A published Q&A review is a source-bounded examination of every question, answer, attachment and referenced correction made available to the relevant bidder population. Its work product is an impact register that preserves the official item identifier, exact source, publication context, affected scope, candidate bid objects, accountable owner and next control. It includes questions raised by other suppliers and records gaps instead of equating an alert, an export or the questions your team submitted with the full record. It identifies potential effects; it does not decide document precedence, silently amend the tender or authorize a response change.
An airport is procuring maintenance for baggage-screening equipment across three terminals. The bid team submitted one question about access badges and later opened the email announcing the buyer’s reply. It did not revisit the portal’s full clarification area. In that area, question 18 states that technician cover runs from the first screening lane opening until the final daily closure, question 41 says specified consumables belong inside the fixed service fee, and question 57 points to a corrected spare-parts workbook. All three were asked by other bidders. The team’s staffing model, price and workbook remain based on the original pack, even though nobody has deliberately accepted that position.
Treat the official Q&A area as a procurement source, not an inbox. Declare which procedure, stage, lot, bidder population, portal area, attached schedule and observation time the review covers. Enumerate every visible item before judging importance. Preserve the buyer’s words, their relationship to the question and any referenced document separately. Then classify a credible effect and assign an owner who must prove no change, open the proper downstream control or retain the issue as unresolved. A notification can tell you to look; it cannot prove that nothing else was published.
Review boundary
Complete means every official location for the relevant bidder population
A Q&A log is rarely just the screen that first opens. The portal may have an interactive forum, a downloadable schedule, notices in a message centre and corrected files in the document area. A later stage can have a separate forum visible only to invited suppliers. The review boundary must name these locations, the account and role used, the relevant stage and lots, and the time at which they were observed. A generic statement that the portal was checked cannot be reproduced.
For the airport procurement, the declared boundary includes the clarification tab for the maintenance tender, all 68 visible threads, the downloadable schedule named in the buyer message, the document list for corrected attachments and the participant inbox. It excludes a separate construction procurement with a similar title. If a team member cannot access one named location, the record says source unavailable and assigns an access action. It does not reduce the expected set to whatever that person could see.
Capture the visible total, first and last identifiers, page count, active filter, unanswered count, export name and observation time. These facts do not prove that the buyer will publish nothing later. They prove what was inspected. The next check remains tied to the tender’s final-answer milestone, a buyer message, a document change or another procedure-specific trigger.
| Boundary element | Airport example | Unresolved state |
|---|---|---|
| Procedure and stage | Baggage-screening maintenance, tender stage | Identity or stage mismatch |
| Audience | All suppliers admitted to submit a tender | Audience not confirmed |
| Interactive forum | 68 visible threads, filters cleared | Page or filter coverage missing |
| Downloaded schedule | Clarification schedule dated 2 September | File unavailable or count differs |
| Referenced files | Spare-parts workbook revision 3 | Target not found or version unclear |
| Observation | Authorized portal role and timestamp retained | Access context not reproducible |
Notifications
An alert prompts a source check; it does not define the published record
Email alerts are useful precisely because a bid team cannot stare at a portal all day. They are still secondary evidence. An alert can be batched, routed to one mailbox, suppressed by a user preference, limited to the team’s own thread or stripped of attachments. It can arrive after a corrected file is already live. Record the message identifier and receipt time, then follow its official link through the authorized route and inspect the surrounding Q&A set.
Current platform documentation makes this distinction concrete. Simap states that notifications about new answers and document changes are sent daily at 17:00. That describes a notification schedule, not a promise that the email contains every current item at every moment. Current UK guidance also favors electronic communications and documenting relevant information. Neither source allows a bidder to replace the procurement-specific portal record with its inbox.
Build an operational rule that survives absence of an alert. For example, check the official Q&A area at the start of each working session during the final response week, after any buyer message, at the published final-answer date and before the response baseline is frozen. The exact cadence comes from the procurement timetable and materiality, not from a universal daily formula.
Source capture
Keep the supplier’s premise separate from what the buyer actually confirmed
A question can contain a proposed interpretation, a preferred outcome and several factual claims. The authority may answer only one of them. Store the question and answer in separate fields and quote each exactly enough to preserve meaning. Add the item identifier, publisher, publication or observation time, source link, stage, lot and cited clauses. Do not rewrite both sides into one smooth sentence, because that often upgrades the supplier’s premise into a buyer decision.
Suppose question 18 asks whether technician cover can end at 20:00 because only two lanes are scheduled later. The buyer replies: “Cover is required from the opening of the first operational lane until closure of the final operational lane on every service day.” The supported proposition is the coverage window. The statement about only two late lanes remains part of the question, not a demand forecast endorsed by the buyer. Staffing review should use the buyer’s words and seek the actual operating schedule from its controlling source.
Retain partial and non-answers. “Refer to the specification” is not empty if it points to an exact clause that resolves the issue, but it is not resolution merely because it sounds final. A response such as “details will follow” stays pending. A deleted or withdrawn question remains part of the observation record if the portal exposes it, with no invented answer. These states stop neat summaries from hiding uncertainty.
| Field | What to preserve | What not to infer |
|---|---|---|
| Question | Exact text and visible identifier | Its premises are buyer facts |
| Answer | Exact buyer text and answer state | Every subpart was resolved |
| Context | Procedure, stage, audience, lot and publication time | The answer applies more broadly |
| References | Clause, table, file, link and attachment names | The target is current or was obtained |
| Observation | Source URL, account role, retrieved time and local copy | No later publication can exist |
Impact screening
Read every answer against the bid, not against who asked it
The identity of the questioner is not a relevance test. Another bidder may expose an inconsistency your team missed, obtain a quantity needed by every supplier or prompt a corrected form. Screen the complete set against named bid domains: eligibility and compliance, technical solution, delivery and resources, price and quantities, contract position, evidence and claims, submission objects and timetable. One answer may enter several queues, but each link needs its own reason.
Use bounded first-pass states. no_effect_evidenced means a reviewer compared the answer with a named current object and recorded why it remains valid. review_required means a credible relationship exists but its consequence has not been decided. incomplete_answer preserves an open point. referenced_object_missing means the answer points to material not yet obtained. conflict_candidate routes a possible contradiction to the dedicated authority analysis. superseded_reference means the answer cites an object that appears replaced and requires baseline review. None of these labels silently edits the response.
Question 41 affects the airport price model because the answer assigns listed consumables to the fixed service fee. Question 57 affects the workbook because it names revision 3. Question 18 affects the coverage design. A fourth answer confirming the published badge application lead time may be no effect only after comparison with the mobilization plan. The register shows all four decisions and their evidence, not only the three changes.
Cross-references
A Q&A row is not complete until its referenced object is resolved
Answers often work by reference: see clause 7.4, use the revised pricing table, refer to drawing B12, or note the amendment issued today. Record the visible target exactly, then locate it in the authorized current document set. Confirm that the identifier, filename, edition, lot and content match. Preserve the access location and time. If the target is missing or the label could describe several files, keep the row open.
Do not copy a changed number from the Q&A into an old workbook when the buyer issued a replacement workbook. The native file may contain validations, formulas, new rows or instructions that the answer does not describe. Likewise, a Q&A reference to an amendment is a direction to obtain and control that amendment. AN-116 records the dependency and routes it. The amendment-lineage control determines whether the package is complete.
Sometimes the answer and target appear inconsistent. Perhaps the answer says revision 3 while the portal exposes revision 2, or the cited clause still contains the old coverage window. Label the candidate conflict and preserve both sources. Do not select a winner based on filename recency, convenience or a guess about buyer intent. That question belongs to the source-authority and conflict process.
| State | Evidence | Next action |
|---|---|---|
| target_resolved | Exact current object and locator obtained | Continue impact review |
| target_not_found | Named location checked without a match | Retrieve or escalate access |
| target_ambiguous | Several plausible objects or identifiers | Resolve identity before use |
| version_mismatch | Answer and available object name different states | Open version review |
| conflict_candidate | Buyer statements appear incompatible | Open authority reconciliation |
| outside_boundary | Reference belongs to another stage or procedure | Record exclusion evidence |
Ownership
An impact exists only when a named object and owner can be followed
“Commercial to review” is not a controlled disposition. Name the price model version, the affected consumables row, the answer identifier, the owner and the decision needed. For coverage hours, link the current shift model and service narrative. For the corrected workbook, link both the submitted candidate file and the buyer’s replacement. The owner can then prove no change, make an authorized update through the proper control or identify a blocker.
Keep intake authority narrow. The reviewer may capture the answer and propose affected objects, but a technical lead decides feasibility, a commercial authority approves price exposure, legal counsel interprets a material contract issue and the release authority accepts the resulting bid baseline. Assignment does not grant those rights. An automated comparison can nominate text and numbers for review, but it cannot convert similarity into scope or legal effect.
Give each candidate link a due time derived from the response calendar and consequence. A changed mandatory form needed for final production requires earlier closure than a low-impact background explanation. If the correct owner cannot act in time, escalate the blocked object and its bid consequence. Do not close the Q&A row because a message was sent.
Procedural effect
Publication answers the information question; it does not settle every legal effect
Public procurement rules support equal access to relevant information, but the effect of an answer depends on the governing regime and the procurement documents. The WTO Government Procurement Agreement requires replies to reasonable requests for relevant information without giving one supplier an advantage, and requires written transmission of modifications. UNCITRAL’s Model Law similarly distinguishes clarification of solicitation documents from modifications and requires shared communication in its stated setting.
Under the UK Procurement Act 2023, section 31 governs modifications to the terms of a covered procurement. Current Cabinet Office guidance expressly gives the example of a supplier clarification that requires an amendment and explains when affected tender documents must be updated and republished or provided again. That does not mean every answer is a section 31 modification. The intake register therefore records answer_published and modification_referenced as separate facts and routes the legal or procedural conclusion where material.
Timing also stays source-specific. A corrected requirement may require the authority to consider the time suppliers need to respond; it does not allow the bidder to invent a new submission deadline. Preserve any explicit extension and route the timetable effect. If none is stated, keep the published deadline until an authoritative source changes it.
Closure
Close the intake only after counts, objects, owners and future checks reconcile
Reconcile the number of visible official items with distinct rows. Explain duplicate identifiers, combined answers, withdrawn entries and unanswered questions rather than forcing the numbers to match. Reconcile every attachment and referenced object to obtained, unavailable, ambiguous or outside-boundary status. Then reconcile every candidate effect to an exact bid object, owner, next control and current state.
Coverage is not the same as completion of downstream work. AN-116 can close an item as routed_to_solution_change while the solution action remains open in its own control. It can close an item as no_effect_evidenced with the comparison attached. It cannot mark everything incorporated merely because each answer was read. Keep the source-intake state and response-change state separate.
Finally, record the last observed identifier and time, the buyer’s published final-answer date if one exists, the next source check and the event that ends monitoring. Preserve the reviewed snapshot. A later Q&A publication reopens intake; a changed answer or replacement schedule moves to version control. The final pre-release check should be explicit even when every notification mailbox is quiet.
| Control | Closure evidence | Still open when |
|---|---|---|
| Source coverage | Locations, counts, filters, pages, files and observed time | Expected location or item is inaccessible |
| Item capture | Question, answer, identifier, context and references | Answer or context is incomplete |
| Impact screening | Named comparison and bounded disposition | Credible effect has no conclusion |
| Ownership | Exact bid object, owner, due time and downstream control | Assignment lacks acceptance or authority |
| Cross-references | Every target has an evidence state | Attachment, clause or version is unresolved |
| Recheck | Next trigger and monitoring endpoint recorded | Publication window remains active without a check |
What good looks like
Useful outcomes from review published tender Q&A log
- The reviewed Q&A boundary names the procedure, stage, lot, bidder population, official locations, access context and observation time.
- Every visible question and answer has a stable register row, including items submitted by other bidders.
- Question text, buyer answer, attachments, cited clauses and any stated correction remain distinguishable.
- Pagination, filters, collapsed threads and attached clarification schedules are checked rather than assumed away.
- Each answer receives a supported initial disposition such as no bid effect found, review required, incomplete, conflict candidate or superseded reference.
- Candidate effects are linked to exact requirements, solution elements, prices, plans, evidence, forms or deadlines.
- An accountable owner and next control exist for every row that may alter the response.
- Notifications, local exports and prior reviews are retained as evidence without being mistaken for the current official set.
- The review closes with a count reconciliation, open-item list and a scheduled recheck through the relevant cutoff.
Operating model
How to run the work
- 01
Declare the official review boundary
Fix the procurement, stage, lot, audience, authorized account, portal section, attached schedules, relevant messages and observation time before collecting items.
- 02
Enumerate the complete visible set
Clear filters, expand threads, traverse every page, inspect downloads and record visible counts, identifiers, unanswered items and access gaps.
- 03
Capture question and answer in context
Preserve exact text, publisher, publication time, source locator, cited document, lot and attachment without blending the supplier premise into the buyer response.
- 04
Screen for credible bid effects
Test each answer against the response baseline for compliance, solution, delivery, price, evidence, contract, submission and timetable consequences.
- 05
Link affected objects and owners
Name the exact response object, current basis, reason for review, accountable owner and specialist control required to resolve the candidate effect.
- 06
Reconcile cross-references and files
Follow every referenced clause, table, workbook, drawing, correction and amendment, recording missing or mismatched objects as open rather than filling the gap.
- 07
Prove coverage and disposition
Reconcile portal counts, register rows, attachments, open questions and owner outcomes so that omissions and unresolved items remain visible.
- 08
Schedule the next official check
Record the last observed item and time, relevant buyer-answer milestone and next source-check trigger until the controlled Q&A intake period closes.
Evaluation
Questions that change the decision
- Which official Q&A views, files and participant messages are within this review boundary?
- Which supplier population and procurement stage was each answer made available to?
- Does the visible item count reconcile after filters, pagination and collapsed threads are removed?
- Is the buyer answer complete, pending, withdrawn, replaced, linked elsewhere or ambiguous on its face?
- What precise proposition comes from the buyer rather than from the supplier’s question?
- Which lot, terminal, equipment family, period, form or requirement does the answer expressly cover?
- Could it change compliance, solution, staffing, schedule, price, evidence, contract treatment or submission mechanics?
- What existing bid object uses a basis that the answer might confirm, narrow, expand or displace?
- Which specialist control owns the next decision and what evidence will close the row?
- When must the official source be checked again, even if no further alert arrives?
Failure modes
Where teams lose control
The team reads only answers to its own questions and misses a material answer given to another bidder.
An email digest is mistaken for the full portal record even though it was delayed, filtered or limited to new activity.
A portal default filter hides answered, older, withdrawn or another-stage threads.
The buyer’s answer is paraphrased together with the bidder’s assumption, making the premise look authoritative.
A linked workbook or drawing is not downloaded, so the team records a correction without obtaining the corrected object.
An answer that merely points back to the pack is labelled no impact without checking the cited passage.
A broadly worded answer is applied to every lot or terminal even though the published scope is narrower.
A clarification is treated as a formal amendment, or an amendment reference is treated as optional context.
A candidate effect is sent to a generic distribution list and no named person proves the resulting disposition.
The register is declared complete before the buyer’s final-answer milestone or without a recheck trigger.
Measurement
Measure the finished job
Measure the completed workflow, including review effort and exceptions. Output volume on its own is not evidence of a better process.
- Official Q&A locations and attached schedules included in the declared boundary
- Visible portal items reconciled to distinct register rows
- Questions from other bidders reviewed as a share of the complete set
- Rows with exact answer text, source locator, publication context and scope
- Answers referencing a file or clause whose target was successfully resolved
- Candidate bid effects linked to an exact object and accountable owner
- Rows closed as no effect with recorded comparison evidence
- Open, incomplete, conflict-candidate and missing-object rows past their review time
- Elapsed time from official publication to controlled intake
- Scheduled source checks completed through the relevant clarification milestone
Questions
Common questions
Do we need to read questions submitted by other bidders?
Yes, when they are part of the official information made available to your bidder population. Relevance comes from the answer’s scope and effect on your bid, not from who submitted the question.
Is the notification email enough evidence of the latest answers?
No. Retain it as a source-check signal and receipt record, then inspect the official Q&A locations, surrounding entries, attachments and document changes through the authorized route.
How do we prove the whole Q&A log was reviewed?
Record the official locations, stage, audience, access role, filters, page count, visible count, first and last identifiers, attachments and observation time, then reconcile them to distinct register rows.
Does every published answer change the tender documents?
No. Preserve the answer and screen its effect. Where it references or appears to make a formal change, obtain the controlling document and route the procedural conclusion instead of assuming it.
What if the buyer only says to refer to the specification?
Follow the cited provision and test whether it resolves the actual question. Keep the row incomplete if the reference is vague, unavailable, inconsistent or silent on a material part.
Can one answer affect several parts of the response?
Yes. Create separate links to each affected requirement, solution element, price, plan, form or deadline, with its own owner and disposition rather than hiding them in one impact note.
When can an answer be classified as having no effect?
Only after a reviewer compares the exact answer with named current bid objects and records why their supported basis remains valid. Familiar wording or lack of an obvious keyword is insufficient.
When does Q&A monitoring stop?
Use the procurement’s own final-answer milestone, formal closure message, submission control and any later authorized update route. Always retain an explicit final source check and its timestamp.
Sources
Primary references
- Procurement Act 2023, section 12 on covered procurement objectives UK Legislation
- Procurement Act 2023, section 31 on modifying a procurement UK Legislation
- Procurement Act 2023, section 54 on time limits UK Legislation
- Cabinet Office guidance on covered procurement objectives, updated 15 April 2026 Cabinet Office
- Cabinet Office guidance on modifying a competitive procurement, updated 17 August 2026 Cabinet Office
- Cabinet Office guidance on time periods, updated 17 August 2026 Cabinet Office
- Cabinet Office guidance on electronic communications, updated 17 August 2026 Cabinet Office
- High-value school procurement guidance on supplier questions and shared logs Department for Education
- Current invitation stating that responses are issued through a log visible to all tenderers Find a Tender service
- Current published instructions on sharing clarification issues with tenderers Find a Tender service
- Current notice requiring written questions and shared responses Find a Tender service
- Agreement on Government Procurement, Article X on information and modifications World Trade Organization
- UNCITRAL Model Law on Public Procurement, articles 15 and 16 United Nations Commission on International Trade Law
- World Bank Procurement Regulations, seventh edition, September 2025 World Bank
- Directive 2014/24/EU, articles 18, 22, 47 and 53 European Union
- Simap online help on Q&A display and notification timing Simap
About the author
Tony Kim
Tony writes about applied AI, dependable product engineering and the systems that turn complex response work into controlled delivery.
Ziva
Proposal software for source-grounded RFP, RFI, DDQ and questionnaire response work.
Bid, proposal, presales, security and compliance teams. Start with the workflow, constraints and evidence you already have.