Your chart doesn't reflect what the data actually tells us. If overall number of queries has grown, then the fact that Google's share has increased only slightly is less significant than the fact that thtee other services did not lose share.

In other words, their overall query traffic must have increased too. For example, if Microsoft took 10.6% of 6.831 billion queries (that is, 724,086,000 queries for Microsoft in Janaury) then 10.5% of 6.9 billion queries indicates they served 724,500,000 queries in February.

We don't have to quibble over statistical significance and semantics. My point is that people are using all but one of the major search services more than before.

Yahoo!, Microsoft, and AOL are therefore growing their query traffic (this metric, however, is not very useful for measuring market share despite the fact it is preferred).

Ask appears to have actually lost query traffic.

These properties should be measuring their markets in terms of unique visitors, repeat visitors, and unique visits -- all with respect to number of queries per visit and visitor. They need to look at how long people use their query services and why those people stop using their query services.

And while I don't expect to see the data any time soon, it would helpful if we could see breakdowns of the data by types of search: Web, Blog, News, Image, and Local (I realize there are other verticals but these are the most important in my opinion).

NOTE: And if I were IAC, I'd be co-branding Ask Search into every property they operate. Their failure to do so -- even to cobrand Ask search into AskCity -- should go down as one of the greatest marketing blunders in Web history.

Comment by Michael Martinez | March 23, 2007 5:25 PM

> In other words, their overall query traffic must have increased too.

No, that's not always the case. In some months, share drops even though queries have gone up. In other months, it's the opposite. I might do a future look at number of queries versus shares in more depth. It is important to understand if query volume has actually risen even if share has dropped. That's why I started exploring that issue in this post.

There are, of course, many other ways you can measure traffic and value aside from these metrics. I'll try to get into those more in the future, as well. But share remains a good starting point, and my focus with share remains trying to put the figures in perspective by showing trends and comparing those trends across multiple rating services.

Comment by Danny Sullivan | March 24, 2007 1:11 PM