How the tender answer checker works
Everything runs on this page in your browser. Words are counted by splitting on spaces, the same way Microsoft Word does. Characters are counted with and without spaces, because portals differ. Sentences end at a full stop, question mark, exclamation mark or line break, and paragraphs are separated by a blank line.
Reading ease uses the Flesch formula: 206.835 minus 1.015 times the average sentence length, minus 84.6 times the average syllables per word. Higher is easier. Reading time assumes about 238 words a minute, a common estimate for adult silent reading. The weak-phrase check matches patterns that evaluators routinely mark down: hedging, unprovable claims, vague quantifiers and likely passive voice. It flags; you decide.
Why word limits matter in UK tenders
Since 24 February 2025, most new UK public procurements are run under the Procurement Act 2023, which requires contracting authorities to award to the most advantageous tender against the award criteria they publish. Word and character limits are part of how buyers keep that assessment fair: every supplier gets the same space to make their case. See the Cabinet Office's Procurement Act 2023 guidance on GOV.UK for how assessment and award work.
Limits are set in the tender documents, not the Act, so they vary. Read the instructions to tenderers for three things: the limit per question, whether it is words or characters, and what happens if you exceed it. Many buyers say evaluators stop reading at the limit. That turns your final paragraph, often the summary of your best evidence, into zero marks.
Worked example: fixing a 1 out of 5 answer
Press “Try a sample” above. The first two paragraphs are typical: “we believe”, “passionate”, “world-class”, “regularly”, “where possible”. Nothing there can be scored. The last paragraph has no flags, because every clause is a commitment with a number. Here is the pattern:
Before:We believe our team is passionate about quality.
After:Our 14 supervisors audit every site monthly; average score 96% in 2025.
Before:We will endeavour to respond where possible.
After:We will respond to all reactive requests within 2 hours, 24/7.
Before:Issues are dealt with quickly.
After:The contract manager fixes any failed audit item within 4 working hours.
Notice the rewrites are barely longer. Specific writing is dense writing, which is why fixing weak phrases usually helps you hit the word limit rather than blow it. For more side-by-side rewrites, read our bid writing examples and how to answer tender questions.
Five rules for answers that fit and score
1. Plan to 90% of the limit. Split the limit across the parts of the question in proportion to their weight, then write. 2. Lead with the answer. Your first sentence should state your headline commitment. 3. Keep sentences under 25 words. The checker warns when your average creeps above that. 4. Replace adjectives with evidence. A KPI, an audit score or a named standard. 5. Mirror the question. Use the buyer's headings in the buyer's order so evaluators can find each part.
New to public-sector bids? Start with how to write a tender response, then how to write a winning bid. Ready to find something to bid on? Browse live UK contracts.
Questions, answered
What happens if my tender answer goes over the word limit?
It depends on the tender instructions, so read them. Many buyers state that evaluators will stop reading at the limit and ignore anything after it, which means your strongest closing points score nothing. Some treat an over-length answer as non-compliant. Either way, going over never earns marks. Aim to finish within the limit with a little room to spare.
Do headings, tables and bullet points count towards the word limit?
Usually yes. Most buyers count every word in the response box, including headings, bullets and table text. Some exclude diagrams or attached appendices, others ban attachments entirely. The instructions to tenderers will say. If they are silent, ask a clarification question through the portal before the deadline, and assume everything counts until you get an answer.
Does this tool count words the same way as the tender portal?
Close, but portals differ. This counter splits on spaces, like Microsoft Word, so "24/7" and "ISO-9001" count as one word each. Some portals count characters including spaces, and some count line breaks as characters. When a limit is in characters, switch the limit to characters here and leave a margin of around 2% for portal differences.
Should I use every word of the limit?
Not for the sake of it. Evaluators score evidence against the question, not length. But if you finish well under the limit, you have usually left out detail that would earn marks: a number, a named role, a frequency, or a result from a similar contract. Aim for 90 to 100% of the limit, with every sentence giving the evaluator something to score.
What is a good Flesch reading ease score for a tender answer?
Aim for 50 or above, and 60 or above where you can. Scores of 60 to 70 are generally treated as plain English. Technical answers on method or specification will sit lower, which is fine, but a score below 30 means long sentences and heavy jargon. The quickest fix is to split any sentence over 25 words.
Why does the checker flag phrases like "we believe" and "world-class"?
Because evaluators can't score them. "We believe" signals doubt. "We will endeavour" and "where possible" are promises nobody can check. "World-class" and "best in class" are claims with no proof. Replace each with a commitment the buyer could hold you to: a time, a target, a named standard, or a figure from a comparable contract.
Is my text safe in this tool?
Yes. All counting and checking runs in your browser using JavaScript on this page. Your answer is never sent to our servers or stored anywhere, so it is safe to paste commercially sensitive drafts. Close the tab and it is gone.
Comparing options? See the AI tender software roundup or check pricing.