Google Japan’s year-end announcement on Dec 22nd, 2008 shocked the industry, when they loudly proclaimed that their search market share had reached 41%, which is only 3 points behind the 44% of Japanese portal giant Yahoo! Japan, using data from Nielsen/NetRatings (report on @IT [J]).
The news raised some eyebrows because no such indications had been observed elsewhere, though Google seems to be incrementally increasing their share in recent years, and collaboration with the #1 and #2 cellphone carriers (NTT Docomo and KDDI/au, respectively) gives Google some power to strike back at Yahoo in the mobile search market (N.B. NetRatings’ statistics do not include mobile data).
Then, a month later, the same Nielsen/NetRatings’ newsletter showed totally different, but rather familiar, numbers: Yahoo! 52.5% vs. Google 38.1%. (*1) (*2)
In the newsletter, they acknowledged this discrepancy, and explained how the two totally different results came from the same data source. Yes, NetRatings.
According to the newsletter, the number which Google announced, might have been taken from the “Channel Search/Search Category”. In these stats, all of the top page views on search sites like Google, Ask.jp and Baidu are counted as “search” activities. However, for portal sites like Yahoo! and MSN, top page views are excluded because their top pages are also used by non-search users.
But Google Japan, which is different from the U.S. Google, has some portal-like features including Gmail/YouTube links. And of course, many people perform searches from Yahoo! Japan’s top page. So this “search”/”portal” distinction makes little sense.
The newsletter alternatively suggested that search “results” page views be counted, which leads to percentages of 52.5% vs. 38.1%.
@IT (Atmark IT), the original source of Google’s big news, and a popular online IT news media website, admitted that they did not confirm Nielsen/NetRatings’ figures and reported only what Google Japan had announced. They also followed up with NetRatings’ statement that “Google’s announcement had not been checked by NetRatings, which is normally required when their data is used for public release, and the numbers include some unfair interpretation.”
(*1) The newsletter only shows the top 10 search engine players. SEM Bar blogger Kuroyagi totalled the top 10’s access data for all the search page views. So if you include unknown small players ranked lower than 11th place, both Yahoo’s and Google’s share would be a bit lower than the real number, but it should not make a big difference.
(*2) ComScore, a competitor of Nielsen/NetRatings, showed similar numbers: Yahoo! 51.2% vs. Google 39.0% in September 2008 [J]