SEO Lab

SEO as the Base, AIO/GEO on Top

Clicks grew +86% in the 30 days after the update. The real driver: new pages that hit page 1 the day after publishing. We dissect the conditions for that initial velocity, with first-party data.

9 min read2026-07-21

Bottom line: new pages published with an SEO base (site structure, internal links, technical quality) plus an AIO/GEO overlay (up-front direct answers, structured data, E-E-A-T, llms.txt) reached page 1 within a day to a few days of publishing, in the post-May-2026-core-update environment. Pages with only the base, or only the overlay, did not reproduce that velocity. AIO/GEO and SEO are not an either/or — they are layers.

Observation

What Happened in the 30 Days After the Update — the Shape Matters More Than the Total

First, the timeline, per the official Google Search Status Dashboard records: the May 2026 core update began May 21 and completed June 2 (an 11-day, 21-hour rollout), followed by the June 2026 spam update on June 24–26. Our observation target is a developer-focused media site we operate (separate from this tool; the "observed site" below). We compare the 30 days before rollout start with the most recent 30 days after completion.

Metric (Google Search / GSC)Change over 30 days post-update
Clicks+86%
Impressions+122%
Average position~11 → ~10 (recently touching the 8s)
CTR5.2% → 4.3% (diluted by new impressions)

Looking only at totals, it's tempting to say 'the core update lifted us.' But daily data tells a less simple story: the real step-ups in clicks appeared not right after rollout completion (June 2) but progressively from about two weeks later — and those steps coincided with the dates the observed site published new pages, which the algorithm alone can't explain.

What This Article Dissects

We split the +86% into (1) recovery of existing pages and (2) launch velocity of new pages. Spoiler: the biggest finding isn't the total — it's that new-page velocity split sharply by when and how each page was built.

Verification 1

New-Page Velocity Split Cleanly into Four Groups

Across roughly two and a half months spanning the update, the observed site kept publishing both tool pages (single-purpose, in-browser tools) and article pages. We grouped them by publish timing and construction, then compared time-to-page-1 and CTR once on page 1.

GroupPublishedReached page 1Early CTR
Tool ADays after rollout completedSettled at #5–8 within days15〜30%
Tool B~2 weeks after completion#3–5 the day after publishing17%+ (peaking at 50%)
Tool C~3 weeks after completion~#6 almost immediately11〜16%
Tools D/E/F (control)Before rollout startNo page-1 presence in the 30-day window
Article G (new topic)~1 week after completionPage-1 range in ~a week3〜5%
Articles H/I (added to existing cluster)~6 weeks after completionClicks from day one (#6–7)7〜11%

Every post-update tool (A, B, C) reached page 1 within a day to a few days, then stuck and kept growing — Tool A's clicks still haven't plateaued weeks in. Meanwhile, tools built much the same way but published before the update (D, E, F) still hadn't established page-1 presence in the same 30-day window.

Verification 2

The Velocity Gap Wasn't About Position — It Was CTR

Here's the easy-to-miss part: Article G also entered page-1 range in about a week — time-to-rank was similar to the tools. The gap opened after that. On the same page 1, the tools drew 15–30% CTR while Article G drew 3–5% — roughly a 3x difference. From the moment they appeared in results, tools got chosen; the article got skipped.

Our interpretation: for do-intent queries like '[x] generator' or '[x] tool,' searchers want something to use now, not something to read. A tool's title matches that intent literally, so its click rate is structurally higher at the same position.

The Exception That Proves the Rule

Articles H and I got clicks from day one — but both were additions to an existing topic cluster built by the observed site's strongest pages, borrowing powerful internal links and accumulated topical authority (the 'base') from the start. The gap versus stand-alone Article G is exactly the presence or absence of that base. Articles need a base first to launch fast; tools can launch alone on intent-matched CTR. That, we believe, is the structural difference.

Breakdown

A 2×2 of Base × Overlay Reveals the Velocity Conditions

By base we mean SEO fundamentals — site structure, internal links and topic clusters, technical quality (speed, mobile, crawlability). By overlay we mean AIO/GEO-oriented quality — up-front direct answers, structured data, FAQs, author info (E-E-A-T), llms.txt. The post-update tools were born with both. Mapping the observations onto four quadrants:

Base ✓ × Overlay ✓

Tools A, B, C (post-update)

Page 1 in a day to a few days, 15–30% CTR, still climbing. The fastest launches we observed.

Base ✓ × Overlay weak

Articles H, I (existing cluster)

Clicks from day one, but CTR in the single digits to ~11%. They launch on borrowed base, but the growth slope trails the tools.

Base weak × Overlay ✓

Article G (stand-alone topic)

Ranks within ~a week, but 3–5% CTR and a sluggish click ramp. Overlay alone doesn't create a reason to be chosen.

Published pre-update (control)

Tools D, E, F

Similar construction, yet no velocity. Suggests the environment — re-evaluation timing — plays a role in launch speed.

In its core update guidance (updated December 2025), Google notes that major recovery for existing pages may require "waiting until the next core update." Existing pages are re-evaluated on the update cycle; new pages are assessed from zero in whatever environment exists at publish time. Shipping pages with both base and overlay into a freshly-updated environment — that combination, per our observation, gets rewarded fastest.

Byproduct

AI Referrals Grew Too — but That Wasn't the Main Battlefield

The overlay elements were originally designed for AIO/GEO — getting cited by AI search. So how did the AI side do? Referral sessions from ChatGPT, Gemini, Perplexity, Copilot, etc., measured in GA4, grew about 1.9x (+87%) versus April. We've also directly confirmed ChatGPT presenting the observed site's tool pages in its answers.

An honest note on scale, though: while absolute volume grew about 1.9x, AI referral sessions have stayed within 1–3% of the observed site's total traffic. And with the Google-side denominator growing +86%, some months the share actually got thinner. In other words, AI traffic is not what drove the +86%. For now, the biggest return on AIO/GEO work shows up not as AI traffic but as the same quality signals being rewarded in Google's re-evaluation — that's the honest state of play our data supports.

This is consistent with the position we took in "Don't Be Fooled by AI Search Numbers": AI-search growth rates look flashy, but they shouldn't take center stage without checking the denominators and the causality.

Limits

What This Observation Proves — and What It Doesn't

What it shows

  • Only post-update pages with base + overlay reached page 1 within days (observed fact)
  • The velocity gap showed up in CTR (~3x), not time-to-rank
  • Articles needed an existing cluster (base) to launch fast

What it doesn't

  • A causal claim that the core update changed launch velocity. This is correlation; the timing overlap could be coincidence
  • Elimination of demand-side confounds — some post-update tools rode topics with growing search demand. (Though demand can't explain ranking #3–5 the day after publishing)
  • Generalization. One site, a handful of pages; reproduction on other sites and niches is unverified

So this article isn't selling a 'how to win core updates' recipe. The one thing we can say: shipping every page with both base and overlay was what got rewarded fastest — at least in the observed site's environment.

Reproducibility

Pre-Publish Checklist — 4 Base Items, 4 Overlay Items

Base (SEO fundamentals)

  • 1. Internal links to/from the existing cluster
  • 2. A title matching intent (for do-intent, words that signal 'usable now')
  • 3. Technical quality (speed, mobile, crawlable HTML)
  • 4. Sitemap inclusion and index submission

Overlay (AIO/GEO quality)

  • 1. Up-front direct answer (what the page answers, in the first viewport)
  • 2. Structured data (Article/FAQPage/HowTo as fits the page)
  • 3. Visible author info and dates (E-E-A-T)
  • 4. llms.txt listing and AI-parseable structure

Implementation steps for the four overlay items are in the AIO Optimization Guide, the full checklist rationale in AIO Check Items Explained, and our measured CTR impact of structured data in the Structured Data Impact Report.

Direbase's Position

This data sides neither with 'SEO is dead, AIO/GEO is the future' nor with 'AIO is a buzzword — it's just SEO as usual.' What we observed is a layered structure: without the SEO base, neither rankings nor clicks get off the ground; without the AIO/GEO overlay, you don't get chosen in the results. Only pages born with both worked on page 1 from the day after publishing.

Direbase audits technical foundation, structured data, and E-E-A-T signals in a single diagnosis — as the verification tool between GA4 and GSC — precisely so these layers don't get split into separate projects. Base and overlay aren't two initiatives; they're one publish-time quality bar.

For the reading of the update's official statements themselves — what "no specific recovery action" and "a drop isn't necessarily a problem" mean — see part one: Google May 2026 Core Update — Official Statement vs Measured Data. This article is the 30-days-later follow-up.

Check Base and Overlay in a Single Audit

Enter a URL for a 45-item audit of technical foundation, internal links, structured data, and E-E-A-T signals — and confirm your next page ships with both layers.

今井政和

Written by

今井政和

SEO Director / Frontend Developer

SEO Director with 20+ years of web industry experience. Creator of Direbase and the official WordPress plugin "ORECTIC SEO CHECK." Author of a book on web strategy inspired by Edo-era merchant principles.

@imai_director

FAQ

Is it advantageous to publish new pages after a core update?
In our observation, only new pages published after the rollout completed reached page 1 within a day to a few days, while a control group published before it failed to launch under similar conditions. This is a single-site correlation, not causal proof. Practically, rather than timing publishes around updates, we recommend making sure every page ships with both the base (internal links, technical quality) and the overlay (direct answers, structured data, E-E-A-T) — publish timing is one variable you barely control; page quality always works.
Does AIO/GEO work also help Google rankings?
The elements you implement for AIO/GEO — up-front direct answers, structured data, FAQs, author information (E-E-A-T), llms.txt — largely coincide with the quality signals Google uses. In our data, new pages born with these launched fastest in the post-update environment, and AI referral sessions also grew about 1.9x. There's no need to run 'AIO projects' and 'SEO projects' separately; they're two faces of the same quality bar.
What should I include to maximize a new page's launch velocity?
As a pre-publish checklist: four base items — (1) two-way internal links with an existing cluster, (2) a title matching search intent, (3) technical quality (speed, mobile, crawlable HTML), (4) sitemap inclusion and index submission — and four overlay items — (1) an up-front direct answer, (2) structured data (Article/FAQPage etc.), (3) visible author info and dates, (4) llms.txt listing. In our observation, only pages with both layers reached page 1 within days; either layer alone failed on ranking or CTR.
Should I build tool pages or article pages?
They play different roles, so it's not either/or. Tools matched do-intent queries literally in their titles and launched stand-alone at 15–30% CTR from day one. Stand-alone articles ranked but stalled at single-digit CTR, while articles added to an existing topic cluster got clicks from their first day. Tools are assets that launch alone; articles are assets that work on top of a cluster base. Establishing a beachhead with a tool, then thickening the cluster with articles, is the sequence consistent with our data.
If a page launches slowly, should I rewrite it right away?
We advise against immediate rewrites. Google's official core update guidance recommends avoiding changes to content that's already performing well and notes that major re-evaluation may wait until the next update cycle. First diagnose with GSC's position × CTR: is it an intent/title mismatch (ranks but low CTR), a missing base (no internal links or cluster), or a missing overlay (no direct answer or structured data)? Fixing only the identified gap is what aligns with both our observations and the official guidance.