Why AdMob’s reported iPhone and Android market shares are inflated

Interesting read.<p>I wonder why he didn't include <a href="http://www.canalys.com/pr/2010/r2010021.html" rel="nofollow">http://www.canalys.com/pr/2010/r2010021.html</a> as reference for smartphone share. Also i think, when he is trying to "guess" the real marketshare, he is underestimating symbian and also i think he is not really aware of the huge divergence of marketshare in different countries all around the world. He assumes what is in the US is true elsewhere which is absolutely not the case. Even more so when talking about featurephones. For example, Symbian/Nokia has huge marketshares outside of the US, but RIM is much bigger in the US then anywhere else.<p>Unfortunately i don't find the statistics anymore, but the mobile market looks quite different on every continent.<p>edit: From admob too, but look at the totally different marketshares of symbian and blackberry on the different continents! -&#62; <a href="http://metrics.admob.com/2010/01/december-2009-mobile-metrics-report/" rel="nofollow">http://metrics.admob.com/2010/01/december-2009-mobile-metric...</a>

Article Summary:<p>AdMob is an advertising network that specializes in showing ads for mobile phones. Their ads can be shown in web pages loaded in mobile web browsers, or the ads can be embedded in mobile applications that run natively on the phones. AdMob does not segment ads in browsers from ads in applications so a single device running a browser and several applications with AdMob ads count count as several devices in AdMob’s figures. This likely leads to a significant inflation of the total number of iPhone and Android users.<p><a href="http://hnsummary.com/2010/03/30/why-admob%E2%80%99s-reported-iphone-and-android-market-shares-are-inflated/" rel="nofollow">http://hnsummary.com/2010/03/30/why-admob%E2%80%99s-reported...</a>

A misleading title. Peter-Paul himself says that AdMob is being very clear about what it measures. The real issue is misinterpretation of the data by analysts.<p>In response to Peter-Paul's question, yes, AdMob could separate out impressions to applications and impressions to web browsers. The user agents are different.<p>However, I believe some (not all) application usage should be legitimately considered the equivalent of browser usage when determining browsing share by platform - separating out in-application and in-browser impressions would be equally misleading. Use of the New York Times' application or Instapaper's application, for instance, is a direct substitute for web browsing, and should be taken into account.