How Much Traffic Does an AI-Run Blog Get? 345 Page Views in 28 Days, Up 4x
TL;DR results
Period: 31 Aug to 27 Sep 2026 (28 days), compared with 28 Jul to 24 Aug 2026
- Page views in 28 days
- 345
- Growth vs first month
- 4x (from 86)
- Posts published
- 116
- Engaged sessions
- 183 of 258 (71%)
First-person write-ups of a test we actually ran drew the most readers. Generic news reaction posts drew almost none.
What we did
We run three blogs where AI does the research, the writing, the editing and the posting, and a person only sets the direction. Each blog has its own subject:
- an AI news and analysis blog,
- an investing research blog that writes up strategy tests,
- a product comparison blog for everyday household items.
The first posts went out in late July 2026. By 3 October 2026 the three blogs had 116 posts between them, published on a schedule without anyone writing a draft by hand.
We measure readers with an analytics tool that filters out bots, and we look at a rolling 28-day window every week. In the 28 days from 31 August to 27 September 2026 the three blogs had 345 page views from 258 sessions, and 183 of those sessions (71%) were engaged, meaning the reader stayed or clicked further. That is four times our first measured month (28 July to 24 August), which had 86 page views.
The steps
- Give each blog one clear subject. A reader should know what the blog is for from any single post. We keep the AI news, investing research and product comparisons on separate blogs rather than mixing them.
- Pick topics from a real source, not a blank page. The AI gets its material from something concrete: a research paper or news item, a strategy test we ran, or a product we compared. That gives every post a fact to stand on.
- Draft with AI, then run a separate editing pass. Ask for a second pass that checks the headline, cuts filler, removes claims the source does not support and makes sure the post answers one question.
- Publish on a schedule. Posts go out automatically at a steady pace, so the blog keeps growing even on days nobody looks at it.
- Link each post to a matching video. Newer posts link to a short video on the same topic, and the video links back, so each piece sends readers to the other.
- Measure with bot-filtered analytics every week. We compare the same 28-day window week after week and look at which individual posts people actually read.
What moved the number
The biggest change was the kind of post, not the number of posts.
| Post type | Example angle | Page views in the window |
|---|---|---|
| First-person test write-up | "I backtested a strategy: here's what happened" | 34 and 16 for the top two |
| Specific, practical analysis | A single named risk and what it changes | around 8 to 10 |
| Generic news reaction | "The latest shift: what it means" | mostly 1 to 2 each |
What the winners had in common:
- A real result in the headline. "I tested this" gives a reader a reason to click that a summary of the news does not.
- One specific topic per post. Posts built around one named idea were read; broad weekly round-ups were not.
- A clear home page. The research blog's home page was its most-viewed page, with 67 page views, so it is worth keeping it tidy as an index of your best posts.
Both of the top two posts were published in September, and between the two windows the research blog that carries them went from 29 to 191 page views, the biggest jump of the three.
What we would do next
- Write more "I tested it" posts. They are our best format, so the next posts on every blog will lead with something we did and its result.
- Retire the generic news format. We will turn weekly summaries into single-topic posts with one clear takeaway.
- Build search slowly on top. Each post will target one question people type into search, so search traffic can add to the readers we already have.
- Keep the 28-day check. The same window every week makes it obvious which change worked.
Free tools for this: Word Counter
Frequently asked questions
Can an AI-written blog rank in search?
It gets indexed, but ranking takes longer than two months. In our 28-day window Google showed our posts only a handful of times and sent no clicks, so all 345 page views came from outside Google search. Plan to bring the first readers yourself, through links from your videos and social posts, while search catches up.
How many posts does an AI-run blog need before it gets traffic?
Our three blogs had 116 posts between them when we measured. The count mattered less than the kind of post: a few posts with a clear, specific angle were read more than dozens of generic ones.
Should I trust my blog platform's built-in view counter?
Use it as a rough signal only. Our platform's counter showed about 1,480 views for the last 7 days, but it counts bots. Analytics that filter bots showed 345 page views over the last four weeks, roughly 86 a week. Make decisions on the filtered number.
What kind of AI blog post gets read?
Posts that report something you did and what happened. Our two most-read posts were first-person write-ups of tests we ran, with 34 and 16 page views in the window. Posts that only summarised the week's news mostly had one or two views each.