Small marketing teams often treat “publish something new” as the default answer to a weak content result. Sometimes the right move is a precise refresh, a consolidation or a better internal link. The decision should follow the page’s intent, evidence and commercial role, not the number of unused ideas in a topic bank.
1. Define the job and the constraint
Write the page’s intended job: answer a question, support comparison, prove expertise, collect a request or guide a sales conversation. Add the audience, market, owner, money-page route and available capacity. State the constraint that makes the choice important, such as limited subject-matter access, a migration deadline or a thin editorial team.
Choose one decision: refresh, consolidate, redirect, create new, or hold. “Do both” is not a decision until the relationship between the pages is explicit. Keep the old URL and any publication hold visible while the choice is being made.
2. Inspect existing performance and intent
Google’s Search Console performance report can provide page and query trends for clicks, impressions, CTR and position. Export the relevant period, property, country, device and page group. Record changes in title, canonical, template, links and tracking before interpreting movement.
Look for impressions with a weak answer, clicks with poor next steps, stable qualified leads, declining relevance, queries that belong to another page and evidence that the page already serves a useful audience. A low click count alone does not prove that the page should be replaced.
3. Test whether the intent is still valid
Write the current primary question in the language a reader would use. Compare it with the page title, introduction, headings, examples, evidence and next step. Mark whether the page answers the question completely, partially, indirectly or not at all.
If the underlying job is unchanged and the evidence is stale, a refresh may be enough. If the job has split into two distinct questions, consider separate pages only when each can offer a genuinely different outcome. If the query now belongs to another service or audience, a redirect or consolidation may be safer than another rewrite.
4. Check originality and usefulness
Google’s people-first content guidance asks whether a page offers original information, substantial coverage and a satisfying reader experience. Use these questions to inspect both the existing page and the proposed new brief.
List the evidence a refresh can add: a current process, original analysis, a decision table, limitations, worked example or clear ownership. Google’s SEO Starter Guide gives a baseline for making pages understandable to search systems, but it is not a substitute for editorial judgment. If the proposed new article would repeat the same definitions, examples and CTA, it is not yet a new content function. Strengthen the angle before commissioning it.
5. Compare overlap and page roles
Build a small map of existing URLs, titles, primary queries, audience, intent, stage, money-page route and status. Mark direct overlap, complementary support, historical legacy and unknown. Include old drafts and pages not currently in the main navigation.
Decide how a reader should move between related pages. A comparison page can link to a service page; a diagnostic can link to an implementation guide. Two pages that compete for the same question without a clear role create editorial and measurement noise even if their titles differ.
6. Estimate effort and evidence access
Score each option for research effort, subject-matter access, technical work, review time, internal-link changes, redirect risk and measurement readiness. Use qualitative bands rather than an invented hourly benchmark. A small refresh that needs legal or product review may take longer than a new page with a clear source set.
Record what is known, assumed and missing. If no owner can verify the claim, either narrow the claim or hold the work. A content calendar should not convert an unresolved evidence gap into a publication deadline.
7. Use a decision matrix
| Situation | Best first action | Why | Required evidence | | — | — | — | — | | intent stable, evidence stale | refresh | preserve useful history | new sources and changed sections | | two pages answer one job | consolidate | reduce competition | overlap map and redirect plan | | intent split into distinct jobs | create new plus link | clarify roles | separate audience and outcome | | page has demand but weak path | refresh CTA or route | protect existing discovery | lead-path test | | no clear audience or evidence | hold | avoid thin expansion | owner and research brief | | outdated or unsafe claim | remove or replace | prevent misleading content | review and redirect decision |
Assign an owner and release condition to every row. The matrix is a prioritization tool, not an automatic publishing approval.
8. Run a bounded content experiment
Choose one page family and one intervention. For a refresh, preserve the URL and log each material change. For a new page, define the unique function, canonical intent, internal links and control URLs. Keep publication, indexing and redirect actions behind their separate approval gates.
Measure reader behavior, valid inquiries, qualified outcomes and overlap exceptions separately. If a refreshed page gains visibility but creates no useful next step, inspect the offer and audience. If a new page attracts the same queries as an existing page, pause expansion and revisit the role map.
9. Record the final choice and revisit it
Create a decision record with current URL, proposed action, intent, audience, evidence, overlap neighbors, owner, source register, QA status and next review date. State what would change the decision. This makes a later refresh explainable instead of cyclical.
The best choice is the smallest change that gives the reader a clearer answer and the team a more trustworthy evidence path. New content is valuable when it adds a distinct job; a refresh is valuable when it restores usefulness. Both deserve the same discipline around sources, overlap and commercial claims.
How did this article land?
Choose one reaction. You can change it anytime.
