Every new content site faces the same fork: cover one topic deeply, or cover many topics shallowly. The honest answer is not “balance” — it depends on which stage your site is in, and the 2026 ranking environment has made the stages much less forgiving. Below is a stage-based decision rule, a content-plan template you can paste straight into your repo, and the exact Search Console query that tells you when to flip from depth to breadth.
TL;DR
- For a brand-new site, go deep first inside one hub until you reach roughly 25-30 substantive articles, then open a second hub. Avoid spreading thin across 5 hubs in the first 90 days.
- This is not a style preference anymore. Google’s March 2026 Core Update (rolled out March 27 - April 8, 2026) reinforced E-E-A-T and hit sites hardest when they relied on templated, keyword-swapped, or AI-generated pages published without expert review across many shallow topics.
- A cluster of ~20 interconnected articles on one subject typically outranks a single 5,000-word guide on that same subject. Depth compounds; breadth without depth dilutes.
- Use Search Console impressions and average position — not gut feel — to decide when hub 1 is “deep enough” and which hub 2 deserves the spend.
Why order matters more in 2026
In the first 60 days a search engine has almost no signal about what your site is about. Go too broad and you create topical confusion across half-built hubs; go too narrow and there is nothing to crawl. The pattern that wins is: depth first inside one hub, then breadth across hubs, then depth again to fill gaps. Most indies do this in the wrong order because the right order feels slow.
What changed in 2026 is the penalty for getting it wrong. After the March 2026 Core Update, sites that published comprehensively within a single subject — covering every angle and sub-topic — outperformed broad sites that touched many topics shallowly. Google’s own guidance frames this as topical coherence: a domain that covers one subject thoroughly enough to be treated as a reference point earns a compounding advantage, where new articles rank faster and pull up adjacent pages on the same domain. The same update applied the steepest visibility drops to keyword-swapped templates and AI-generated content shipped without a named human review. Breadth-first is now the high-risk play.
(One related 2026 change worth knowing: Google retired FAQ rich results in May 2026. You can no longer shortcut your way to a richer SERP listing with FAQ schema — rankings now come from genuine depth and authority, which is exactly what depth-first builds.)
The stage-based decision rule
Find your site below and follow the matching move:
| Your situation | Move |
|---|---|
| < 60 days old, < 30 articles total | Go deep in ONE hub. Ignore every other idea. |
| One hub with 30+ articles, nothing else | Open a second hub for breadth. |
| 3+ hubs, each with only ~10 articles | Stop. Go deep on the hub with the best early traffic. |
| Search Console impressions concentrated on one hub | Double down on that hub before opening new ones. |
| Average position in the lead hub trending below 30 | Depth is paying off — keep going. |
Before you start
- Pick the anchor hub honestly — strongest by your knowledge AND real search demand, not just personal enthusiasm.
- Have a content-collection schema that enforces single-hub assignment, so a stray article can’t quietly start a second hub.
- Commit to using Search Console data, not vibes, to decide when to open hub 2.
Step by step
- Pick a single anchor hub. Score candidates on a 1-5 scale for knowledge and demand, minus competition. Highest score wins:
# scripts/hub-score.yml
candidates:
- id: indie-dev
knowledge: 5
demand: 4
competition: 3 # lower = better
score: 6 # knowledge + demand - competition
- id: ai-tools
knowledge: 3
demand: 5
competition: 5
score: 3
- id: prompt-library
knowledge: 4
demand: 4
competition: 4
score: 4
# pick the highest score- Plan 25-30 long-tail articles inside the anchor. Keep a short YAML plan in the repo so the cluster is a checklist, not a vibe:
# content-plan/indie-dev.yml
hub: indie-dev
pillar: "How to launch an indie content site in 2026"
cluster:
- astro-deploy-firebase
- firebase-hosting-go-live-checklist
- firebase-custom-domain
- firebase-cache-and-deploy-update
- firebase-route-404-causes
- what-is-firebase-hosting
- firebase-hosting-free-tier
- is-vercel-good-for-content-sites
- vercel-deploy-astro
- vercel-custom-domain
# ... 15-20 more
cadence: 3/week
target_complete: 2026-07-15-
Set a publishing pace and hold it for 8 weeks. Two to three articles per week is sustainable for a solo builder. Do not open a second hub until hub 1 has ~25 articles.
-
Wire a hub-progress guard into prebuild so you can’t accidentally drift wide:
// scripts/hub-progress.mjs
import { readdirSync, readFileSync } from 'node:fs';
import matter from 'gray-matter';
const counts = {};
for (const f of readdirSync('src/content/articles/en/indie-dev')) {
const { data } = matter(readFileSync(`src/content/articles/en/indie-dev/${f}`, 'utf8'));
counts[data.category] = (counts[data.category] || 0) + 1;
}
const lead = Math.max(...Object.values(counts));
const anyOther = Object.values(counts).filter(c => c !== lead).some(c => c > 10);
if (lead < 25 && anyOther) {
console.warn('WARN: opened a second hub before anchor hit 25');
}- Choose hub 2 from real demand signals, not vibes. Pull the queries Search Console is already showing for hub 1 (the API is unsampled and returns up to 25,000 rows per call, so it sees far more than the UI):
curl -X POST "https://www.googleapis.com/webmasters/v3/sites/$SITE/searchAnalytics/query" \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
--data '{
"startDate":"2026-03-22","endDate":"2026-05-22",
"dimensions":["query"],"rowLimit":200
}' \
| jq -r '.rows[].keys[0]' | head -40Look for repeated nouns that don’t fit hub 1 — those are your hub 2 candidates. You are letting Google tell you where demand already exists instead of guessing.
-
Alternate publishing 60/40 in favor of hub 1 for the next 8 weeks, until hub 2 also reaches ~25 articles. Do not let the exciting new hub starve the one that already has traction.
-
At ~100 articles, audit hub traffic and double down on the winner. Join your prebuild counts to a 28-day Search Console pull:
hub articles impressions(28d) clicks(28d) CTR
indie-dev 34 42,000 1,260 3.0%
ai-tools 28 11,200 180 1.6%
prompt-library 22 4,300 62 1.4%
# → indie-dev wins; allocate the next 20 articles there- Only after 150-200 total articles consider opening hubs 3 and 4 in parallel. By then each established hub is self-sustaining and a new one won’t strand the others.
What “pillar and cluster” should actually look like
Depth-first only pays off if the cluster is wired into a pillar. The structure that builds topical authority in 2026 is the classic hub-and-spoke:
| Page type | Length / count target | Job |
|---|---|---|
| Pillar page | ~3,000-5,000 words | Cover the whole topic at a high level; link OUT to every cluster page |
| Cluster pages | 10-20+ per pillar, each genuinely useful | Answer one long-tail query well; link BACK to the pillar |
A 300-word subtopic stub does not build authority. When clusters link to the pillar and the pillar links to the clusters, the crawler can map the structure and read which subject you specialize in. Done correctly, sites implementing content clusters report roughly a 40% average lift in organic traffic, with stronger cases hitting 50-300% over 6-12 months.
Implementation checklist
- Anchor hub chosen via written scoring, not vibes.
- Content-plan YAML lists at least 25 cluster slugs for the anchor.
- Prebuild warns if a second hub opens before the anchor hits 25.
- A real pillar page links to every cluster; every cluster links back.
- Search Console data — not enthusiasm — drives the hub-2 decision.
After-launch verification
- After 8 weeks, the anchor hub has 20+ indexed articles.
- Average position for cluster queries trends from above 50 toward 20.
- Search Console impressions concentrate (over 70%) on the anchor hub — that is what depth looks like in the data.
Common pitfalls
- Opening 5 hubs in month 1 because you have ideas in many directions. You end up with topical authority nowhere, which is exactly the profile the March 2026 update demoted.
- Going so deep that you cover sub-sub-sub topics no one searches for. Cap depth at “real queries from Search Console plus adjacent autocomplete”.
- Abandoning hub 1 the moment hub 2 feels exciting. Half-built hubs do not rank.
- Calling a hub “deep” with 8 articles. In 2026, depth means at least 20-30 useful articles tied to a pillar.
- Treating hub-2 selection as a creative decision when it is a data decision Search Console can make for you.
- Faking breadth with unreviewed AI output. That is the precise pattern that took the steepest visibility hit this year.
FAQ
- What if I have multiple hubs I’m equally excited about?: Pick the one with the strongest combination of search demand and your own knowledge edge. The others can wait 90 days without losing anything — and you’ll choose better once hub 1’s Search Console data is in.
- How do I know when a hub is “deep enough”?: When you have 25-30 articles covering the main long-tail variations plus at least one strong pillar (aim for 3,000-5,000 words) tying them together with internal links.
- Does Google really reward topical authority?: Yes, and more so since the March 2026 Core Update. Sites that cover one subject thoroughly get a compounding edge: new articles rank faster and pull up adjacent pages on the same domain.
- Can I use AI to fake breadth?: Not reliably. The March 2026 update applied the steepest visibility drops to AI-generated and templated content shipped without expert review. Depth that survives needs fact-checked, human-edited, structured input.
- Should the pillar article come first or last?: Last. Write the cluster articles first; the pillar emerges from what you actually learned. Drafting the pillar first leads to abstract, low-utility writing.
- How long until a new hub starts ranking?: Plan for 8-12 weeks of consistent publishing before cluster queries climb from above position 50 toward 20. Indexing of individual pages is faster, but authority — the thing that lifts the whole hub — accrues over weeks, not days.
Related
- How to design content-site sections so they scale
- Planning a long-tail keyword site from day one
- How to pick a niche that has search demand
- Judge search demand before building
- Set publishing cadence
External reference: Google’s own Creating helpful, reliable, people-first content guidance is the canonical source on the E-E-A-T signals that depth-first building is designed to satisfy.
Tags: #Indie dev #Website planning #SEO #Pillar / Cluster #Content ops