A quarterly content audit for a small website that leads to decisions
Run a quarterly content audit for a small website using four decision lanes, bounded evidence, and a realistic 90-day action queue.
The first content audit usually begins with good intentions and ends with a spreadsheet nobody wants to reopen. Every URL gets twelve columns, three weighted scores, a traffic snapshot, a color, and a recommendation. By the time the inventory is complete, the team has spent its energy describing the site instead of improving it.
A small website does not need an enterprise audit. It needs a reliable way to answer four questions:
Which pages should stay as they are?
Which pages deserve a focused improvement?
Which pages should be combined?
Which pages should be retired?
That is the job of a quarterly content audit. The deliverable is not a perfect database. It is a short, defensible queue of decisions for the next 90 days.
Start with the decisions, not the columns
Open a sheet and create one row per indexable page. Add only the fields you will use during the review:
URL and page title
Page job
Primary audience or query
Current evidence
Decision
Reason
Owner and due date
“Page job” matters more than most audit scores. A product comparison, a tutorial, a definition, and an opinion article are not supposed to behave the same way. If you compare them only by traffic, the audit will punish valuable pages that support a sale, answer a narrow customer question, or establish expertise.
Write the job in a verb phrase: help a founder choose, explain a setup, resolve an objection, capture a qualified search, or route a reader to a deeper resource. A page with no clear job is already giving you useful information.
Gather enough evidence to make a choice
For a small site, evidence can come from four places:
Search visibility: impressions, clicks, click-through rate, and the queries connected to the page
On-site behavior: engaged visits, key actions, and useful next-page movement
Commercial usefulness: whether sales calls, customer support, onboarding, or outreach relies on the page
Editorial quality: accuracy, originality, clarity, and fit with the business today
Google’s Search Console guidance shows how to examine top pages, queries, low click-through rates, and changes over time. It also warns, in effect, against treating one movement as proof of one cause: competitors, demand, and other outside events can change at the same time.
Keep that uncertainty visible. Use confidence labels such as observed, reported, and inferred:
Observed: the page received impressions for a set of queries.
Reported: two sales calls mentioned the article.
Inferred: a clearer comparison may improve qualified clicks.
This is more honest than converting unlike signals into one impressive-looking score.
Use four decision lanes
Keep
Keep a page when its job is clear, the information remains accurate, and there is no stronger overlapping page. “Keep” does not mean ignore forever. It means no meaningful work belongs in the next quarter.
Record why it stays. A simple note such as “best explanation of the setup question; still accurate; receives qualified product-assist traffic” prevents the same debate next time.
Improve
Improve a page when the job is still valuable but execution is weak. The change should be named before it enters the queue.
Weak recommendation: “refresh article.”
Useful recommendation: “rewrite the opening to answer the decision sooner, add the missing cost boundary, verify the outdated screenshot, and test a clearer search title.”
An improvement should have a hypothesis. If the page has strong impressions and weak clicks, a title or description test may be reasonable. If readers arrive but do not understand the next step, the content and routing may deserve attention. Do not make ten unrelated changes and then pretend the result taught you which one mattered.
Combine
Combine pages when they compete to do substantially the same job for the same reader. This is not a word-count exercise. Two short pages can be distinct; two long pages can be redundant.
Before combining, name the strongest contribution from each page. Decide which URL should become the durable destination. Then plan the merged structure, internal-link updates, and the correct URL treatment. Editorial consolidation and technical consolidation are connected, but they are not identical decisions.
Retire
Retire a page when its job no longer matters, its information cannot be maintained, and it contributes no unique value worth preserving. Retirement still needs care. Check whether the page has backlinks, internal links, recurring direct visits, or a better destination for its audience.
Do not use retirement to hide an uncomfortable maintenance problem. A page that is difficult to update may reveal a missing owner or an unstable promise. Fixing that operating issue can be more valuable than deleting the evidence of it.
Run the audit as a 90-minute workshop
A bounded session keeps the review oriented toward action.
Minutes 0–15: scan the site shape
Group pages by job or topic. Look for obvious clusters, orphaned pages, outdated product language, and multiple URLs answering the same question.
Minutes 15–55: classify the clear cases
Move quickly through pages with strong evidence. Mark uncertain cases without solving them immediately. The goal is to build momentum and reserve discussion for choices that truly need it.
Minutes 55–75: resolve the uncertain cases
For each uncertain page, ask what evidence would change the decision. If a ten-minute check will answer it, do the check. If not, assign a confidence label and choose the lowest-risk next action.
Minutes 75–90: build the quarter queue
Select only the work the team can finish. A 40-item audit backlog is another abandoned spreadsheet. Five well-defined actions with owners and dates are a content plan.
Handle two difficult edge cases
The page with no traffic but strong customer value
Search traffic is not the only legitimate job. A niche implementation guide might save support time or help a late-stage buyer. Keep or improve it if that value is real, and record the evidence. Do not disguise the decision as SEO.
The page with traffic but weak business fit
A popular page can attract the wrong audience, create support demands, or pull the brand toward a topic it no longer wants to own. Traffic makes the decision consequential, not automatically positive. Consider whether the page can be reframed, routed, or responsibly retired.
Apply a people-first quality check
Google’s people-first content self-assessment asks whether a page adds original value, demonstrates trustworthy expertise, serves an intended audience, and helps someone achieve a goal. Those questions are useful beyond search because they test whether a page deserves continued maintenance.
For every improve decision, write one sentence finishing this prompt:
After the change, the right reader will be able to…
If the team cannot complete the sentence, the work is not ready for the queue.
Leave a one-page closeout record
At the end of the audit, summarize:
Pages reviewed
Decisions by lane
The five to ten actions accepted for the quarter
Important uncertainties
Pages intentionally left unchanged
The next review date
That closeout note is the bridge to the next quarter. It lets the team compare decisions with outcomes instead of repeating the same inventory exercise.
A good quarterly content audit should feel slightly incomplete. You will not investigate every anomaly or rewrite every weak sentence. The discipline is choosing the most useful actions, preserving the reasoning, and returning after enough time has passed to learn from the result.
