New editorial guide, published 17 October 2026. The historical sources linked below are preserved separately and are not represented as current product or service information.
A website inventory can become an impressive spreadsheet that changes nothing. Hundreds of addresses, titles and dates are collected, but nobody knows which page to fix first or why two pages disagree. The missing layer is usually decision-making. A useful inventory records what a page does, what evidence supports it and what should happen next. This article describes a small editorial review method using an invented resource website. It does not assume that old pages are bad, that low traffic proves low value, or that deleting content will improve search performance. The intended output is a reviewable work queue, not a numerical score pretending to settle every editorial question.
Define the collection before counting it
Write down which addresses belong in the review. A public page, a redirected address, a downloadable file and an internal search result are different kinds of object. Include the ones relevant to your task and label their types. Decide how you will handle query strings, language variants and trailing slashes before merging records. Otherwise, one reviewer may count several addresses as one page while another treats them as separate items. Save the collection date and method so a later reviewer knows what the inventory could and could not observe.
For the invented resource site, begin with the public navigation and an available sitemap, then compare those addresses with an approved content export. A page missing from navigation may still have a legitimate audience through direct links. A sitemap entry may no longer return useful content. Record those observations rather than automatically removing either item. Keep a raw address column beside any normalised address. Normalisation helps organise the work, but the original address is needed to investigate a redirect chain or determine where an old link actually sends someone.
Record purpose and audience in plain language
For each page, describe the reader task in one sentence. Examples might be finding eligibility requirements, downloading a current form or understanding a historical announcement. A page labelled general information has not yet been understood well enough to review. Record who is responsible for its factual accuracy and who can approve an editorial change. If no owner can be identified, mark the gap. Do not invent responsibility by copying a name from an old byline or assuming that the person who uploaded the page still manages the subject.
Separate current usefulness from historical value. An old event announcement may no longer support registration but still document what was announced at that time. Its appropriate treatment could be an archive label and removal of an obsolete action, not a rewrite that makes the event appear forthcoming. Conversely, a current instructions page may need prompt revision even if its publication date is recent. Age is a clue, not a verdict. The inventory should help the reviewer ask whether the page’s wording, actions and date labels match its present role.
Use evidence fields that support a decision
Collect enough evidence to answer the review question, not every available metric. Useful fields can include the visible title, stated task, response status, canonical address, last substantive review and references requiring verification. If analytics are available and permitted, record their definitions and period. A low observed visit count is not a direct measure of accuracy or importance. A page needed only during an unusual incident could still matter greatly. Missing analytics should remain missing rather than being replaced with an invented zero that makes the page look unused.
Use a note explaining the evidence behind a proposed action. Keep because it answers a unique current question is more useful than keep because score seven. Revise because the only download points to an obsolete form identifies both a problem and a checkable repair. Combine because two pages answer the same task needs evidence that their audiences and requirements really overlap. Do not let a colour scale hide uncertainty. If a reviewer has not inspected a page’s main content, mark it unreviewed rather than treating a completed metadata row as a completed editorial assessment.
Choose an action with a reason and a destination
Use a small action set: keep, revise, combine, archive, retire or investigate. Define each term for the project. Archive might mean retaining historical wording with clear context; retire might mean removing an obsolete public page after handling its links and records. The two should not be interchangeable. A combine decision needs a proposed surviving destination and a list of information that must not be lost. A retire decision needs a check for dependencies such as navigation, external references, downloadable files and material that must remain available for a separate reason.
Avoid redirecting every removed address to the homepage. Ask whether a genuinely relevant replacement exists and whether the destination fulfils the old link’s likely purpose. If there is no replacement, the site should communicate absence honestly rather than claiming success with unrelated content. Implementation details depend on the hosting system, so record the desired behaviour for technical review. Do not change live routes simply because the spreadsheet suggests a tidy structure. The inventory proposes decisions; the release process verifies their effect on readers and the rest of the site.
Prioritise consequences, dependencies and effort
A queue should distinguish an urgent misleading instruction from a harmless stylistic inconsistency. Consider the consequence of leaving the page unchanged, the number of other pages depending on it and the effort needed to resolve its uncertainty. A small correction to a widely reused definition may unblock several articles. A large redesign of a low-impact index may be less urgent. Record the reasoning rather than relying on a precise-looking priority number. The team should be able to explain why the first task comes before the second without invoking a mysterious formula.
Break large actions into verifiable steps. Revise the entire advice section is difficult to finish or review. Confirm the current eligibility rule, update three affected paragraphs, replace one download and test the links is a clearer task. Identify what can proceed independently and what requires a subject owner’s answer. A blocked factual question should not prevent unrelated link repairs, but a cosmetic repair should not be presented as resolution of the factual problem. Give each item a next action and owner, not merely a status that says in progress indefinitely.
Close the loop after publication
Once a change is released, inspect the actual public page. Check its main text, destination links, date labels and mobile layout against the approved decision. An updated source document is not proof that the public website changed. Record the release address and verification time in the inventory. Keep the prior decision note so a later reviewer can understand why material moved or became historical. If a release reveals an unexpected dependency, reopen the affected task with that evidence rather than silently changing the definition of done to match what happened.
Schedule future attention according to the page’s subject and risk of change, not one universal annual date. A stable explanatory essay may need a different review trigger from operating instructions tied to a changing product. Google’s people-first guidance discourages changing dates merely to make unchanged content appear fresh. A review record should distinguish inspected without change from substantively revised. The inventory becomes valuable when it preserves these decisions over time: what the page was for, what evidence was available, what changed and how the published result was checked.
References & further reading
From the recovered archive
- Why Your Blog Isn't Growing - 15 Trivial Reasons · captured 1 July 2021
- How To Create A Blog From Scratch And Make Money From It: The New Blogger's Guide · captured 25 May 2021