30 Lessons From Failed Tenders

Most health and social care bids do not fail because the provider cannot deliver the service. They fail on avoidable, well-documented mistakes: a missed threshold, an unsupported claim, a recycled answer, a formatting breach. The evaluator feedback we review names the same errors again and again, which is good news: a predictable failure is a preventable one.

At a glance

• Most losses are self-inflicted and appear in written feedback in almost identical wording from one bid to the next.
• A single sub-threshold score, or one failed pass/fail gate, can end a bid before its quality is ever compared.
• Evaluators reward evidence, specific outcomes, data and worked examples, and mark down bare assertion.
• Local accuracy is a differentiator: recycled, outdated or invented local detail is routinely scored 0.
• Statutory precision (correct section 42 roles, the Mental Capacity Act, current legislation, the right regulator) signals a provider who knows the work.
• Mechanics protect earned marks: formatting breaches and internal contradictions lose points you had already won.
• Written feedback is intelligence, and providers who log it and act on it progress at the next round.

Read the scoring scheme before you write a word

1. A single sub-threshold score can sink the entire bid. Many quality-weighted frameworks set a per-question minimum, often 2 out of 5, sometimes 3 out of 5. Score below it on one question and the whole submission can be excluded before it is compared with rivals. We have seen a bid scored 1 on one staffing question rejected in full.

2. Weight your effort to the marks on offer. Read the weightings before drafting, then budget your time to them. Providers pour hours into a low-weighted question they find comfortable and starve a high-weighted gate they find hard. A 5% question and a 20% question do not deserve the same word count.

3. Percentage thresholds often apply to each category, not just the total. Some frameworks require a minimum such as 60% in every category applied for, not only across the bid as a whole. A strong overall score will not save you if one category falls short; we have seen a category scored 37.47% fail an otherwise competitive submission.

4. Reverse-engineer every answer from the scoring descriptors. The published descriptors tell you exactly what a top-band answer contains. Read the wording for a 5 out of 5 and the wording for a 2, then build your response to hit the higher band point by point. An answer written to impress in general drifts to the middle of the range.

5. Answer every printed sub-bullet, in the order it appears. Questions in care tenders list the points the panel must see covered, and evaluators score against that list, often ticking it off literally. A missing sub-bullet is a missing mark. Mirror the structure of the question and make each point easy to find.

Clear the compliance gates first

6. Treat pass/fail questions as the foundations of the bid. Some requirements are mandatory: fail one and the bid is, in the panel's words, not evaluated against the award criteria at all. Identify every pass/fail question at the outset, confirm you meet each one, and never assume quality elsewhere will rescue a gate you have missed.

7. A case study for the wrong client group scores zero. A provider once won 2 lots of a children's homecare framework but lost a third, because both mandatory case studies scored 0: one described an adult and adult services, evidencing none of the children and young people requirements. Match every case study to the exact service and cohort in the spec.

8. Answer every clarification by the deadline, and label the lot it covers. A late or ambiguous clarification response can score a whole section 0. In one bid a Social Value answer was scored nil because the panel could not tell which lot it applied to. Respond within the deadline, and say plainly which lot or question each response addresses.

9. Read location and proximity requirements literally. Where a question scores proximity, the registered address on record is what counts. We have seen a CQC registered office judged to sit outside the contract area scored 0, whatever the local capacity behind it. If the spec rewards proximity, make sure the verifiable address falls inside the area.

10. Meet contractual obligations about premises head-on. If the contract requires an office within the contracting borough, a bid that commits to an office in a neighbouring borough has not met the requirement. Read the draft terms and conditions, list every binding obligation, and confirm your response satisfies each one.

Evidence, not assertion

11. Replace every claim with a worked example. The most common criticism in the feedback we review is assertion without evidence. One panel scored an answer 25 out of 100 and recorded "no confidence", because it was built from generic statements with no context or examples. For each claim, show the instance: what happened, what you did, what changed.

12. Lead with outcomes and hard data. Panels want the result, not the aspiration. "We promote independence" earns little; a named measure, a baseline, a figure after a stated period and the number of people it applied to earns marks. Quantify wherever you honestly can, and connect each intervention to the outcome it produced.

13. Prove continuous improvement with a documented before-and-after. "We are committed to continuous improvement" is not evidence of it. Feedback repeatedly notes "little or no evidence of continuous improvement", and no example of how an improvement helped a person you support. Show one change you made, what triggered it, and the difference it made to someone's life.

14. Choose one deep example over many shallow ones, and close the loop. A single relevant example, carried through to its outcome with the lessons you took from it, scores better than a list of thin illustrations. Finish the example: outcome, review, and what you changed to put it right.

15. Evidence your workforce model in numbers, not adjectives. "Suitably qualified and experienced staff" tells a panel nothing. Weak answers lose marks for no management structure chart, thin detail on training, supervision and staff welfare, and staffing so vague it is unclear how demand would be met. Give the structure chart, ratios, supervision frequency and training matrix.

Local knowledge and tailoring

16. Never recycle an answer from another area. Evaluators catch reused content instantly. In one mobilisation response the contracting authority was never named, while the data, recruitment and staffing detail all referred to 2 other boroughs; it scored 0. If an answer would read identically for any commissioner, it is not yet an answer to this one.

17. Get the local safeguarding pathway exactly right. Naming a MASH where the council does not operate one, using a superseded referral form instead of the current portal, or citing organisations irrelevant to the area all signal that you have not researched the patch, and they damage your safeguarding score. Confirm the current route, board and portal first.

18. Do not claim a local presence you do not have. Case studies that imply you already work in a district when you do not are a credibility problem, and a panel that knows its own patch will flag it. Be precise about where you operate now and where you would mobilise.

19. Mirror the commissioner's own terminology. Read how the authority names its services, teams, documents and pathways, and use those terms exactly. If the specification calls the service "Home First", answer about "Home First", not a generic label of your own. Matching their language lets the evaluator map your response to the specification without effort.

20. Make local knowledge a deliberate differentiator. Where local knowledge is strong it is often the winning margin: specific transport constraints, rural coverage, community assets, relationships with named local teams. Where it is absent the loss is doubled, because a rival who has it pulls ahead on the same question.

Get the statutory and service-model detail right

21. Describe safeguarding as a complete escalation route. Answers lose marks for stopping at "we would report it". Show the full route: immediate safety, any medical intervention needed, escalation to the police where a crime may have occurred, and the correct local referral pathway. Feedback often cites missing escalation, no medical response and incorrectly named safeguarding boards.

22. State the Care Act section 42 roles correctly. Under section 42 of the Care Act 2014, the local authority makes, or causes to be made, the safeguarding enquiry. Your role is to secure the person's immediate safety, record, refer without delay, share information and support the enquiry. Answers that claim the provider "leads a section 42 enquiry" misstate the law.

23. Cite current law and name the right regulator. Stale templates are obvious to a panel. The Data Protection Act 2018 replaced the 1998 Act; the Mental Capacity Act 2005 has 5 statutory principles and a 2-stage capacity test, not a "4-part test". Name the right regulator, CQC for regulated adult social care in England, Ofsted for children's social care, CIW in Wales.

24. Match your model to the exact service commissioned. Proposing six-weekly reviews and reablement content in a Home to Assess service shows you have not grasped the scope. A Home to Assess service, a supported accommodation service under the Supported Accommodation (England) Regulations 2023, and a children's home are different things. Evidence what the spec expects: health action plans, hospital passports, best-interest and DoLS processes, move-on planning.

25. Write in plain English, accessible to the client group. If the specification requires information accessible to people with a learning disability, an answer written in dense professional prose fails its own test. Show that you can produce easy-read and accessible formats, and write the response itself clearly. A panel reads accessibility as a proxy for real communication.

Submission mechanics, and turning feedback into wins

26. Obey the formatting rules to the letter. The most avoidable loss we see is formatting non-compliance. In one bid, responses were not double-line-spaced, ran over the page limit, and everything past the limit was excluded from evaluation; the missing content dragged score after score below the threshold. Check spacing, font, margins and page limits before you submit.

27. Use the full allowance, and never over-run it. Under-using the word or page limit leaves marks on the table; over-running it loses the content that spills past the cap. Both are penalised. Fill the space with evidence, not padding; if you are over, cut the weakest material yourself rather than letting the limit cut your strongest.

28. Eliminate internal inconsistencies before submission. Contradictions destroy confidence. One timeline suggested 500 hours in one place and an aim of 1,000 hours in another, creating confusion the panel scored against. Numbers, dates, staff figures and commitments must agree across every answer. Read the whole bid as one document and reconcile every figure.

29. Cost and localise your Social Value, and track it. Social Value that is high-level, uncosted, generic and unmeasured scores poorly. Panels want commitments specific to the area, costed or resourced, and tracked with metrics and timescales you will report against. Tie each commitment to the local community the contract serves, and say how you will evidence delivery.

30. Stress-test your risk and continuity, do not gloss them. Business continuity plans that never test their actual risk areas, and risk or contingency sections that are "not explored", read as box-ticking. Name your real risks (staff sickness, referral surges, a failure in the supply chain), and show the specific, resourced response to each.

FAQ

Why do social care tenders fail? Most fail on avoidable, evidenced mistakes rather than an inability to deliver: a score below the per-question minimum, a failed mandatory gate, assertion instead of evidence, recycled or inaccurate local detail, statutory errors, or formatting breaches. These patterns recur across the feedback we review, so they can be designed out before submission.

Can one low score fail an entire bid? Yes. On many quality-weighted frameworks each question carries a minimum, often 2 out of 5, and dropping below it on a single question can exclude the whole submission outright. Pass/fail questions work the same way: fail one and the bid may not be evaluated against the award criteria at all.

Does formatting really affect tender scores? Directly. If your response is not formatted as instructed, for example not double-line-spaced, it can exceed the page limit, and any content past the limit is usually excluded from evaluation. Excluded content means missing evidence, and missing evidence pulls scores below the threshold.

How much evidence should a tender answer include? Enough to prove, not just assert, every claim. Favour one deep, relevant example, carried through to its outcome and the lessons learnt, over a list of shallow ones. Use real figures a commissioner could audit, and make sure each example matches the exact service and client group in the specification.

Should you request written feedback after losing a tender? Always. Written feedback names the specific reasons you lost marks, which is the most precise improvement plan you will ever receive. Log it, address each point directly, and carry the corrections into the next bid.

The pattern runs both ways. A provider that lost a supported living DPS round, asked for written feedback, corrected the specific gaps it named and re-applied later reached the presentation stage on the next attempt. The same feedback that explains a loss is the blueprint for the next win, and the 30 lessons above are simply that feedback, gathered before you need it.