Used AI to write 450 articles in 3 months. Got to 50K monthly visits. Then Google killed my site overnight.

🎨 Content & Creativeby ai_content_farm · 3w ago

▸ What I did

Built a niche content site about home improvement tips. Used Claude and ChatGPT to generate 5 articles per day — 1,500 words each, targeting long-tail keywords. Did minimal editing: fixed obvious errors, added a few images, published. 450 articles in 90 days. Set up AdSense for monetization.

▸ What I expected

The articles were "good enough." Not great writing but factually accurate and keyword-optimized. I figured even mediocre content at scale would capture enough long-tail traffic to make AdSense revenue worthwhile. Target: $1,000/month passive income.

▸ What actually happened

It actually worked at first. Month 1: 5K visits. Month 2: 22K visits. Month 3: 51K visits. AdSense revenue peaked at $380/month. I was planning to scale to 1,000 articles. Then a Google core update hit. Traffic dropped from 51K to 4,800 in a single week. AdSense revenue: $31. My site got hit by what I'm pretty sure is the "helpful content" classifier. 90% of my pages disappeared from Google's index entirely. 3 months of "work" effectively deleted overnight.

▸ What I've tried so far

Tried noindexing the lowest-quality pages. Tried adding "human touches" — personal anecdotes, opinion sections. Tried submitting a reconsideration request (Google doesn't actually have one for algorithmic penalties). Considered manually rewriting all 450 articles but at 2 hours each that's 900 hours of work. Currently deciding between: a) abandon the site and start fresh, b) nuke 400 articles and try to save 50 good ones, or c) give up on SEO-based content entirely.

Progress Updates (0)

0 / 3000

3 Replies

seo_slow_burn · 3w ago✓ the fix

This is the exact opposite of my story. I wrote 65 articles in 8 months by hand and hit 10K/month. You wrote 450 articles in 3 months with AI and got nuked back to 5K. The math is clear: 65 genuine articles > 450 AI articles in the long run. Google's helpful content system isn't just checking individual articles — it evaluates your ENTIRE site. One bad section drags down everything. Option B is your best bet: nuke 400, keep 50, and manually improve those 50 until they're genuinely the best content on the internet for those keywords.

algo_survivor · 3w ago

I've seen a dozen AI content sites go through this exact cycle. The pattern is always: rapid growth → core update → cliff. The ones that recover do ONE thing: they become a REAL site. Real author with a real name and real expertise. Real original research or data. Real community engagement (comments, newsletter). Google doesn't care if AI wrote the draft — they care if the site has genuine expertise and authority. An anonymous content farm with 450 articles from nobody has neither.

content_grinder · 2w ago

Unpopular opinion: AI content CAN work, but not the way you did it. 5 articles/day with minimal editing is not "using AI as a tool" — it's running a content factory. The people succeeding with AI-assisted content are using it to write 5 articles/WEEK with heavy human editing, original research, and personal expertise layered on top. AI handles the draft, human handles the value. You skipped the value part.

Status

✗ stuck

Stats

Me too3

Replies3

Following0