SEO & content marketing: frameworks & methodology

You have budget for 40 content briefs next quarter and a keyword export with 3,000 rows in it. The decision is not how many pieces to commission. It is which shape of page each row deserves, and which rows deserve nothing at all. Teams that skip that decision spend a quarter producing articles that rank for people who will never buy. What follows is the method for making the call: classify the query, pick the page shape it demands, then build at the scale the query volume justifies.

The three objects this method runs on

*Search intent* is the outcome someone wants from the results page, not the words they typed. The working classification has four classes: informational (learn something), navigational (reachreachThe number of unique people exposed to your message in a given period. Unlike impressions, reach counts each person once, no matter how often they see it.View full definition → a specific site), commercial investigation (compare options before choosing), and transactional (act now). Intent is observable rather than guessed. Read the current top ten results for the query, because Google has already decided which shape of page satisfies it. If page one is nothing but guides and comparison lists, your product page will not rank there however well it is written.

*A topic cluster* is a pillar page covering one broad topic at breadth, plus cluster pages that each go deep on a single subtopic, with every cluster page linking up to the pillar and the pillar linking down to all of them. The internal linking is the mechanism, not decoration: it concentrates authority on the page you want ranking for the head term and tells crawlers which page is your canonical answer for that topic.

*A programmatic page template* is one page design populated from a structured data set, so a single build produces hundreds or thousands of pages that share a shape and differ in content. "[App A] + [App B] integration". "[City] payroll rules". "[Job title] salary benchmark". The unit of work becomes the template and the data behind it rather than the individual page.

Those three give you the vocabulary. The method is deciding which one applies to a given query, in what order you build them, and how much of the budget each earns.

Step 1: mapmapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.View full definition → the query inventory to intent and page shape

Work the list in this order.

• Group the 3,000 rows into 30 to 60 topics. Rows that share a top-ten result set are the same topic, whatever the wording.
• Label intent per topic by reading the results page, not by pattern-matching on words like "best" or "how".
• Assign the shape. Informational topics get cluster pages under a pillar. Commercial investigation topics get comparison, alternatives and pricing-explainer pages. Transactional topics get product or solution pages. Competitor brand searches get an alternatives page, written so a sales rep would be comfortable sending it.
• Score commercial value before you look at volume.

The scoring arithmetic is simple enough to do in a spreadsheet column. Expected annual sessions equals monthly search volume, times the click-through rateclick-through rateClick-Through Rate (CTR) is the percentage of people who click a link, ad, or call to action out of those who viewed it.View full definition → for the position you can realistically hold, times twelve. Published click studies put position three at roughly a tenth of clicks and position eight in the low single digits, so use those as your bands. A 200-a-month comparison query held at position three yields about 240 sessions a year; at a 2% demo rate that is five opportunities from one page that you write once. A 40,000-a-month informational term where you would land eighth yields perhaps 14,000 sessions, and at a 0.05% demo rate roughly seven opportunities, from a 4,000-word page that needs refreshing every year and links to compete at all. The 200-a-month query is the better trade, and it is the one most content calendars skip.

Step 2: build one cluster completely before starting a second

Pick the product category where you can plausibly outrank what is already there. Build the pillar against the head term, then six to twelve cluster pages, one per distinct query group. Two rules keep it from collapsing: one query group per URL, and every cluster page links to the pillar with descriptive anchor text. When two of your own pages chase the same query, neither ranks well; consolidate them into one and redirect.

Feed the cluster from one research investment rather than twelve separate briefs. Semrush (which sells the keyword and rank-tracking data this whole method depends on, so treat its published guides as marketing for its own tooling) runs its annual State of Content MarketingContent MarketingA strategy of creating and distributing valuable content to attract, engage and retain a defined target audience, rather than pitching products directly.View full definition → survey and then extracts a dozen or more derivative pieces from the same data set, each aimed at a different query group and channel. One study, one cluster, one round of promotion. Semrush also bought Backlinko outright in 2022, which is the fastest version of the same move: acquiring a cluster that already ranks and already has links pointing at it.

Step 3: programmatic pages, and the quality floor they have to clear

Programmatic only works when four conditions hold: you have a data set that is yours or licensed, each page has a reason to exist beyond a swapped noun, the template surfaces something specific per page, and you control indexing.

Zapier is the clearest working example. Its app directory generates a page for every application it connects to and for app pairs, which runs into many thousands of URLs across a catalogue of several thousand apps. Those pages hold up because the data underneath is real and unique per page: the actual triggers and actions available for that combination, plus ready-made Zap templates the visitor can start from. Someone searching for a way to connect two specific tools gets an answer and a product action on the same page. The commercial logic is neat, too: the query is transactional and low volume, thousands of times over.

Index discipline is what separates this from a spam problem. Ship a few hundred pages first, watch impressionsimpressionsThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → and crawl behaviour in Search Console for four to six weeks, then release the rest. Anything that has accumulated no impressions after a quarter should be noindexed. Google's March 2024 spam policy update named scaled content abuse explicitly, and thin templated pages with a city name swapped in are exactly what it targets.

Step 4: put measurement in the brief, not in the quarterly review

Every brief should name the query group it targets, the position it needs, and the action the page asks for. Then track three layers: rankings, impressions and share of voiceshare of voiceYour brand's share of total advertising or conversation volume in your category, measured against competitors over a defined period.View full definition → at the top; content-assisted pipelinepipelineAll active sales opportunities across the stages of the sales process, together with their combined potential value and probability of closing.View full definition → and lead quality by entry page in the middle; closed revenue and cost per organic session versus paid at the bottom.

The horizons differ, which is why single-metric reporting fails. Ranking velocity is readable at 60 days. Engagement and conversion rateconversion rateThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.View full definition → per page are readable at 30. Link acquisition is a monthly number. Pipeline influence takes six to twelve months on most B2B sales cycles. Set a verdict date per page, usually 180 days, at which point the page is kept, refreshed, consolidated into a stronger sibling, or deleted. Without that date, libraries accumulate hundreds of pages nobody will ever fix.

Real-world cases

Zapier's growth engine has been a template plus a catalogue for most of its life. Adding one integration adds pages, and each page is a low-volume transactional query answered by a product that does the job immediately. The lesson for anyone with a directory, a marketplace or a taxonomy of jobs: your database is a content programme you have not built yet.

Semrush shows the other pattern, the cluster plus original data. Its guides own the head terms in its own category, the tool pages sit one click away, and the annual research gives the whole structure something to link to that competitors cannot copy.

CMO action items

• Score your existing library against the intent classes this quarter, and calculate what share of it targets commercial investigation queries. In most B2B programmes the answer is under 10%, while the majority of pipeline touches come from that layer.
• Rebuild one cluster end to end before commissioning anything new: one pillar, six to twelve cluster pages, cannibalisation cleaned up, internal links in place. Judge it at 90 days on rankings and at 180 on pipeline.
• Ask which data set you already own that could become a template. Integrations, locations, job families, product configurations, regulatory rules. If one exists, prototype 200 pages, not 20,000.

Common mistakes that kill results

Optimising for volume instead of intent match. A company selling enterprise HR software does not need 50,000 monthly visitors; it needs the 500 who are HR directors with budget. High-volume informational terms flatter the dashboard and starve the funnel.

Running search and content as two functions. When one team picks keywords and the other picks topics, you get two disconnected outputs and a library that ranks for nothing in particular. Keyword research has to produce the brief, and both teams sit in the same planning meeting.

Building programmatic pages before the template is worth reading. A template with one variable swapped and no unique data per page is the pattern search engines now demoted by name. Fix the data layer first; scale second.