Introduction

Every few weeks, someone on a sales call asks me how many pages they need to rank. Sometimes they’ve counted a competitor’s blog posts. Sometimes a previous agency sold them eight articles a month, and they want to know if that’s enough.

I get the appeal. A page count is easy to budget, assign, and report to a CEO. The trouble is that it measures effort, and neither Google nor AI search engines reward effort. They reward the page that best answers the question in front of them.

“How many pages?” is an outdated question. A better one is: “What does our audience need that our site doesn’t already handle well?” Start by evaluating the content you already own before deciding what to publish next.

Why More Pages Stopped Meaning More Traffic

Years ago, every keyword variation got its own URL. “Local SEO services,” “local SEO company,” and “local SEO agency” became three nearly identical pages, and the traffic chart went up and to the right.

Search engines are now better at recognizing that those queries share one intent. Splitting them across three URLs doesn’t triple your chances. Instead, the pages often compete with each other, dilute internal links, and leave Google guessing which version to rank.

We see this all the time with multi-location brands: near-identical location pages with swapped city names and a dozen blog articles answering the same five questions. Nobody planned it that way. The pages piled up one “quick post” at a time, and now every URL needs maintenance.

Google’s March 2024 core update targeted unhelpful, unoriginal content, and Google later reported a 45% reduction. The update also introduced the scaled content abuse policy, targeting content produced at scale to manipulate rankings, whether written by humans, machines, or both.

AI Overviews and AI Mode added another dimension. According to Google’s documentation, both can use “query fan-out” to run related searches across subtopics and build an answer. Your page doesn’t need to match a keyword exactly. It needs to provide a strong answer to part of what the person is trying to figure out.

But the fundamentals haven’t changed. In January, Google’s Danny Sullivan said Google doesn’t want publishers to break content into bite-sized chunks for LLMs or to create separate versions for AI and people. The focus remains on helpful, reliable, people-first content. One well-built page can answer several related questions without needing a separate URL for each.

Audit What You Have Before You Add Anything

Before anyone writes a new brief, pull data on the pages you already own. Search Console, analytics, and your SEO platform will give you impressions, clicks, rankings, CTR, and engagement. Your CRM will tell you which pages show up in the journeys of people who actually became customers. That second step gets skipped a lot, and it matters.

Traffic alone will mislead you. A personal injury firm’s article on how long a claim takes might get modest organic traffic and almost never produce a lead on the first visit. But if prospects read it before booking a consultation, it’s doing real work. Meanwhile, a viral article can drive thousands of visits from people who will never buy anything.

For every page that matters, write down:

  • The primary topic and search intent.
  • Who it’s for and what buying stage they’re in.
  • The business goal it supports.
  • How it’s performing now.
  • What you recommend doing with it.

That turns a messy content library into a decision tool. Group the pages into topic clusters while you’re at it. Gaps and overlaps become obvious once related URLs sit side by side: three articles chasing the same intent, an orphaned guide nothing links to, an old article taking impressions from a newer, better one.

One newer check before you cut anything: See whether AI assistants are citing the page. Search Console won’t show you that, and a page with thin organic traffic can still be the source ChatGPT or Perplexity points people to.

Then ask the simplest question in the whole audit. If this page disappeared tomorrow, would a customer, prospect, or salesperson miss it? If the honest answer is no, polishing it won’t change much.

Every Audit Should End in One of Four Decisions

Here’s how I think about each call.

Update Pages with Untapped Potential

Pages ranked page-two, impressions without clicks, and relevant visitors who don’t convert. Those pages have a foundation worth building on. Refresh outdated stats, examples, and screenshots, get the answer near the top, fill in missing subtopics, fix the CTA, and link to related service pages.

A refresh isn’t changing the date and calling it done. Read the page as a skeptical customer would. What would send you back to Google? Fix that first.

Consolidate Pages That Compete with Each Other

When several pages answer the same question, merge them. Keep the URL with the best history, backlinks, and performance, fold in the useful material from the others, cut the repetition, and 301 the weaker URLs to it.

Clients get nervous here because the page count goes down. That’s kind of the point. One thorough guide usually beats four thin ones splitting the same signals.

Watch the intent, though. “How to choose an SEO agency” and “SEO agency pricing” serve the same buyer, but they answer different questions.

Remove What No Longer Has a Job

Discontinued products, expired events, thin tag pages, and articles on topics you have no business covering are all candidates. Low traffic by itself isn’t a reason. Check backlinks, conversions, and referral value first.

If you remove a page, redirect it only to a genuinely relevant replacement. If nothing fits, a clean 404 or 410 beats dumping visitors on your homepage.

Preserve What’s Working

Some pages rank well, match intent, and drive business. Leave them alone. Check accuracy, links, and technical health, but don’t rewrite them just to look busy. Instead, document why they work (structure, examples, internal links, and tone) and use that as the model for everything else.

When Net-New Content Earns a Spot

A new page makes sense when your audience has a real question your site can’t answer clearly. The gap might be a service, an industry, a location, a comparison, or a buying stage you’ve ignored.

The best ideas usually come from conversations. Sales calls, support tickets, and proposal objections tell you exactly where buyers get stuck. Keyword data is good for confirming demand. It’s a poor place to start.

Score each opportunity on three things: business impact, realistic search demand, and the effort it takes to make something genuinely useful. Low volume isn’t disqualifying. For an agency, “SEO for multi-location franchises” is worth more than a long-shot run at “SEO,” because the people searching it are the ones who sign contracts.

Then ask what you can say that the current top results can’t. Original data, client examples, real screenshots, and opinions you’re willing to defend give a page a reason to exist. If all you can do is summarize what already ranks, skip it.

One last filter before anything gets briefed. Finish this sentence: “This page helps [audience] do [task] by explaining [something distinct].” If it describes a page you already have, expand that page instead.

Prioritize by Impact

Turn the audit into a ranked list. Quick wins go first: fixing titles, repairing internal links, correcting facts, and merging obvious duplicates. Bigger projects, like new topic clusters, original research, or service page rewrites, get scheduled after.

A rough score of expected impact divided by effort keeps the team honest. It won’t be precise, and that’s fine. It’s there so the loudest stakeholder doesn’t set the roadmap. Give every item an owner and a date, and revisit the list monthly.

Measure What Changed, Not What You Shipped

Track qualified impressions, relevant clicks, rankings for priority topics, lead quality, sales, assisted conversions, and branded search. Watch for your pages showing up on long-tail queries you never targeted. That’s usually a sign your topical coverage is working.

For AI search, track how often your brand gets cited or mentioned around your priority topics. The specific tool matters less than having a baseline and checking it on a schedule.

These days, when a prospect asks me how many pages they need, I turn it around: What changed for your customers the last time you published something? If one rewritten page brings in qualified leads, it’s worth more than 30 new articles nobody reads. Build enough to cover what your buyers actually need to know. After that, put the effort into making those pages better.