Articles that rank are rarely accidents, and they are rarely the product of a secret trick. They are what you get when a site applies a repeatable system: choose topics you can genuinely own, structure them into clusters so the site itself demonstrates expertise, match each article to the intent behind its query, write introductions that earn the read, and keep old articles alive through refreshes. This pillar guide lays out that system end to end — the same system this site is built with.
- Ranking articles come from a system, not a trick: own a topic, structure it into clusters, match each article to its query's intent, write introductions that earn the read, and refresh old posts instead of always publishing new ones.
- The five steps reinforce each other — clusters demonstrate the expertise that thorough topical coverage claims, and refreshes compound it.
- A realistic first run is one complete cluster over about 90 days, judged on Search Console evidence rather than vibes.
- What the system deliberately leaves out matters too: no link schemes and no secret factors — only what Google documents and what your data shows.
The system at a glance
- Topical authority: cover one subject area thoroughly instead of scattering single posts across twenty subjects.
- Clusters: one pillar page per major subject, with focused spoke articles around it, all linked both ways.
- Intent matching: five minutes reading the SERP before outlining saves five weeks of ranking failure.
- Introductions: the first screenful decides whether the rest of your good work gets read at all.
- Refreshing: your back catalogue is an asset; updating proven pages often beats publishing new ones.
Step 1: Own a topic, not a keyword
A site with one article about roses, one about sourdough, and one about tax deductions tells Google nothing about what it is for. A site with thirty interlinked articles covering rose care from every practical angle makes a legible claim: this is a rose site, and its rose content should be taken seriously. That legibility is the practical core of "topical authority" — not a metric in any tool, but a property of your coverage: complete, organised, and deep within a defined area.
Practically, this means choosing a topic area narrow enough that a small site can plausibly become one of the best resources on it, then mapping its subtopics before writing. Our full explainer, with a mapping exercise you can do in an afternoon, is topical authority, explained practically.
Step 2: Structure with hub-and-spoke clusters
Topical coverage needs architecture, or it decays into a pile of posts. The architecture that works: for each major subject, one hub (pillar) page gives the broad, authoritative overview and acts as the index of the subject; around it, spoke articles each answer one narrow question in full. The hub links out to every spoke; every spoke links back to the hub and sideways to relevant siblings. Two-way links are not decoration — they are how readers and crawlers move through the cluster and how the hub accumulates the cluster's collective weight.
You are reading a cluster right now: this page is the hub, the five guides in the "In this guide" section below are its spokes, and each spoke links back here. The dedicated guide — including a map of this site's four real clusters as the worked example — is hub-and-spoke content clusters, explained with a live example.
Step 3: Match intent before writing
Every failed article that "should have ranked" deserves an intent check. Before outlining, read page one for your target query and note the dominant page type, format, and angle; then write that kind of page, better. Intent work also sharpens titles: a title that promises exactly what the SERP rewards gets clicked, and clicks are the one ranking-adjacent signal fully in your control at publish time. The routine is detailed in matching search intent before you write, and the failure mode is dissected in wrong search intent.
Step 4: Write introductions that hook
Your introduction has one job: convince the searcher, within a few lines, that this page will answer their question — and deliver a first instalment of the answer immediately. Bury the answer under three paragraphs of throat-clearing and readers bounce back to the results, an outcome no amount of later brilliance repairs. State the answer or the payoff early, name the specific problem the page solves, and set honest expectations about scope. Techniques, examples, and the common intro patterns that quietly kill good articles are in writing introductions that hook (and answer).
Step 5: Refresh instead of always publishing new
Publishing schedules worship new content, but portfolios rank on their whole catalogue. An old article with history, links, and impressions sitting at position 9 is usually one good refresh away from page one — a far better return than a brand-new page starting from zero. Refreshing means re-checking intent, filling coverage gaps, updating stale facts, and improving the introduction, then watching Search Console for movement. When to refresh, when to leave alone, and when to merge instead: refreshing old articles, with the operational version in our content refresh checklist.
Putting the steps in order
For a new site, the sequence is: pick the topic area → map the first cluster → write the hub → write spokes in order of search demand → run the pre-publish checklist on each → after a month, review Search Console and refresh the near-misses. For an existing site with a pile of unranked posts, start instead with diagnosis (why your articles aren't ranking), reorganise what you have into clusters, and only then write new pieces into the gaps.
None of these steps is glamorous. Together they are the difference between a blog that publishes and a site that compounds.
In this guide
Topical authority, explained practically
Topical authority, explained without buzzwords: what it really means, why complete coverage of one subject beats scattered posts, and how to map your topic area.
Read the guide
Hub-and-spoke content clusters, explained with a live example
Hub-and-spoke content clusters explained with a real, live example: this site's own four clusters mapped out — pillar pages, spokes, and the two-way linking rules.
Read the guide
Matching search intent before you write (a five-minute routine)
Match search intent before you write and your articles start life rankable. A five-minute SERP routine: read page types, formats, and angles before outlining.
Read the guide
Writing introductions that hook (and answer)
Article introductions that hook: give the answer early, name the reader's exact problem, and cut the throat-clearing. Patterns, examples, and what to avoid.
Read the guide
Refreshing old articles: the highest-return work on your site
Refreshing old articles often beats publishing new ones. Learn which posts to refresh (Search Console signals), what a real refresh includes, and when to merge.
Read the guideSources (5)
Google Search Central's guides to content and site structure (developers.google.com/search/docs), Google's people-first content guidance, and the cluster structure of this site itself, described in the clusters spoke.
- Google Search Central — Creating helpful, reliable, people-first content
- Google Search Central — SEO Starter Guide
- Google Search Central — Link best practices for Google
- Google Search Central Blog — Google Search's guidance about AI-generated content
- Google Search Central — Google Search Essentials
What this system deliberately leaves out
Notice what is not in the five steps: link building campaigns, keyword density targets, social media promotion schedules, and posting frequency quotas. That is deliberate. Off-page promotion has its place once a site has something worth promoting, but it is not a writing system, and for the small sites this guide addresses it is usually a distraction from the unglamorous truth that their coverage is patchy and their structure is flat. Likewise frequency: the system cares that what you publish lands inside a cluster and passes a checklist, not that you publish on Tuesdays. One well-placed spoke per fortnight compounds; daily unclustered posts evaporate.
How the steps reinforce each other
The steps are not a menu. Topical authority gives clusters something to organise. Clusters make intent matching easier, because each spoke has one narrow query to match rather than one sprawling topic. Intent-matched spokes give the refresh process raw material — when one underperforms, its siblings and the SERP tell you why. And introductions work harder inside clusters, because a reader who arrives mid-cluster and trusts the intro can be routed to three more relevant pages. Remove any one step and the others get weaker: that interdependence, more than any single tactic, is what "a system" means here, and why copying one visible element of a successful site (their headings, their word counts, their posting cadence) so rarely copies their results.
A realistic first 90 days
Days 1–14: choose the topic boundary and map cluster one (hub plus 5–8 spokes). Write the hub. Days 15–45: publish two spokes a week, each through the pre-publish checklist, links completed both ways. Days 46–75: finish cluster one; start reading Search Console patterns; write cluster two's hub. Days 76–90: first refresh pass on the earliest spokes using real impression data. By day 90 you own one complete, interlinked body of work on one subject — a small thing that outranks bigger, messier sites more often than any single brilliant article does.
The mindset underneath the system
Every step in this system assumes something unfashionable: that the searcher is the customer and Google is the distributor. Sites get this backwards — writing for the distributor (keyword placements, score chasing, publishing cadence theatre) while treating the searcher as an afterthought. The reversal shows in small decisions. A system-led writer asks "what would make this the obvious best answer to this question?" and lets rankings be the receipt. A distributor-led writer asks "what does the algorithm want?" and produces pages that read like they were written for a crawler, because they were. Google's own guidance has pointed the same direction for years — build for people, organise it well, and the machinery mostly follows. The five steps are simply that principle with the working parts exposed: authority is "be genuinely about something", clusters are "organise it so people can navigate", intent is "answer the question actually asked", introductions are "respect the reader's first ten seconds", and refreshing is "keep your promises current". Strip the SEO vocabulary and it is just good publishing — which is why it keeps working when tactics expire.
Proof discipline: what 'working' looks like
Because this system asks for months of trust, it should also tell you what evidence to collect along the way. Track three things per cluster in Search Console: total impressions (the cluster is being discovered), the number of distinct queries earning impressions (coverage is broadening), and the hub's position for its head query (the pillar is consolidating). If, after a full cluster has been live for a quarter, all three are flat, the honest conclusion is not "the system failed" but "one of the five diagnosis causes is active in this cluster" — and you now own a pillar guide for exactly that investigation. Systems earn belief by surviving their own diagnostics.
One cluster, completely
If this pillar persuades you of anything, let it be the modest version: build one cluster completely before starting a second. The discipline of finishing — hub polished, every spoke linked, refresh dates in the diary — is where the system's benefits actually live.
Finishing, it turns out, is also a ranking factor — the only one entirely within your control.
Frequently asked questions
How many articles do I need before rankings improve?
There is no threshold number. Improvement usually follows structure, not volume: one complete cluster of 6–10 interlinked pages routinely outperforms thirty scattered posts.
Should every article belong to a cluster?
Nearly all should. Occasional standalone pieces (news, opinion) are fine, but a site where most content is unclustered is leaving its main structural advantage unused.
Do I write the hub or the spokes first?
Draft the hub first so the cluster has a centre, but expect to expand it as spokes are finished. Some publishers publish two or three spokes first to start their indexing clocks; either order works if the links are completed at launch.
How long until this system shows results?
On an existing indexed site, refreshed pages can move within weeks; new clusters typically need 2–4 months to mature. The system compounds — month six looks nothing like month one.
Is this just for blogs?
No. Service businesses, shops, and documentation sites all use pillar-and-cluster structures. Anywhere one broad topic breaks into specific questions, the model applies.
Pick one topic you can genuinely own and build its first complete cluster before publishing anything scattered. The system does not need your whole site reorganised; it needs one cluster finished, measured, and then repeated.
Which topic could your site realistically own end to end — and what would its hub page be?