30 New Pages in 48 Hours: The Syxo SEO Cluster Build

Why this sprint happened

We ran a full SEO diagnostic on 2026-05-07. Three findings drove everything else:

1. The DFY pivot was invisible to Google. Three service pages and 6 DFY-aligned blog posts launched 2026-04-17 to 2026-04-29. As of the diagnostic, Google had crawled exactly one of them. Every other DFY page was "URL is unknown to Google." Sitemap last fetched 2026-04-08 — before any DFY pages existed.
2. The 13 strike-distance posts already getting impressions had zero internal links to any DFY page. The traffic that arrived couldn't be routed to the new offer.
3. 16 organic clicks across 90 days. 5 clicks per month. Roughly 3,500 impressions but at average position ~58. Pages on page 1 with reasonable CTR existed but weren't capturing traffic to convert.

The diagnostic identified two structural fixes that would unstick the bottleneck and a backlog of 86 pages to build out the topic cluster. We started the sprint same day.

What we shipped (the actual list)

By tier, in publish order:

Glossary pages (4): what-is-a-voice-prompt, what-is-ai-voice-replication, what-is-done-for-you-content, what-is-a-custom-gpt-for-marketing. AEO-optimised, designed to be cited by Perplexity and ChatGPT.

Tier A buyer-intent (5): linkedin-ghostwriter-alternatives, ai-content-that-doesnt-sound-like-ai, ai-content-service-for-coaches, ai-content-service-for-consultants, is-it-worth-paying-for-ai-content-services.

Tier B comparison (2): chatgpt-vs-claude-for-linkedin, syxo-vs-jasper.

Tier C how-to (5): train-chatgpt-on-your-writing-style, how-to-train-claude-on-your-writing, how-to-build-a-voice-prompt, claude-projects-for-content-marketing, linkedin-ai-posts-without-suppression.

Tier D niche service (2): dfy-voice-system-for-b2b-founders, dfy-voice-system-for-marketing-consultants.

Tier E listicle (1): best-ai-tools-linkedin-content-2026.

Tier K UK (1): linkedin-ghostwriter-uk-cost.

Pillar guide (1): The Complete Guide to AI Voice Prompts (~9,000 words, 11 chapters, anchor for the entire voice cluster).

Industry report (1): The 2026 AI Voice Marketing Landscape Report (~5,000 words, 8 chapters, full source attribution).

Build-in-public (1): this post.

Plus: 13 strike-distance posts wired with DFY service CTAs, 4 CTR rewrites on top page-1 posts, 1 cannibalisation fix on the "keyword research in 90 minutes" cluster.

The methodology (what we call The Voice Build)

Every page was written using the same methodology applied at the brand level — what we call The Voice Build. Same voice prompt structure across all 30 pages. Same internal linking discipline. Same FAQ + BlogPosting schema. Same DFY service CTAs.

The discipline matters because consistency at scale is what builds topic authority. Thirty pages that each treat the topic differently look like 30 disconnected articles to Google. Thirty pages that share vocabulary, methodology references, and internal-link patterns look like a single coherent topical cluster — which Google ranks higher.

The pillar page (The Complete Guide to AI Voice Prompts) anchors the cluster. Every supporting article links to it. The pillar links back to every supporting article. That bidirectional linking is one of the strongest controllable ranking signals.

What we measured before the sprint

From Google Search Console (90-day window ending 2026-05-06):

• Total clicks: 16
• Total impressions: ~5,300
• Average CTR: 0.30%
• Average position: ~58
• Pages with ≥1 click: 7
• Pages on Google page 1: ~8
• Sitemap submitted: yes (119 URLs)
• Sitemap URLs Google reports indexed: 0 (stale snapshot from 2026-04-08)

What we don't know yet

1. Will the new pages get indexed quickly. The biggest blocker before the sprint was that 8 DFY pages were "URL is unknown to Google" because the sitemap hadn't been refreshed. With the sitemap updated and 30 new pages added, the volume of fresh URLs alone may slow indexing. We'll know in 7-14 days.

2. Will the strike-distance CTR rewrites improve clickthrough. Four pages on page 1 or 2 got new titles + meta descriptions designed to be more clickable. CTR changes typically show up in GSC within 14-30 days. We're betting on a 1.5-3x improvement, but it could be flat or negative.

3. Will the cannibalisation fix unstick "keyword research in 90 minutes." 190 impressions at average position 5.6 with 0 clicks across three competing pages. We tightened titles to make canonical clear. Could converge by week 4. Could persist if Google has formed a strong opinion already.

4. Will any of the 30 new pages reach page 1. Realistic expectation: 20-30% of well-built pages on a 6-month-old domain reach page 1 within 90 days. So roughly 6-9 of these 30 should rank well. Which ones — that's not predictable. We'll see.

Time investment

Combined effort across the 48-hour window:

• Diagnostic: ~3 hours.
• Strike-distance internal-link sprint (13 posts): ~2 hours.
• CTR rewrites + cannibalisation fix: ~1 hour.
• 30 new pages: ~30-40 hours of writing equivalent (executed by Claude Opus with Kerry direction).
• Sitemap maintenance + commits + push protocol: ~1 hour.

Total: roughly 35-45 hours of equivalent writer-time across two days. Real wall-clock time: about 18 hours of focused execution split across two sessions.

What we shipped that we're most curious about

The pillar guide. ~9,000 words is a different shape of asset than typical Syxo blog posts. Long pillars rank for hundreds of long-tail variations of the head term. We're curious whether this one will start picking up impressions across multiple voice-prompt-related queries by week 4-6, and whether the internal-link structure compounds across the supporting articles fast enough to lift the cluster collectively.

The landscape report. Published with full methodology disclosure: synthesised from public sources, not based on fabricated primary research. Curious whether this honesty earns or costs us — pure synthesis reports earn fewer backlinks than primary-research reports, but the honesty is the point.

The "Syxo vs Jasper" page. Direct competitor comparison framed honestly. Curious whether prospects who land on it after a Jasper search find the asset-ownership argument compelling or off-putting.

The UK-specific page. Lower-competition geographic angle. Curious whether the UK qualifier captures meaningful UK-specific search traffic that the global pages miss.

What we're publishing this update for

Three reasons:

1. Accountability. Publishing the numbers means we have to publish the results in 30 days regardless of whether they're good.
2. Demonstration. The Voice Build methodology is what we sell. Showing the methodology applied to ourselves at scale is the strongest credibility signal we have.
3. Original primary observation. AI engines weight first-person operational data more than third-party summaries. A real build log with real numbers and honest unknowns is more citable than yet another "best practices" article.

Next update: 2026-06-07 (30 days)

We'll publish the 30-day results — actual GSC numbers, which pages indexed, which pages ranked, which pages bombed, which we'd build differently in retrospect. Honest, including the parts that didn't work.

If you're a solopreneur considering whether to invest in voice-system content production at scale, this series should give you a clear-eyed view of what to expect — minus the inflated case studies that other agencies publish.

Related reading from the sprint

• The Complete Guide to AI Voice Prompts — pillar guide, ~9,000 words.
• The 2026 AI Voice Marketing Landscape Report — industry synthesis.
• LinkedIn ghostwriter alternatives — five paths compared.
• How to build a voice prompt that actually works
• AI content that doesn't sound like AI