30 Lessons From Won Tenders

Winning tender responses are rarely the most eloquent ones. They are the most evidenced and the most deliberately engineered to the scoring scheme, and the evaluator feedback we read on won bids keeps rewarding the same handful of moves. Master those moves and you turn good practice you already deliver into marks that place you on the framework.

At a glance

• Evaluators score against named criteria, not overall impression. Map your answer to every sub-point and write to the top band.
• A specific, anonymised case example is the single biggest score driver. Generic assurances score as "acceptable" at best.
• Quantify outcomes. "Improved" is not evidence; 100% retention or a measured independence score is.
• Tailor to the named authority and area. Local pathways, partners and recruitment are where winning bids separate from the field.
• Get the statutory detail exactly right. Outdated legislation and the wrong regulator make evaluators doubt everything else.
• Compliance is a gate, not a formality. A missed clarification or an incomplete submission can drop you a tier on its own.

Engineer to the scoring scheme

1. Open every answer with a direct thesis sentence. State your conclusion in the first line, then prove it. Evaluators skim for whether you have answered the question before they read for quality, and a clear opening claim tells them the marks are there to be found.

2. Answer every named sub-point, in the question's order. Evaluators map marks to specific criteria, so a top answer addresses the question criteria in full and is often praised point by point, for example against "point 3 of the criteria". Signpost each sub-point using the commissioner's own numbering, so nothing is missed or has to be hunted for.

3. Write to the top band, not the pass line. Read the scoring descriptors and learn what separates a 3 that meets the minimum from a 5 that exceeds requirements and adds significant benefit. Top marks go to answers that show clearly and convincingly how every requirement will be delivered, backed by detailed plans and methodologies, not intent.

4. Use the full word or page allowance. A recurring note on strong-but-improvable answers is "only just over 2 pages used" where 5 were available. Unused space is unclaimed marks. If you have room, add another worked example, a metric or a contingency; the evaluator can only score what is actually on the page.

5. Mirror the commissioner's own terminology. Where the specification uses a particular phrase for an outcome, a pathway or a cohort, use that exact phrase rather than your house synonym. It helps the evaluator match your answer to their criteria at speed, and it signals close reading rather than a recycled template.

Evidence and outcomes

6. Anchor every claim to a specific, anonymised case example. A named individual, usually written as "P", with a real situation, action and result is the single biggest score driver we see. Generic statements without examples are marked as descriptive and score poorly; one relevant, detailed example routinely lifts an answer from "acceptable" to "good".

7. Quantify outcomes and retire the word "improved". Replace narrative claims with figures: a 100% staff retention rate, a fall in incidents, an independence score measured against a baseline, quarterly progress reports. Feedback penalises evidence that is "mostly descriptive and not quantified". A number the evaluator can check beats a paragraph of adjectives.

8. Go deep on one example rather than thin across many. A single case followed through from referral to outcome, including what you changed along the way, evidences competence far better than a list of shallow mentions. Depth shows assessment, delivery, review and result in one thread, the exact arc evaluators score.

9. Evidence reflective practice and lessons learnt. Show what did not work first time and what you changed. Winning answers make the learning explicit, with lessons learned evidenced and incorporated; weaker ones leave it "unclear which strategies have worked well and which have not". A named improvement cycle turns a good service into a maturing one.

10. Set a baseline and measure the distance travelled. Capture where the person started, then track change against it with a repeatable measure: an independence rating, a mood tracker, a loneliness early-warning indicator reviewed on a set cycle. Distance travelled persuades more than a snapshot, and it gives you the quarterly data evaluators reward.

Local knowledge and tailoring

11. Write to the named authority, not a generic council. The clearest winning differentiator we see is genuine local understanding: the area's pathways, demography and pressures named specifically. Where it is missing, feedback reads "further focus was required regarding the area". Replace every "the Council" placeholder with something only a local provider would know.

12. Evidence a recruitment and retention plan for that labour market. Councils know their own workforce shortages. Show where you will recruit locally, your pay position, and how you will retain staff in that specific area, rather than a national statement. A credible local supply of carers separates a deliverable bid from an aspirational one.

13. Name the local community assets and multi-agency partners you will use. Reference the actual voluntary sector organisations, day services and networks in the area that support community access and reduce isolation. Concrete local connections evidence that your model works on the ground here, and they strengthen your social value and outcomes answers too.

14. Show you understand the specific local challenges and how you mitigate them. Rural travel times, a tight care market, transport gaps or particular demand pressures should each appear by name, paired with your mitigation. Evaluators reward research into local challenges and penalise its absence, because it reassures them you can actually deliver here.

Getting the statutory detail right

15. Cite current legislation and check every reference is in force. A single outdated citation, such as the Data Protection Act 1998 in place of 2018, makes evaluators question how often all your policies are reviewed. Sweep every answer for superseded Acts, old statutory guidance and lapsed frameworks before submission. Accuracy here protects the credibility of everything else you claim.

16. Get the Section 42 safeguarding split right. Under the Care Act 2014, the local authority makes, or causes to be made, a Section 42 enquiry. Your role is to secure immediate safety, record, refer, share information and support that enquiry. Name your Designated Safeguarding Lead, the correct local referral route and a clear escalation path. Precision here signals a service that has done this for real.

17. State the Mental Capacity Act 2005 correctly. It rests on 5 statutory principles and a 2-stage test of capacity. State both precisely, because muddling the number of principles or stages is a common and visible error. Pair it with Making Safeguarding Personal to show that you start from the person's own outcomes. Getting the framework exactly right separates practitioners from template writers.

18. Use the correct regulator for the setting, and never mix them. CQC regulates adult social care in England; Ofsted regulates children's social care and supported accommodation for looked-after 16 to 17 year olds under the Supported Accommodation (England) Regulations 2023; CIW regulates in Wales. Citing the wrong regulator tells the evaluator the content was lifted from another sector.

19. Write child-centred content in children's tenders. Adult social care wording pasted into a children's response is penalised repeatedly. Children's frameworks need child-focused language, the correct statutory context, and evidence such as accessible one-page profiles written in plain language with pictures and a quote from the child. Match the content to the cohort, not to your nearest existing template.

Models, workforce and systems

20. Name recognised support models and show how you apply them. Positive Behaviour Support with both proactive and reactive strategies, low-arousal approaches, trauma-informed and strengths-based practice, and RRN-trained staff all score well when tied to a real example. Name the model, then show it working for one person. A named model with no application reads as a buzzword.

21. Co-produce plans and evidence genuine service-user voice. Show the person, or the child, helping to write their own plan, and co-production carried up to board or committee level. Evaluators look for real involvement, not consultation after the fact. A one-page profile in the person's own words is stronger evidence than a policy statement about co-production.

22. Detail the whole workforce model, not just training. Show values-based recruitment, a training matrix, specialist and refresher training, regular supervision, appraisals, wellbeing support and career progression as one connected cycle. This is where deliverability is judged, and an answer that only lists mandatory courses leaves most of the available marks on the table.

23. Show sufficiency, ratios and a management structure. Include a clear staffing establishment, realistic ratios for the cohort, named roles including the Registered Manager, and a management structure chart. Evaluators are testing whether the model is safely staffed and led, and a chart of who covers whom answers what narrative alone cannot.

24. Evidence digital systems with proof, not headlines. Electronic call monitoring, a digital social care record, electronic medicines administration records and role-specific dashboards should be evidenced with screenshots, data or named products, not asserted. Some frameworks score technology as a hard evidence item and award 0 where it is unproven. Show the system in use and the real-time data it gives managers.

25. Set defensible crisis, risk and continuity timescales. Give clear thresholds and realistic timings: de-escalation steps, mood trackers, crisis coping cards, and risk management plans produced within a stated number of days. Timings must be defensible; one bid's 72-hour case strategy meeting was judged "a long timescale for a critical incident". Stress-test business continuity against its own named risks.

Social value, price and compliance

26. Cost, localise and track your social value. Commit to specifics such as care-leaver and local employment, apprenticeships with a credible supply-chain plan, and environmental measures you will monitor and report. Attach metrics and a reporting cycle. Vague pledges score poorly; a costed, locally focused commitment with named measures reads as something you will actually deliver and be held to.

27. Price to the weighting with a clear audit trail. Understand how price is scored, then build transparent, itemised pricing linked to the Real Living Wage with a moderate, proportionate surplus. An auditable model that shows how every pound is spent wins price marks and survives scrutiny. Undercutting on an unexplained rate invites questions about safe staffing.

28. Do not bank on a strong price rescuing a weak quality answer. On some heavily price-weighted or pass-threshold frameworks, a good price plus a bare "meets minimum" quality score is enough to win a place. Treat that as a framework-design quirk to understand, not a strategy. Write every quality answer as if it must clear a threshold, because on most frameworks it must.

29. Treat your CQC rating as a hard gate where it is one. On regulated adult social care frameworks in England, a "Good" or "Outstanding" rating and a Registered Manager in post can decide your tier, and a weaker rating can drop you down one. Know the gate before you invest in the bid, and evidence your current registration status clearly.

30. Complete the whole submission and hit every deadline. One bidder was demoted from Tier 1 to Tier 2 purely for not submitting a video question and missing a clarification deadline. Answer clarifications on time, supply insurance and policy evidence promptly, and submit a realistic, tailored mobilisation plan with a Gantt timeline and no internal discrepancies. Assurance of deliverability is itself scored.

FAQ

What makes a tender response score well? Evaluators score against named criteria, not overall impression, so the highest-scoring answers map directly to every sub-point of the question and write to the top band. They lead with a clear claim, prove it with a specific, quantified case example, and use the full word count. In our experience, evidence and structure beat eloquence almost every time.

How important are case studies in care tenders? They are usually decisive. A specific, anonymised case example is the single biggest score driver we see, because it shows the evaluator your model working for a real person rather than in theory. One deep example, followed from referral to a measured outcome, lifts an answer further than a page of general assurances.

Can a strong price win a tender with a weak quality score? Occasionally. On frameworks that are heavily price-weighted, or that only require quality to pass a threshold, a good price with a minimum quality score can secure a place. This is a quirk of how that framework is designed, not a reliable strategy. Treat every quality answer as if it must clear a threshold, because on most contracts it does.

Why do care providers lose marks on safeguarding answers? Usually because the statutory detail is wrong or vague. Getting the Care Act 2014 Section 42 split right, stating the Mental Capacity Act 2005 correctly with its 5 principles and 2-stage test, and naming your Designated Safeguarding Lead and local referral route all signal a service that does this for real. Outdated legislation undermines the whole answer.

Do I need a Good CQC rating to win a care contract? On some regulated adult social care frameworks in England, yes. A "Good" or "Outstanding" rating and a Registered Manager in post can determine which tier you are placed in, and a weaker rating can drop you a tier or gate you out entirely. Check the qualification rules before you commit time to the bid.

The pattern across won bids is consistent: the providers who place are not writing better prose, they are proving, in the evaluator's own terms, that the service already works. Build every answer to be scored, and the marks follow.