Website knowledge-centre publishing Q&A

Question 1

What should the footer link to the tender Q&A resource be called?

Accepted Answer

Use a descriptive label such as 'Tender & Bid Knowledge Centre' and link to a stable route such as /knowledge/.

The landing page should explain coverage and provide search and topic navigation.

This is a proposed implementation choice, not a ranking formula.

Keep the label understandable to visitors unfamiliar with internal content or technical terminology.

Question 2

Should a million Q&As be rendered on one web page?

Accepted Answer

No practical implementation should send the entire collection to every visitor.

Use a searchable store and load a bounded set of relevant answers or category pages.

Treat Markdown as an editorial exchange format.

The appropriate database and hosting design depend on the existing website, but users should reach individual answers without downloading the whole corpus.

Question 3

How should question synonyms be handled on the website?

Accepted Answer

Map alternative phrasings to one canonical answer when they express the same intent.

Include synonyms in search indexing without creating a public page for every wording.

Split questions only when the substantive answer changes.

This keeps the knowledge centre easier to maintain and helps avoid contradictory copies of the same guidance.

Question 4

What metadata should every published answer display?

Accepted Answer

Display a clear title, applicable jurisdiction, source links, review date and whether the answer is a rule, documented product claim, practical advice or an example.

Show important limitations near the answer.

Do not present a machine-generated timestamp as proof of substantive review.

The metadata should help a reader decide whether the answer applies to their situation.

Question 5

What is the difference between source-checked and professionally reviewed?

Accepted Answer

Source-checked means the material was compared with the cited source within the stated scope.

It does not mean a qualified lawyer, procurement specialist or product tester has independently approved every conclusion.

Record professional review separately with the real reviewer and date when it occurs.

Never invent reviewer names, credentials or approval records.

Question 6

How should stale answers be handled?

Accepted Answer

Track which answers depend on each material source and reassess affected records when it changes or a review date passes.

Preserve earlier versions and use an appropriate warning or publication hold where needed.

Apply changes to the correct jurisdiction, product and plan rather than updating every answer mechanically.

A successful fetch or rebuilt page does not establish substantive verification, and the displayed review date should change only when an actual content check occurs.

Question 7

How should a knowledge centre distinguish UK, EU and German answers?

Accepted Answer

Use an explicit jurisdiction field and visible filters, with cross-links where comparisons help.

Explain narrower applicability inside the answer.

Language is a separate field: an English-language answer about Germany remains German-jurisdiction guidance.

Avoid changing only a country name in a generic paragraph and treating it as a new legally verified article.

Question 8

Should all draft answers be indexed by search engines?

Accepted Answer

Publish and index only content that meets the site's editorial criteria and offers standalone value.

Keep drafts behind access control where appropriate.

If a public page needs noindex, allow the crawler to fetch that instruction; a robots.txt block can prevent it from being seen.

Noindex does not make confidential information private.

Question 9

Will FAQPage markup produce FAQ rich results on Google now?

Accepted Answer

Google's documentation update says FAQ rich results stopped appearing from 7 May 2026; a later update records removal of the feature documentation.

As reviewed on 3 November 2026, do not promise this rich result as a benefit of the website.

Publish useful answers and use only structured data appropriate to the actual content and currently supported use case.

Question 10

Does a large AI-generated knowledge centre automatically violate Google's policies?

Accepted Answer

Size or AI use alone does not establish a violation.

Google's scaled-content policy focuses on producing many pages primarily to manipulate rankings without providing user value.

For this resource, a sensible editorial policy is to require distinct intent, useful substance and evidence before publication.

A numerical target should not override those criteria.

Question 11

How should sitemaps work if the knowledge centre becomes very large?

Accepted Answer

Google limits a single sitemap to 50,000 URLs or 50 MB uncompressed.

Split larger sets and use a sitemap index as appropriate.

Include the canonical pages you actually want considered for indexing, not every draft or search-filter combination.

Sitemap submission helps discovery but should not be described as a guarantee of indexing.

Question 12

Should every answer have its own URL?

Accepted Answer

Use a stable URL when the answer has enough standalone substance and a distinct user need.

Closely related short questions can live on a coherent topic page with accessible anchors.

This is an editorial and product decision.

Do not create thin pages solely to increase the URL count; test how users find and navigate the material.

Question 13

How should canonical URLs be chosen?

Accepted Answer

Choose a consistent primary route for each substantive page and link to it internally.

Use canonical signals carefully for duplicate versions, preserving genuine language or jurisdiction differences.

Google treats canonical declarations as signals rather than guarantees.

Avoid making different substantive answers canonical to the same page simply to conceal a weak content structure.

Question 14

What should the website search support?

Accepted Answer

Support relevant keywords, common abbreviations, topic and jurisdiction filters, and a clear no-results state.

Rank direct matches and useful related answers before obscure tangents.

Keep a record of failed searches without unnecessarily collecting personal data.

Search quality should be tested with realistic user questions rather than only the exact titles in the database.

Question 15

How should Markdown be converted safely for the website?

Accepted Answer

Parse it using a maintained renderer and apply an appropriate sanitisation policy.

Treat imported text as data, not executable code or instructions to fetch arbitrary resources.

Validate links and metadata and escape content for its output context.

The exact implementation depends on the site's stack; review that stack's current documentation before adding code.

Question 16

How should accessibility be built into the FAQ interface?

Accepted Answer

Use clear headings, labelled controls, keyboard-operable interactions, visible focus and readable presentation.

Ensure collapsed answers can be opened without a mouse and that status is communicated accessibly.

WCAG 2.2 provides the reference framework.

Test the actual rendered interface; choosing an accessibility-friendly component name is not proof of conformance.

Question 17

What should a report-an-error feature collect?

Accepted Answer

Ask for the answer identifier, the suspected problem and supporting information if available.

Keep the form simple and avoid requesting confidential tender documents.

Route reports into a review queue with a clear owner.

A correction process improves the resource only if issues are assessed and resolved; do not display unverified user allegations as factual corrections.

Question 18

How should the knowledge centre handle citations?

Accepted Answer

Keep stable source identifiers in the data and display descriptive links beside the answer.

Retain source title, publisher, access date and a note about what was checked.

Distinguish a vendor's own product statement from independent evidence.

A citation should support the associated claim, not simply point to a reputable site's home page.

Question 19

How can Codex import the resource without changing the rest of the site?

Accepted Answer

Have it inspect the existing architecture, create a dedicated knowledge route, import the validated records and add the footer link through the site's normal component.

Preserve source metadata and draft status.

Verify navigation, search, rendering and representative citations in a preview.

Deployment should follow the user's established repository and hosting workflow rather than assuming a new platform.

Question 20

How should future expansion towards one million entries be governed?

Accepted Answer

Add distinct questions supported by real user need and evidence, deduplicate by intent, and pass the relevant review gates before publication.

Store pending candidates separately from approved answers and report actual counts.

The target is a long-term capacity objective, not evidence that a million useful topics currently exist or that this edition already contains them.

Website knowledge-centre publishing Q&A

What should the footer link to the tender Q&A resource be called?＋

Should a million Q&As be rendered on one web page?＋

How should question synonyms be handled on the website?＋

What metadata should every published answer display?＋

What is the difference between source-checked and professionally reviewed?＋

How should stale answers be handled?＋

How should a knowledge centre distinguish UK, EU and German answers?＋

Should all draft answers be indexed by search engines?＋

Will FAQPage markup produce FAQ rich results on Google now?＋

Does a large AI-generated knowledge centre automatically violate Google's policies?＋

How should sitemaps work if the knowledge centre becomes very large?＋

Should every answer have its own URL?＋

How should canonical URLs be chosen?＋

What should the website search support?＋

How should Markdown be converted safely for the website?＋

How should accessibility be built into the FAQ interface?＋

What should a report-an-error feature collect?＋

How should the knowledge centre handle citations?＋

How can Codex import the resource without changing the rest of the site?＋

How should future expansion towards one million entries be governed?＋