record a raw mark, moderate it where the procedure requires, then convert the agreed result through the published weighting. The tender documents control each step.
Suppose a panel uses a 0 to 5 quality scale and agrees a score of 4 for a criterion worth 20%. The normalised result is 80%, giving that criterion a contribution of 16 points out of 100 before the remaining criteria and price are included.
The arithmetic is simple. The assumptions require control.
From individual marks to contribution points
A panel might record implementation scores of 3, 4 and 4. If discussion produces a consensus score of 4, the criterion contributes 16 points at a published weighting of 20%.
| Measure | Raw values | Calculation | Weighted result |
|---|---|---|---|
| Implementation evaluator marks | 3, 4, 4 | Panel discussion produces consensus of 4 | 4 out of 5 |
| Normalised implementation score | 4 out of 5 | 4 ÷ 5 × 100 | 80% |
| Weighted implementation result | 80% | 80% × 20 | 16 points |
| Remaining criteria | Their published scores | Apply each stated weighting | Add to total |
For any other scale, normalise before applying the weighting. Divide the evaluator score by the maximum possible score, multiply by 100, then multiply by the criterion weight. Rounding can change a close ranking, so retain the unrounded value where the tender methodology requires it.
Price is a separate calculation
A common low-price method awards the lowest compliant tender the full available price marks. Other bids receive a proportional result using lowest price divided by tenderer price, multiplied by the available price marks, as described in the published Contracts Finder evaluation documentation.
Buyers may instead publish another formula, rank prices, apply a cost model, or combine price and quality through a stated MEAT methodology. Use the formula in the invitation. A familiar calculation is not a substitute for the buyer's method.
Keep raw scores, weightings, calculations, moderation notes, and permitted changes together. Consensus should record why the panel changed or confirmed a mark, not merely replace individual scores without an audit trail.
Teams reviewing their assessment process can use improve team workflows with this guide as practical context. The same discipline applies to bid strategy. Assign ownership, record decisions, and make the path from evidence to score visible before the response is written.
Pitfalls That Destroy Your Final Score
A bid can lose marks before the final edit begins. Vague criteria invite inconsistent interpretation. Overlapping sub-criteria allow the same evidence to influence several marks, while excessive sub-questions split the response and make consensus harder.
Fix the answer before polishing the prose
Polished wording cannot rescue a missing resource plan. A strong policy does not show that the delivery team can implement it. Evaluators award marks for evidence tied to the question, not for capability that appears elsewhere in the organisation.
Build a compliance map connecting:
- Claim: What the response promises.
- Evidence owner: Who verifies the claim.
- Source location: The relevant case study, credential, process, CV, or approved internal record.
- Criterion link: The precise question or sub-criterion supported.
- Approval status: Whether delivery, commercial, and operational owners have approved it.
Use the buyer's terminology where it improves clarity, then support each phrase with substance. Mark cross-references clearly, so evaluators can reach supporting evidence without searching through appendices. A response that forces the panel to hunt for proof creates scoring risk, even when the proof exists.
Challenge the comfortable assumptions
Test generic assertions, unsupported percentages, guaranteed outcomes, and precise claims without a source or delivery mechanism. Confirm that the team has costed each promise, approved it, and defined how performance will be measured.
Run a pre-submission scoring rehearsal using the published anchors. Investigate every result below the acceptable mark, then check the formula as closely as the wording. Pass or fail gates, lots, variants, evaluation stages, normalisation, rounding, and tie-break provisions can change the bid strategy before any answer is written.
The question isn't whether the answer sounds strong. It's whether an evaluator can award the intended mark from the evidence on the page.
Keep internal bid targets separate from contractual commitments. A stretch ambition can guide planning, but it should not become a guaranteed outcome unless the relevant owners have approved it and can deliver it.
Using Bidwell to Score-Strategy Every Bid
Scoring strategy starts before the opportunity reaches the writing team. Bidwell's tender monitoring helps identify live UK opportunities and review the published criteria early, so bid managers can judge whether the weighting favours their actual capabilities rather than reacting after the deadline is close.
The knowledge base then gives the team a working evidence inventory. Store credentials, past responses, case studies, evaluator comments, and approved delivery material in one place. When a tender contains a heavily weighted implementation or technical criterion, the bid lead can find relevant proof before assigning the response.
AI response generation has a narrower, more useful role than producing fluent text. It can draft against the live evaluation schema, mirror the buyer's terminology, and surface evidence patterns that match the criterion. Human reviewers still need to verify accuracy, delivery commitments, pricing assumptions, and compliance.
Scan, mine, draft
- Scan the opportunity. Use monitoring to find tenders, inspect the quality and price structure, and reject work that doesn't fit your evidence or delivery capacity.
- Mine the knowledge base. Map each weighted criterion to approved case studies, credentials, people, methods, and operational proof.
- Draft and rehearse. Generate a criterion-led response, then score it internally using the buyer's scale and weighting before submission.
Bidwell's tender use cases fit this workflow by putting opportunity selection, evidence reuse, and response preparation in the same operating process. Teams comparing wider procurement tools may also consider Software for Government contracting, particularly when they need a separate view of public-sector contracting software.
The useful shift is upstream. The scoring model determines which tenders you pursue, which evidence you prioritise, and how you review every answer. Writing quality still matters, but it works hardest when the architecture has already been understood.
Bidwell monitors UK tender portals, organises your approved bid knowledge, and uses AI to generate responses aligned with published evaluation criteria. Visit Bidwell to see how your team can turn tender scoring rules into a practical scan, evidence, draft, and review workflow.
