To improve tender writing, find where you lose marks using assessment summaries, then plan every answer against the scoring criteria, replace claims with evidence, write for a tired evaluator and review independently. Small, repeated fixes to process raise scores more than better prose.
Why now
How to improve tender writing: diagnose first¶
Collect your last five assessment summaries and list scores by question type: method, quality, social value, mobilisation, staffing. Patterns show quickly. If method answers score low everywhere, the problem is planning or evidence, not wording.
Tender writing tips: planning¶
- Read the scoring descriptors before the question.
- Turn sub-questions into headings in the same order.
- Allocate words by weighting.
- List your evidence before you draft.
- Pick win themes and place them deliberately.
Tender response tips: writing¶
- Answer in the first sentence.
- Use the buyer's terms from the specification.
- Replace adjectives with numbers and names.
- Explain how, who, when and how it's measured.
- Write about the buyer's outcomes.
- Keep paragraphs short; use bullets for lists.
How will you recruit and retain staff for this contract?
We will recruit 18 care workers through local job centres and our existing bank staff, starting eight weeks before go-live. All staff will complete the Care Certificate within 12 weeks. Retention will be supported by guaranteed-hours contracts, paid travel time and monthly supervision, with turnover reported to you quarterly.
Why it scores
- Specific numbers and timings
- Named standard
- Reporting to the buyer
Illustrative example. Use your own verified figures.
Tender writing tips: evidence and review¶
- Keep an evidence library of case studies, KPIs, policies and certificates.
- Write case studies as challenge, action, measured result.
- Have someone who didn't write the answer score it.
- Run a separate compliance check on limits, formats and attachments.
- Search for other buyers' names before submitting.
How to improve tender writing in six steps
Step 1
Review past scores
Collect assessment summaries and find weak question types.
Step 2
Build an evidence library
Gather proof in one place and keep it current.
Step 3
Plan to the criteria
Storyboard each answer against the scoring descriptors.
Step 4
Write with proof
Lead with the answer, then how and evidence.
Step 5
Review as an evaluator
Independent scoring before submission.
Step 6
Learn from results
Update answers and library after every bid.
Why tenders lose marks: the six most common causes¶
Before changing how you write, it helps to know what evaluators actually mark down. Across sectors, the same handful of faults appear in low-scoring answers again and again. Most of them are process faults that better wording cannot fix.
- Missed sub-questions. A question with four bullet points is four questions. If you answer three, the evaluator usually cannot award the top band, however good the rest is.
- Generic content. Answers lifted from a previous bid often mention the wrong service, the wrong location or, worse, the wrong buyer. Evaluators notice copy-and-paste quickly.
- Claims without proof. Phrases like "highly experienced team" or "robust processes" carry no marks. Evaluators are told to score what is evidenced, not what is asserted.
- Describing the company instead of the contract. Two paragraphs of history and values use words that should explain how you will deliver this specific requirement.
- Ignoring the specification. If the specification sets response times, reporting frequency or staffing ratios, a strong answer confirms each one and explains how you will meet it.
- Poor structure. A single block of text makes the evaluator hunt for points. If they cannot find something quickly, they may not credit it.
Map your own assessment summaries against this list. Most suppliers find that two or three causes explain most of their lost marks. Fix those first.
Read the scoring criteria like an evaluator¶
Every UK public-sector tender should tell you how answers are scored. Look for the evaluation methodology, usually in the Invitation to Tender or the tender notice documents. Under the Procurement Act 2023, buyers must set out the award criteria and their relative importance in the tender notice or associated tender documents, so you should never have to guess.
A typical quality scale runs from 0 to 5 or 0 to 10, with a descriptor for each band. The descriptors are the most useful text in the whole pack. A top-band descriptor usually asks for an answer that is fully compliant, gives detailed and relevant evidence, and gives the evaluator full confidence in delivery. A middle band usually describes an answer that meets the requirement but lacks detail or evidence in places.
| Score band | What the descriptor usually asks for | What to put in your answer |
|---|---|---|
| 5: Excellent | Fully addresses every part, detailed, evidenced, high confidence | Every sub-question answered in order, named people, timings, KPIs and a relevant case study |
| 4: Good | Addresses every part with good detail and some evidence | Every sub-question answered, most points supported with specifics |
| 3: Satisfactory | Meets the requirement, limited detail or evidence | Answer is compliant but relies on general statements |
| 2: Weak | Partly addresses the requirement, major gaps | One or more sub-questions missed or answered vaguely |
| 0-1: Poor or no answer | Fails to address the requirement | Off-topic, missing or unacceptable answer |
Illustrative 0-5 scale. Always use the exact descriptors in your tender pack.
Copy the top-band descriptor into your planning document above each question. When you review the draft, test it against those words. If the descriptor says "detailed" and your answer has no dates, names or numbers, it will not reach the top band.
Plan answers with a storyboard¶
Storyboarding means planning the answer before writing it. It is the single habit most strongly linked to better scores, because it forces you to find the evidence first. A simple storyboard for each question fits on one page.
- The question, word or page limit, and weighting.
- The top-band descriptor copied from the scoring guidance.
- Each sub-question as a heading, in the order the buyer asked.
- Under each heading: the key point, the proof (case study, KPI, certificate, named person) and the benefit to the buyer.
- The win theme or themes this answer should back up and repeat.
- Any specification clauses the answer must confirm.
Allocate words by weighting. If a question is worth 20% of the quality score and another is worth 5%, the first deserves far more effort and usually more of its word limit. Within a question, spread words across sub-questions roughly equally unless the buyer signals one matters more.
Write for a tired evaluator¶
Evaluators often score dozens of responses to the same question, sometimes alongside their day job. They mark against a sheet and look for evidence that each point has been met. Make that easy.
- Lead each section with a direct answer: "We will mobilise in six weeks using a dedicated mobilisation manager."
- Mirror the buyer's headings and vocabulary so the evaluator can tick points off.
- Use short paragraphs of two to four sentences, and bullets for lists of actions or features.
- Put numbers in digits and keep them consistent with your pricing and other answers.
- Name the roles responsible and say how often things happen: weekly, monthly, quarterly.
- End each section with the outcome for the buyer, not a summary of what you just said.
- Avoid jargon the buyer did not use, and spell out acronyms on first use.
Before and after: one sentence
Weak: "We have a strong commitment to quality and continuously improve our services." Strong: "Our contract manager will audit 10% of jobs each month against the specification, report results to you in the monthly meeting and agree corrective actions within five working days." The second sentence is longer but every word can be scored. It says who, what, how often and what the buyer gets.
Build evidence that scores¶
Evidence is what separates a 3 from a 5. The best evidence is recent, relevant to the buyer's requirement and measurable. Keep it in a library so writers are not searching old emails the week before a deadline.
| Evidence type | Good example | Where to keep it |
|---|---|---|
| Case study | Similar contract, challenge, action, measured result, buyer name if permitted | Evidence library, one page each |
| KPI data | 12 months of performance against contract KPIs | Monthly export from your operations system |
| Accreditations | ISO 9001, ISO 14001, ISO 45001, Cyber Essentials, sector schemes | Compliance folder with expiry dates |
| Policies | Health and safety, modern slavery, environmental, equality, safeguarding | Policy folder, reviewed annually |
| People | CVs and role profiles for key staff | HR folder, updated when staff change |
| Social value | Local hires, apprenticeships, SME spend, carbon reduction with figures | Social value log updated quarterly |
Only use figures you can prove. Evaluators can ask for clarification and, if you win, buyers may hold you to commitments in the contract. Under the Procurement Act 2023, contracts over £5 million must have at least three KPIs published and performance against them reported, so commitments made in bids are more visible than before.
Review like the buyer will¶
Most suppliers review too late and too lightly. A proper review is a scoring exercise, not a proofread. Use at least two stages.
- 1
Storyboard review
A colleague checks the plan for each high-weighted answer before drafting begins. Fixing a plan takes minutes; fixing a finished answer takes hours.
- 2
Draft scoring
Someone who did not write the answer scores it against the buyer's descriptors and writes down why it did not reach the top band.
- 3
Revise
The writer addresses every comment, then checks word limits, formatting and cross-references.
- 4
Compliance check
A final pass against the compliance matrix: every question answered, every attachment included, every declaration signed.
Give reviewers the scoring guidance and the specification, not just the draft. A reviewer who only reads the answer will judge style. A reviewer who reads the criteria will judge whether it scores.
Sector variations¶
The principles are the same everywhere, but evaluators look for different evidence in different sectors. Adapt your library and your examples to suit each sector.
- Construction: method statements, programme, CDM 2015 duties, site safety record, Constructionline or CHAS accreditation where the buyer asks for it.
- Health and social care: CQC ratings, safeguarding procedures, Care Certificate completion, staff turnover and continuity of care.
- IT and digital: GOV.UK Service Standard, Cyber Essentials or Cyber Essentials Plus, data protection under UK GDPR, accessibility.
- Facilities management: mobilisation plans, TUPE handling, staffing levels, audit regimes, sector certifications.
- Professional services: named team CVs, methodology, quality assurance and conflict of interest management.
Common tender writing mistakes to stop today¶
- Starting to write on the day the tender is published without reading the whole pack first.
- Leaving the highest-weighted question until last.
- Answering the question you wish had been asked.
- Exceeding word or page limits, which can mean text is ignored or the bid is rejected.
- Using a price or number in one answer that contradicts another answer or the pricing schedule.
- Leaving another buyer's name in reused text.
- Submitting at the last minute and discovering a portal upload problem.
Using AI to improve tender writing¶
AI drafting tools can help most with the parts that slow teams down: extracting questions and requirements, producing a structured first draft from your evidence, and checking a draft against the criteria. They do not replace judgement about strategy, pricing or which evidence is strongest.
PPN 017 allows UK public buyers to ask suppliers how they have used AI in preparing tenders. Keep a record of how you use AI, check every fact it produces against your own evidence, and never let it invent case studies, figures or accreditations. Used this way, AI shortens the drafting stage and leaves more time for planning and review, which is where most marks are won.
Worked example: turning a 2 into a 4¶
Take a common question: "Describe how you will manage complaints and ensure continuous improvement (500 words, 10%)." A supplier scored 2 out of 5 with an answer that said complaints were "taken seriously", "logged on our system" and "resolved promptly". The assessment summary said the answer lacked detail on timescales, escalation and how learning would be shared with the buyer.
The rewrite followed four steps. First, the writer split the question into two headings: complaints handling and continuous improvement. Second, they listed what a top-band answer needed: timescales, roles, escalation, reporting and examples of change. Third, they pulled evidence from the library: last year's complaint volumes, average resolution time and one example where a complaint led to a process change. Fourth, they drafted to the plan.
The new answer committed to acknowledging complaints within one working day and resolving them within ten, named the contract manager as owner and the operations director as the escalation point, promised a monthly complaints report with root causes, and described a real change made after a complaint on another contract. Nothing was exaggerated. The difference was specificity and structure, and it moved the answer into the good band.
Use feedback to improve every bid¶
Under the Procurement Act 2023, contracting authorities must give each bidder an assessment summary when they notify the award decision, before the standstill period. It explains how your tender was assessed against the award criteria and, where relevant, how it compared with the winning tender. This is the richest free source of improvement data you have.
- Log every score by question type in a spreadsheet so you can spot patterns over time.
- Note the evaluator's reasons word for word. Phrases like "lacked detail" or "not evidenced" point to specific fixes.
- Update the relevant library answer the same week, while the feedback is fresh.
- Share lessons with everyone who writes or reviews bids, not just the bid manager.
- If something is unclear, ask the buyer for a short debrief call. Many will agree.
Tools that help¶
Compare bid compliance software, bid library software and tender writing services.
Related guides: how to improve bid writing, tender writing guide, how to write a winning tender, tender writing examples.
Sources
FAQs
Frequently asked questions
- Structure every answer around the scoring descriptors and replace claims with evidence. These two changes usually lift scores fastest.
- Read the criteria first, answer in the first sentence, use the buyer's language, prove every claim and review independently.
- Common causes are generic answers, missed sub-questions, no evidence and poor fit with the requirement.
- Compare your scores with the winner's in the assessment summary and fix the weakest question types first.
- For strategic or complex bids it can help. For routine tenders, a good process plus AI drafting is often enough.
- Plan by weighting, cut adjectives and history, and use bullets for lists.
- You can see improvement on the next bid by storyboarding high-weighted questions and adding evidence. Lasting improvement usually takes several bid cycles, as you build an evidence library and learn from assessment summaries.
- A one-page plan for each answer listing the sub-questions as headings, the key points, the evidence for each, the win themes and the word allocation. It is agreed before drafting begins.