The 5-second answer
A small site's analytics showed roughly 5,780 unique visitors and 281,000 requests at the zone level over 30 days. Turn on bot exclusion in Web Analytics, however, and the picture becomes 160 visits, 330 pageviews, with only about 40 pageviews on actual article pages.
So the dashboard effectively says:
Machines: “We come here all the time.”
Humans: “About 40 article pageviews this month.”
That sounds catastrophic until you add the missing context: only about 35 articles are currently public while the site waits for advertising review. A much larger unpublished article inventory is not being judged by those 40 views.
Revenue is equally sobering. A common Japanese rule of thumb for Google AdSense is around ¥0.2–¥0.3 per pageview. At that level, even 100,000 monthly pageviews produces only about ¥20,000–¥30,000.
High-ticket affiliate offers can pay tens of thousands of yen for one approved conversion. But if you start choosing article topics because the commission is large, the monetization mechanism begins controlling the site's purpose.
The clean rule is:
Purpose first. Monetization second.
The 30-second overview
This started with a funny mismatch in analytics.
Cloudflare's zone dashboard displayed large numbers: about 5,780 unique visitors and 281,000 requests. It looked almost respectable.
Cloudflare's own documentation explains why that can be misleading. Zone HTTP analytics can include legitimate users, crawlers, and threats, and a “request” is not a pageview. Loading one page can generate multiple HTTP requests for images, stylesheets, scripts, and other resources.
Cloudflare Web Analytics has a bot-exclusion dimension that the company describes as producing a result closer to real-user traffic. In this case, that reduced the picture to 160 visits and 330 pageviews in 30 days, with roughly 40 article-body pageviews.
Meanwhile, Google says AdSense site review usually takes a few days but can take 2–4 weeks.
So the site is effectively a shop that is not fully open yet, while automated visitors are already circling the building.
The mall is still preparing the shelves. The delivery trucks have arrived 281,000 times.
1. The first surprise: the machines are the regular customers
The 30-day case-study metrics were:
| Metric | 30 days |
|---|---|
| Cloudflare zone unique visitors | ~5,780 |
| Cloudflare zone requests | ~281,370 |
| Web Analytics visits after bot exclusion | 160 |
| Web Analytics pageviews after bot exclusion | 330 |
Pageviews under /articles/ |
~70 |
| Article-body pageviews after removing the index page | ~40 |
| Article-body pageviews in the last 7 days | 6 |
| Article-body pageviews in the last 24 hours | 0 |
The first mistake would be to read 5,780 as “5,780 people read the articles.”
Cloudflare states that its Free-plan zone HTTP analytics include legitimate user requests as well as crawlers and threats. It also notes that a typical pageview requires many requests.
So 281,000 requests does not mean 281,000 pages read.
One human opens a page.
Several resources load.
A search crawler visits.
A monitoring service checks the site.
Other automated clients arrive.
Put them all on one meter and the tiny town suddenly looks like Tokyo Station.
2. How 5,780 “unique visitors” and 160 visits can both be true
Cloudflare offers analytics products with different definitions and purposes.
In Web Analytics, a pageview is a successful HTML response. A visit starts with a pageview that came from another site or a direct link, and one visit can contain multiple pageviews.
The “Exclude Bots” dimension is particularly useful. Cloudflare says enabling it gives a dataset that is a closer representation of real-user traffic.
That still does not mean “160 visits = exactly 160 individual humans.” Analytics systems differ in cookies, client execution, referral handling, bot detection, and identity assumptions.
The practical goal is not to discover one metaphysically perfect human count. It is to track the same definitions consistently over time.
For this site, a sensible baseline is:
Visits → pageviews → article-body pageviews.
At the current level, 330 pageviews divided by 160 visits equals about 2.06 pageviews per visit. That becomes a useful future comparison point.
3. Forty article pageviews across 35 public posts is not a verdict yet
About 40 article pageviews is tiny in isolation.
But the publication state matters.
The site is still in advertising review and only around 35 articles are public. This is not the same situation as publishing roughly 1,200 articles and receiving 40 total pageviews.
If you mechanically divided 40 views by 35 posts, the average would be about 1.14 pageviews per article per month. In reality, traffic is concentrated in a few posts, so that average is mostly comic relief.
Google's AdSense help center says site reviews usually take a few days but can sometimes take 2–4 weeks.
The current numbers should therefore be treated as pre-launch observation data rather than a final business result.
Only 35 products are on the shelves. The warehouse is much larger.
4. If 1,200 published articles produced 40 pageviews, then yes, call an emergency meeting
Now take the nightmare scenario literally.
40 ÷ 1,200 = 0.033 pageviews per article per month.
At that rate, the average article would wait about 30 months for one view.
And if every article took one full week to create manually:
1,200 ÷ 52 ≈ 23.1 years.
Twenty-three years of publishing for 40 monthly pageviews would not be a minor SEO problem. It would be a philosophical event.
Season finale: “We got 40 again.”
Fortunately, that is not the current situation. The distinction matters because judging a large unpublished inventory by a deliberately small review-period catalog would be analytically wrong.
5. What do article count and traffic look like in other blogs?
Productive Blogging's 2026 income survey collected responses from 129 monetization-oriented bloggers, mainly in English-speaking markets. It is not a representative census of all websites, but it provides unusually useful combined data on article count and pageviews.
| Number of posts | Average monthly pageviews | Average total monthly income |
|---|---|---|
| 0–49 | 473 | $36.91 |
| 50–99 | 3,839 | $38.05 |
| 100–299 | 37,370 | $1,987.52 |
| 300–499 | 85,822 | $5,106.66 |
| 500–999 | 166,853 | $9,460.09 |
| 1,000+ | 100,641 | $9,412.50 |
Do not read this as “1,200 posts guarantees 100,000 monthly pageviews.”
These are averages from a self-selected, monetization-focused sample. Total income also includes advertising, affiliates, products, services, and other revenue sources.
The interesting part is that average pageviews actually fall after the 1,000-post mark. The survey author suggests one possible reason: maintaining and optimizing a huge archive becomes difficult, and stale content may drag performance down.
Posts are assets. Unmaintainable posts can also become inventory debt.
6. More posts do not automatically produce more traffic
Ahrefs analyzed roughly 14 billion pages in a 2023 study and estimated that 96.55% received no organic traffic from Google, while another 1.94% received only one to ten monthly visits.
Ahrefs explicitly notes limitations: its traffic numbers are estimates and its sample is not a perfect representation of the entire web.
Still, the basic lesson is hard to miss.
Published does not mean discovered. Discovered does not mean clicked.
A topic may have no search demand.
The article may miss search intent.
Competition may be stronger.
Internal and external links may be weak.
The page may not be indexed.
The information may become stale.
Several similar pages may compete with each other.
After a large release, the important question is therefore not simply “How many posts exist?” It is:
How many were indexed, shown in search results, and clicked?
7. Multiply the ad numbers and reality becomes salty
A Japanese Xserver guide gives a common AdSense rule of thumb: roughly ¥20–¥30 per click, around a 1% click-through rate, and therefore about ¥0.2–¥0.3 revenue per pageview. Google itself does not publish a universal average, so this is an industry estimate, not an official guarantee.
Using that range:
| Monthly pageviews | Approximate AdSense revenue |
|---|---|
| 1,000 | ¥200–¥300 |
| 10,000 | ¥2,000–¥3,000 |
| 100,000 | ¥20,000–¥30,000 |
| 500,000 | ¥100,000–¥150,000 |
| 1,000,000 | ¥200,000–¥300,000 |
One hundred thousand pageviews sounds powerful.
Then the estimate says ¥20,000–¥30,000.
The seasoning suddenly becomes very salty.
The 2026 Productive Blogging survey tells a similar story from another market: Google AdSense averaged $3.26 per thousand pageviews among respondents, far below several premium ad networks.
AdSense is excellent at turning spare page real estate into small automatic revenue. It is not designed to turn low traffic into large money.
8. Do not apply “one week per article” to 1,200 articles unless you enjoy 23-year projects
If every article takes one week, 1,200 articles require about 23 years.
That does not mean deeply researched articles are wasteful. Original experiments, measurements, field experience, interviews, and expert verification can absolutely justify serious production time.
The inefficiency appears when every article is forced to carry the same production weight.
For a large information library, a better model can be:
Publish broadly enough to test demand.
Observe actual impressions, clicks, and engagement.
Then upgrade the winners with original data, diagrams, stronger sourcing, clearer examples, and updates.
In other words:
Do not send every article to the championship match on day one.
Let the qualifiers earn their equipment.
9. High-ticket affiliate commissions can change the math with one conversion
A8.net's high-commission ranking, updated August 10, 2026, included examples such as:
- Career-service free consultation: ¥50,000
- Engineer recruiting interview setup: ¥78,000
- Fiber internet activation: ¥30,000
- Car-lease contract: ¥30,000
If AdSense is worth roughly ¥0.2–¥0.3 per pageview, one approved ¥50,000 conversion is numerically equivalent to about 167,000–250,000 AdSense pageviews. A ¥78,000 conversion corresponds to roughly 260,000–390,000 pageviews.
That looks absurdly powerful.
But affiliate revenue requires the entire chain to work:
Click → application → qualification → approval.
A ¥50,000 commission with zero relevant readers still produces exactly ¥0.
10. If you choose topics because the commission is high, the site's purpose starts to die
This is the important part.
“There's a ¥50,000 offer, so let's manufacture career-change articles” is a dangerous editorial order.
Google's people-first content guidance asks whether a site has a primary purpose or focus and warns against producing large amounts of content across many topics mainly in the hope of attracting search traffic.
A multi-topic site is not automatically a problem. The problem is letting ranking potential or commission size become the main reason a topic exists.
If the site's purpose is to make annoying parts of work and everyday life easier to understand, then write the useful article first.
If a career comparison is genuinely the reader's next step, a career-service affiliate link can be natural.
If a rain-gear article comes from real use and comparison, a product link can be natural.
But if an article about running better meetings suddenly says:
“By the way, here's a ¥50,000 career consultation!”
the reader can feel whose problem the link is solving.
11. The clean model: article purpose first, monetization second
A practical order is:
- What problem does this article solve?
- What is the reader's natural next action?
- Is there genuinely a product or service that helps with that action?
- If yes, consider an affiliate link.
- If no, do not force one.
- Use display ads as baseline revenue only where they do not wreck the experience.
That keeps the commission table from becoming the editor-in-chief.
Display ads become the system that collects a little value across the whole site.
Affiliate links become the high-power system used only when the context is an exact match.
Neither one is the site's purpose. They are revenue mechanisms attached to useful content.
12. Twelve languages are not a 12× traffic button
The same logic applies to multilingual publishing.
Publishing one article in 12 languages does not automatically multiply traffic by 12. Search demand, competition, ad markets, and intent differ by language.
Google recommends using separate URLs for language versions and using hreflang or equivalent signals to show their relationships. Google also determines language primarily from visible page content, so each localized page should actually read as a page in that language rather than a mixed-language container.
Measurement should therefore be split by language as well.
A topic may win in Japanese.
Another may unexpectedly perform in Indonesian.
An English version may have the largest search market but the strongest competition.
The value of 12 languages is not a magical 12× multiplier. It is the ability to test the same useful idea across 12 markets and observe where it resonates.
13. What to record after ad approval and the larger release
After the larger publication wave, monthly pageviews alone are not enough. Track the funnel in order:
| Stage | Metric | What it tells you |
|---|---|---|
| Supply | Published article count | How much content is live |
| Discovery | Indexed URL count | Whether search engines recognize it |
| Exposure | Search impressions | Whether pages reach actual queries |
| Acquisition | Search clicks and click-through rate | Whether titles and intent match |
| Reading | Bot-excluded visits and pageviews | Usage closer to real readers |
| Navigation | Pageviews per visit | Whether internal links move readers |
| Ads | Revenue and revenue per 1,000 pageviews | Advertising efficiency |
| Affiliate | Link clicks, applications, approvals, revenue | Commercial-path efficiency |
| Locale | Impressions, clicks, and revenue by language | Which markets respond |
The most interesting experiment is preserving the same table before and after the large release.
If article-body traffic moves:
40 → 100 → 500 → 1,000 → 10,000
you can investigate what changed at each stage.
If 1,200 articles are public for months and the number is still 40, then the emergency meeting becomes justified.
14. Conclusion — for now, the crawlers love the site more than humans do
The current state can be summarized in one sentence:
Humans have barely discovered it. Machines have discovered it enthusiastically.
With only about 35 articles public during ad review, roughly 40 article-body pageviews are not enough to judge the full project.
The real experiment begins after approval: publish the larger library, then track search impressions, clicks, bot-excluded visits, pageviews, and revenue under consistent definitions.
And monetize only where the monetization method matches the article's actual job.
Ad review: “Checking…”
Crawler: “Visited.”
Crawler: “Visited again.”
Human: “…about 40 article pageviews.”
For now, the shop has more inspectors than customers.
What happens after the doors fully open may be the most interesting data point of the entire project.
Sources
- Cloudflare, Zone Analytics: https://developers.cloudflare.com/analytics/account-and-zone-analytics/zone-analytics/
- Cloudflare, Web Analytics Dimensions: https://developers.cloudflare.com/web-analytics/data-metrics/dimensions/
- Cloudflare, High-level metrics: https://developers.cloudflare.com/web-analytics/data-metrics/high-level-metrics/
- Google AdSense Help, Connect your site to AdSense: https://support.google.com/adsense/answer/7584263?hl=en
- Productive Blogging, Blogging Income Survey 2026: https://www.productiveblogging.com/how-much-do-bloggers-earn/
- Ahrefs, 96.55% of Content Gets No Traffic From Google: https://ahrefs.com/blog/search-traffic-study/
- Xserver, Google AdSense CPC and revenue-per-pageview guide: https://www.xserver.ne.jp/blog/adsense-cpc-average/
- A8.net, High-commission ranking, updated 2026-08-10: https://support.a8.net/as/HintOfProgram/ranking/highprice.php
- Google Search Central, Helpful, Reliable, People-First Content: https://developers.google.com/search/docs/fundamentals/creating-helpful-content
- Google Search Central, Managing multilingual sites: https://developers.google.com/search/docs/advanced/crawling/managing-multi-regional-sites
