The setup

Zero budget. A free GitHub Pages subdomain. Three articles a week, every one researched and refined with AI. No ads, no paid tools, no existing audience to lean on.

The bet was the same one behind every "scale content with AI" pitch. Publish enough useful articles and search traffic follows. So we published — fast.

By month four the archive held 50 articles. Then we opened the workflow that got them live and checked Search Console for how many Google had actually indexed.

📊 The result

50 articles published. 9 indexed by Google. Roughly 82% of everything we'd written was invisible in search — and 43 of those pages had never even been crawled.

The number that reframed everything: 9

Nine out of fifty. But the surprising part wasn't the count — it was the reason. Those 43 missing posts weren't rejected by Google. They were never seen.

Google hadn't judged them thin or low quality. It had never crawled them at all. Barely any search traffic reached anything beyond the homepage, because there was almost nothing for search to send people to.

Discovery vs indexing — the distinction most people miss

Here's the lesson that changed how we work. There are two separate walls, and they get confused constantly.

Indexing is Google crawling a page and deciding whether to store it. Discovery is Google learning the page exists in the first place. You can't be judged on quality if you're never found.

We'd spent months optimising for indexing — cleaner titles, FAQ sections, structured data. None of it reached 43 pages, because they were stuck one step earlier. Invisible, not rejected.

Why it happened

We checked the obvious technical causes first. The links were plain static HTML, not JavaScript. Robots.txt allowed everything. Nothing carried a noindex tag.

So the cause was structural. A young domain with no authority earns a tiny crawl budget. Google fetched the homepage and the blog index, saw fifty near-identical links, and decided the rest weren't worth the queue.

Publishing faster made it worse. Every new post added another undiscovered URL to a pile Google had already declined to work through. This lined up with the data on whether AI content ranks — quality was never the blocker here.

What actually moved the needle

Internal links from pages Google already trusts. The handful of indexed articles became doorways. A contextual link from a crawled page carries a buried post into Google's path far better than a card in a list of fifty.

Manual URL Inspection requests. On a free host, this is the one reliable way to say "please look at this specific page." Slow and capped daily — but it works.

Not the sitemap. On free GitHub Pages the sitemap gets served in a format Search Console can't fetch. That door is closed, so we stopped pushing on it.

The honest takeaway

Publishing volume is a vanity input. The number goes up and it feels like progress. But a post nobody can find is worth zero, however good it is.

If you're building on a new domain, order matters. Get discovered first. Earn a little authority. Then scale output into a system that can actually surface it.

We'd trade thirty of those fifty articles for the discovery work we skipped. That's the real result — and the reason the next four months look nothing like the first.

Frequently Asked Questions

Does Google penalise AI-written content?
That's not what we saw. The 43 missing pages weren't penalised — they were never crawled. Discovery failed before quality was ever assessed. AI content that gets discovered and is genuinely useful can index and rank fine.
Why would Google not crawl published pages?
Crawl budget. A new domain with little authority gets a small allocation. Google crawled the homepage and blog index, then decided fifty near-identical links weren't worth queuing. The pages existed and were reachable — they just sat below the crawl threshold.
How do I check my own indexing?
In Search Console, open the Pages report under Indexing. Compare the indexed count to how many pages you've published. A wide gap means discovery, not quality, is your bottleneck.
What fixes a discovery problem on a free host?
Two levers: contextual internal links from pages Google already crawls, and manual URL Inspection requests for priority pages. Publishing less but linking more beats publishing more into a pile Google won't crawl.