Back when I predicted this in January, I recall worrying I was calling it too early. Now it appears the timing was about right. From Mashable:
…while Google grew from June to July, it still lost market share to its competitors – from 66.1% in June to 64.8% in July, a 1.3 percentage point drop.
From my prediction: 3. Google will see search share decline significantly for the first time ever. It will also struggle to find an answer to the question of how it diversifies its revenue in 2009.
There’s more to be said on that second point, revenue diversification. More on that after the summer break I’m supposedly on.
1.3% is significant?
Agree with the second point, not sure a 1.3% drop is a sign your first point is correct.
Whilst 1.3% may seem like a small figure – it’s not that small to those that benefit from it.
With Bing on 9.0% for July 2009 and Ask.com on 1.7% – a 1.3% increase for a competitor is a significant change.
What I’d question though is the accuracy and relevance of these figures? Primarily, where does Neilsen actually get it’s data – from the search companies themselves? And also, this is just for the US market – I’m sure Baidu is the #2 ranked search engine globally – and I would be interested to see whether there’s a Google-defect on a global scale or whether it’s just the US that are susceptible to marketing campaigns.
Google is becoming more a content provider and less a search service (that sends traffic to other sites).
The metrics that are published by companies like comScore, Compete, Hitwise, and Nielsen really have no relevance to SEARCH — they are just estimating page views within the large networks that companies like Google, Microsoft, Yahoo!, AOL (Time-Warner), and Ask operate on the Web.
There are currently no accurate, reliable measurements for search — a fact I have shared often but which others occasionally acknowledge as well (Cf. http://www.marketingpilgrim.com/2009/08/why-katy-perry-can-predict-search-share-better-than-nielsen-comscore.html).
These companies have tremendously awful metrics. I’d give them plus or minus 10% of reality with a 95% confidence interval. A two percent change is statistically insignificant.
1.3% seems like a small % with the rise of Bing.