We've watched clients publish three posts a week for a year and rank for nothing. Not because the posts were bad exactly — because none of them was better than what already ranked, and volume doesn't substitute for that.

Here's what we do instead. It produces fewer pages, and it works.

Why the volume approach fails

Google's job is to return the best available answer. If twelve pages already answer a question well, a thirteenth adequate page has no reason to rank. There's no credit for effort or frequency.

So the question for every piece isn't "is this good?" It's "is this better than the current page one for this query, in some specific way?" If you can't name the way, don't write it yet.

Ways to be better, in rough order of how achievable they are:

First-hand data or experience. Your own numbers, your own projects. Nobody can copy this.
Actually current. Half the top results for most technical queries reference deprecated tools or superseded rules.
More complete. Answering the follow-up questions the ranking pages leave hanging.
More specific to a segment. "SEO for law firms" beats competing with generic SEO guides.
Clearer. Underrated. Plenty of top-ranking pages are badly organised.

Get topics from sales calls, not tools

Keyword tools give you volume and difficulty. They don't tell you which questions block a purchase.

So we start by sitting in on sales calls or reading through enquiry threads. Every question a prospect asks before buying is a question people type into Google, and those queries convert because the person asking is already in the market.

For a law firm this produced: what does a shareholder dispute cost, how long does it take, what happens if there's an arbitration clause, can I force a buyout. Four pages, all of them commercially valuable, none of which would have surfaced from a keyword tool as a priority.

Then use the tools to check volume and see how the query is phrased. That order matters — sales insight first, validation second.

Read the results before you write

Search your target term and look at what's ranking. Google has already decided what kind of page that query wants.

If the top ten are buying guides, a product page won't rank there regardless of quality. If they're all listicles, a long essay is fighting the format. If they're forum threads and Reddit posts, the query wants a discussion and probably isn't worth a page.

This five-minute check saves the most common wasted effort in content marketing: writing the wrong kind of page for the right term.

Build clusters, not calendars

Individual posts scattered across unrelated topics don't accumulate anything. Clusters do.

A cluster is one substantial page on a broad topic plus five to ten pages on specific sub-questions, all linking to each other and to the parent. It works for two reasons: internal links concentrate authority on the parent page, and covering a subject thoroughly is itself a relevance signal.

A practical example. Parent page: technical SEO. Children: crawl budget, indexing problems, duplicate content, structured data, Core Web Vitals, log file analysis. Each child links up to the parent and sideways to two or three siblings. The parent links down to all of them.

Pick two clusters and finish them before starting a third. A half-built cluster performs like scattered posts.

Write it so it can be scanned and read

People arrive with a specific question and skim for the answer. Then, if you've earned it, they read.

What that means practically:

Answer the title question in the first two paragraphs. Don't warm up. If someone asks what something costs, the number goes near the top.
Descriptive subheadings. Someone should get the argument from the H2s alone.
Short paragraphs. Two to four sentences. Walls of text get abandoned on phones.
Specifics over adjectives. "Three weeks" beats "quickly". "$4,000 to $12,500" beats "affordable".
Vary sentence length. A long sentence carrying a full argument, then a short one. It reads like a person.

And a note on tone: content written to fill a slot reads like content written to fill a slot. Readers can tell, and increasingly so can search engines. If you have nothing specific to say about a topic, the right decision is not to publish about it.

Update before you publish something new

The highest-return content work is usually improving what exists.

Open Search Console, filter to queries where you rank between positions 5 and 15, and list the pages. Google already considers those pages relevant. A modest improvement moves them into the range where people click, and it's far cheaper than earning a new page's authority from zero.

What to do to each one:

Add the sub-questions the page doesn't cover — check "People also ask" for the query
Refresh anything factually stale, and update the modified date honestly
Improve the title tag for click-through, since you now know the exact query
Add internal links from related pages with descriptive anchors
Cut anything that pads without answering

We've seen a single afternoon of this move a page from position 11 to 4. No new content, no links.

Also: prune. Pages with no traffic, no links, and no purpose dilute the site. Merge them into something better or remove them. Fewer, stronger pages outperform more, weaker ones.

A cadence that's actually sustainable

For most businesses: two substantial pieces a month, plus one update to an existing page. That's it.

Two pieces a month is 24 a year. If each is genuinely better than what ranks, that's a serious content asset in eighteen months. Eight thin pieces a month is 96 a year of nothing.

For a professional services firm, the ratio is even steeper. Two well-researched pieces a quarter written by a partner beat twelve outsourced ones, because the expertise is the product and it can't be subcontracted convincingly.

Measure the right things

Not pageviews. Not time on page, which mostly measures how long people spend hunting for what they came for.

Per page: which queries it gets impressions for, average position on the queries you targeted, click-through rate, and whether it produces enquiries or assisted ones. Set up conversion attribution before you start, not six months in when someone asks whether the content is working.

Realistic timeline: a new page takes three to six months to reach a stable position. An updated existing page can move in two to four weeks. That difference is the whole argument for updating first.

The part nobody puts in a proposal

Good content is slow and it doesn't scale the way content marketing decks imply. A genuinely useful 1,500-word piece with original insight takes a day or two, including the research and the interview with whoever knows the subject.

Anyone offering twelve posts a month at a low per-post rate is producing something assembled from what's already ranking. That was viable for a while and it isn't now.

Our content writing work runs at the slower cadence deliberately, and we'd rather write six pieces that rank than thirty that fill a calendar.

Want us to do it