Anyone else seeing better results from smaller AI content batches vs mass publishing?

[mercer]

I’ve been testing this across a few niche sites over the last couple of months, and I’m starting to think the old “publish a lot and let Google sort it out” approach is getting weaker.

What I’m seeing:
– Smaller batches of 5–15 articles per topic cluster seem to index and settle faster
– Pages with stronger internal linking and tighter topical focus tend to outperform larger AI dumps
– Updating and expanding existing pages often gives a better ROI than pushing out new posts every day
– Thin “good enough” AI content still gets stuck unless there’s clear search intent match and some real value added

For context, these are mostly affiliate sites in low-to-mid competition niches, built on WordPress. I’m using a mix of AI drafting, manual editing, and a fairly aggressive internal linking workflow. Nothing fancy, but I do try to keep the structure consistent:
– one primary keyword per page
– supporting entities covered naturally
– links from hub pages to supporting articles
– a few contextual links back into money pages

One thing I’m testing now is whether Google responds better when a cluster is launched in a more “complete” state rather than drip-fed over several weeks. My suspicion is that once the site has enough internal relevance signals, the whole section performs more predictably.

I’m also paying attention to GEO-style optimization, but I’m not convinced it’s a magic lever. In my experience, it helps most when the content is already strong and well-structured. It doesn’t seem to rescue weak pages.

Curious what others are seeing:
1. Are you getting better results from smaller, tighter content batches?
2. Do you launch clusters all at once or publish gradually?
3. Has anyone found a repeatable edge with internal linking patterns?
4. Are you still seeing value in mass AI content production, or has that mostly burned out?

Would be good to compare notes, especially from people running actual sites rather than just testing in theory.

[Author]

Posts