Siobhan.GEditor
Choose which existing SEO pages to refresh by looking for a genuine gap in usefulness, accuracy or customer experience—not simply the oldest publication date. A three-year-old guide may still answer its question beautifully, whilst a service page written last month can already be wrong because the business changed its offer. For a small UK agency, editorial capacity is precious. A sound refresh-priority review should identify which readers are affected, what is actually wrong, who can verify the correction and how the improved public page will be accepted. The goal is not a calendar full of revised dates. It is a website that answers important questions accurately and remains manageable for the people responsible for it.
Begin with the decisions the website helps customers make
Build a short inventory of service, product, location, support and educational pages that matter to the business. A service page may need to explain who qualifies, which activities are included and how to make an enquiry. An educational guide may need to let someone diagnose a real problem without purchasing a service. A specialist support answer can be valuable even when its search traffic is low. Note each page's canonical URL, intended reader, primary question, responsible business owner and current useful related links. This information is often more revealing than a spreadsheet containing only title, keyword and traffic.
Ask whether the original purpose is still legitimate. An article describing a discontinued service can mislead customers even if every sentence is grammatically sound. A page that duplicates two other guides may require a decision about where its unique information belongs before anyone rewrites it. Conversely, an old technical explanation with current sources and a clear answer may deserve no intervention at all. The first priority is usually an incorrect promise or blocked customer journey, not a visually untidy publication date. Note where an authorised employee must confirm current business facts; an editorial persona cannot supply that authority.
Understand the evidence before deciding that traffic is a content problem
Search Console clicks and impressions can reveal where a page appears in search and whether recorded patterns have changed. They cannot by themselves establish that the writing is poor. A drop may involve changing demand, site migration errors, indexing behaviour, search-result presentation, competing information or the effects of seasonality. Compare suitable periods and relevant page groups, noting the search type, country and device when the distinction changes the conclusion. A small specialist guide with very few impressions can show dramatic percentage movement from a tiny absolute change. Avoid using that movement as a convenient justification for an expensive rewrite.
Google's traffic-drop investigation guidance explains that different kinds of decline warrant different diagnostic questions. Check the site's change register: did the page move, lose internal links or acquire an unintended robots directive? Has the service itself changed? Is there evidence that customers are dissatisfied with the actual answer? Separate observations from explanations in the working review. If a technical fault prevents discovery or access, repair that problem before expanding content. If the available measurement is incomplete, record the limitation and identify who can verify it rather than inventing a cause.
Inspect the article for a specific missing answer
Read the actual page as its intended visitor would. Does the opening paragraph identify the issue clearly? Are important eligibility conditions, current steps and exceptions easy to find? Would a customer understand what happens after contacting the business? If a guide promises to explain a technical procedure, does it identify prerequisites, failure cases and a way to check the result? This direct editorial review often reveals a more useful improvement than keyword-density scores. A real gap might be a missing decision table, an inaccurate product reference or outdated instructions—not an absence of generic introductory paragraphs.
Preserve accurate sections rather than treating an update as permission to replace everything. Recheck cited facts and their scope: an official software manual may establish how a feature operates, but cannot prove an individual client's implementation. A company representative can confirm the actual service offer, but cannot guarantee an external search ranking. An illustration can clarify a process but should not masquerade as a client case study. Google's people-first content guidance
Resolve overlap with neighbouring pages before commissioning work
Two articles may share keywords because their topics genuinely relate, or because they repeat the same answer. Read the substantive content before classifying them as competitors. An introductory guide and an advanced implementation checklist can coexist when each serves a distinct task. Two articles with different titles but the same audience, examples and next step may confuse navigation and create unnecessary maintenance. Map the primary purpose of each page and the unique details worth retaining. Where the pages serve different readers, make their boundaries and cross-links clearer instead of merging them automatically.
If consolidation is justified, approve which URL should remain and what useful material must be carried across. Review incoming internal links, important external references and any genuinely appropriate redirects. Google's canonicalisation guidance
Distinguish quick corrections from projects requiring new evidence
Classify candidates into useful treatments: retain, investigate, factual correction, targeted refresh, substantial restructuring, consolidation or retirement. Each has a different commitment. Fixing a broken source link needs a valid replacement reference and a check that the claim remains supported. Correcting a service eligibility statement needs an authorised operational reviewer. Rewriting a guide whose structure fails the task may need fresh research, a new outline and specialist review. Consolidation requires careful URL and content planning. None should be treated as an identical unit simply because the content calendar has one column marked refresh.
Estimate effort and dependencies honestly. A quick paragraph amendment can become complicated if the client has not approved the underlying policy. A seemingly large topic might be straightforward if reliable source material and a knowledgeable reviewer are available. Record the reason for the proposed change in testable language: 'explain the quotation exclusions approved by operations' is more actionable than 'add more SEO content'. If the page already serves its readers well, choose retain and monitor. Protecting good material is a legitimate editorial outcome, not evidence of inactivity.
Prioritise by customer consequence, evidence and delivery readiness
Use a short decision framework rather than a fabricated numeric prediction of SEO gains. Consider the importance of the customer journey, seriousness of the verified gap, confidence in the diagnosis, number and type of affected pages and whether the required approval is available. A critical service page with materially wrong information should generally outrank a low-impact title tweak. A broken contact journey can require prompt attention regardless of search-volume data. A seasonal guide may be reviewed before its relevant period, but only if the underlying subject remains important. The team should be able to explain each decision in ordinary language.
Some apparently attractive projects are not ready. A client may need to confirm service details, a developer to repair the template, or a subject reviewer to validate technical advice. Keep those dependencies visible rather than assigning a writer to invent missing facts. A smaller ready task can proceed whilst an important issue is appropriately escalated. GOV.UK's agile planning guidance
Check the saved revision and, separately, the visitor's real page
For each approved refresh, write an acceptance brief: what must remain, what must change, which source supports the correction, who approves it and what a reader should be able to do afterwards. Confirm the article's exact existing slug and title purpose before saving. If the CMS stores the body, comparison and FAQ separately, verify every intended field under the correct record. A native Production UPDATE followed by exact-ID READ-BACK proves the fields were saved; it does not prove that caching, the public renderer, links, images or forms behave correctly. Keep those acceptance states distinct.
When authorised browser testing is available, inspect the canonical public page at representative desktop and mobile widths, confirm the figure's attribution and source, examine any responsive comparison table and test links or customer actions relevant to the change. Record unavailable tests as unverified, not passed. After deployment, monitor suitable search and business observations without claiming that one edit caused every later improvement. Refreshing content is worth doing when a demonstrated reader need is served better. The clearest priority review often produces fewer changes, each with a better reason and a more defensible verification record.