Japan Online Media Audience Data: Page Views, Unique Browsers and What Publisher Numbers Mean

Short answer: Japanese publishers usually report monthly page views (PV) and unique browsers (UB), and they report them themselves. We found no independent audit standard for these web figures. Third-party tools such as Similarweb estimate visits, which is a different metric again. None of these numbers tells you how many people will read one article about your product. Use publisher figures to judge an outlet's rough scale and fit, and measure actual coverage separately.

Why these numbers confuse overseas brands

When an overseas brand starts planning PR in Japan, it often asks for the "reach" of each outlet. Japanese media kits answer with large numbers: millions of page views, millions of unique browsers. An agency report may then add those numbers up and present the total as the audience of a campaign.

That's misleading. The figures measure a whole site over a month, they're counted differently by each publisher, and almost all of them are self-reported. This guide explains what the metrics mean, what verified figures exist for Japanese tech media, and how to use them without overstating anything.

PV, UB, UU and visits: what each one counts

Page views (PV). A page view is counted each time a page loads, including reloads. One reader who opens a gallery article with ten pages can produce ten PV. Sites that split articles into several pages, or that have many short pages, can report high PV.

Unique browsers (UB). This is what Impress and ITmedia actually report: both label their figures UB, not UU. A browser is identified by a cookie or similar ID. One person who reads on a phone and a PC, or who clears cookies, counts as several browsers. So UB overstates the number of people.

Unique users (UU). In analytics tools, unique visitors are also identified by IDs stored in browser cookies (Adobe Experience League, in Japanese). Many Japanese media kits use "UU" loosely for what is really a browser count. Don't read "UU" as "people" without checking the method.

Visits (sessions). Similarweb and similar tools estimate visits, meaning sessions, using panels and models. A visit can include several page views. Visits are not comparable with PV or UB, and they're estimates, not counts.

Self-reported, estimated and audited figures

Self-reported publisher figures. Most numbers you'll see come from the publisher's own media guide or ad page. Some state their method: ITmedia says its September 2026 report is measured with its own ad-serving server, with bots and crawlers excluded in principle (ITmedia media data, in Japanese). Impress's ad guide pages don't state the tool (Impress ad guide, in Japanese). Others give a number with no period at all.

Third-party estimates. Similarweb's public pages, checked 20 October 2026 and headed September 2026, showed about 95.9 million visits over the last three months for itmedia.co.jp, about 40.2 million for impress.co.jp, and about 3.2 million for ascii.jp (Similarweb, itmedia.co.jp; impress.co.jp; ascii.jp). These are estimates, three-month totals and visits, not PV. The page doesn't say whether subdomains such as pc.watch.impress.co.jp are fully included. The ascii.jp estimate is far below KADOKAWA's own stated figure for the ASCII brand, which shows why estimates and publisher counts can't be compared. Use estimates only for rough relative scale, and always label them.

Audited figures. For print, 日本ABC協会 (Japan Audit Bureau of Circulations, founded 1952) publishes audited magazine circulation. Since 2016, its magazine reports have also carried web and social "brand indicators" for magazine brands, and from the second half of 2017 web figures and off-platform syndication figures. Those web figures appear in publisher-submitted reports, not audited ones. We found no audited page-view standard for Japanese tech news sites comparable to audited print circulation. In practice, Japanese web audience figures are almost all self-reported.

Verified publisher figures for Japanese tech media

The table shows figures we could verify from publishers' own documents, with the period each one covers. They're publisher-reported and unaudited. Don't add them together, and don't compare sites measured in different months or by different publishers as if they were the same metric.

*ITmedia's report shows site names as images; we matched each row to a site by its description text. ねとらぼ isn't in the September 2026 report; its figure comes from the July–September 2025 edition of its media guide and reflects early-2025 data.

Some other publishers state figures with no period. In October 2026, Mediagene's media page gave GIZMODO JAPAN 22.4 million or more monthly PV and 5.9 million or more unique users, and KADOKAWA's guide gave the ASCII brand about 60 million average monthly PV. Without a period, you can't tell whether these describe last month or several years ago. We found no current publisher figures for CNET Japan, Mynavi News, Kakaku.com Magazine, BCN+R or 360LiFE.

Figures change, and periods matter

The same site can report quite different numbers a year apart. Impress's April–June 2025 media guide gave PC Watch 20,996,331 monthly PV and 4,143,284 UB for January 2025 (Impress media guide, 2025, in Japanese); its July 2026 figures are lower on both counts. Always note the month and the source next to any figure you put in a plan.

Aggregators: platform totals are not article reach

News aggregators publish the biggest numbers of all, and they're the easiest to misuse.

Yahoo!ニュース (Yahoo! News Japan). LY Corporation says Yahoo! News averaged about 16.5 billion page views a month between April 2025 and March 2026 (LY Corporation, July 2026, in Japanese). In the same release it launched ニュースPR (News PR), a paid service that places press releases on Yahoo! News for a fee, with no view guarantee.

The first results show the gap between platform and article. Across the first 450 News PR articles, measured to 23 September 2026, the average was 4,260 PV and the median 2,317 PV; the highest reached 92,021 (GameBusiness.jp, October 2026, in Japanese, reporting LY Corporation's figures). A typical paid article on a platform with billions of monthly page views drew a few thousand.

SmartNews. SmartNews says it has more than 3,000 partner media and delivers more than 400,000 articles a day (SmartNews, in Japanese; no date given). We found no current official user figure for SmartNews, LINE NEWS or Gunosy.

When an article is syndicated to an aggregator, the aggregator's total traffic tells you nothing about how many people read that article.

The same gap on paid placements

Paid placements make the difference explicit, because publishers guarantee article page views, not site traffic. Impress's April–June 2025 media guide priced a native tie-up article on a Watch site at ¥1.5 million, tax excluded, with a guarantee of 15,000 PV over a month on sites such as PC Watch, Kaden Watch and AV Watch (Impress media guide, 2025, in Japanese). Those same sites report millions or tens of millions of PV a month.

In other words, the publisher itself expects one paid article, promoted on the site, to reach tens of thousands of views, not millions. An earned article without promotion has no guarantee at all. Paid tie-ups are advertising and must be labeled; see reviews, sponsored editorial and seeding.

How to use these numbers in PR planning

Use them to judge scale and fit, not to forecast readers. A site with 3 million monthly UB and an exact match to your category is a better target than a bigger site that rarely covers your product type. For which sites cover which categories, see our guide to Japanese gadget websites and Japan Tier 1 and Tier 2 media.

Keep every figure labeled. Write the metric (PV, UB, estimated visits), the month and the source next to each number. "PC Watch, 18.2M PV, July 2026, Impress" is useful; "PC Watch reach 18M" isn't.

Never add sites together. Different publishers, months and methods make a sum meaningless, and the same person may visit several sites.

Ask for article-level data where it exists. For paid placements, publishers report the article's own PV. For earned coverage, you usually won't get article data from the outlet; measure what you can see yourself, such as referral traffic to your Japanese product page, branded searches and inquiries. See how to measure product PR ROI in Japan.

Treat estimates as estimates. If you use Similarweb, say so, give the period, and don't present it as the publisher's figure.

For who owns and edits the sites behind these numbers, see Japanese tech media publishers profiled; for how aggregators carry stories, see Yahoo! News, SmartNews and LINE NEWS.

Where Gloture fits

When we plan product PR in Japan for an overseas brand, we choose outlets by beat and recent coverage, and we show any audience figure with its metric, period and source, or say plainly that none is published. We report coverage article by article, and we don't present site traffic as the reach of an article. Product PR in Japan starts from US$2,500. We don't guarantee coverage or readership. To see which outlets fit your product, start a Japan PR Review.

Audience data checklist

Frequently asked questions

Are Japanese media audience figures audited?

Not for tech news sites, as far as we could find. Publishers report their own PV and UB, sometimes describing the method, as ITmedia does. Japan ABC audits print magazine circulation; web figures in its magazine reports are publisher-submitted.

What is the difference between UB and UU?

UB counts browsers. UU in analytics also relies on browser cookies, and many Japanese media kits use the two terms loosely. Either way, one person on two devices counts twice, so neither equals the number of people.

Can we say our article reached a site's monthly audience?

No. A site's monthly PV or UB covers every page on the site. One article reaches a fraction of that. The paid examples show it: Impress's 2025 guide guaranteed 15,000 PV for a tie-up on sites that report millions, and Yahoo! News PR articles had a median of 2,317 PV.