Amr Awadallah Blog

As you know, Google last year introduced a feature that dynamically sets the required Minimum Bid for a given keyword (before that they had a flat $0.05 minimum bid for all keywords). I commented on this change as being smart but evil at the same time, since it lets Google price its keyword inventory to perfection and capture every penny possible, thus leaving a small margin for you, the bidder.

This allowed Google to drop the default minimum bid from $0.05 to $0.01, which allows them to capture more listings from keywords that were not ROI positive at 5 cents and above. However, at the same time it protects the high value keywords since Google can over-ride the minimum keyword bids for those terms individually.

So for fun, I went to Google today and checked what is the minimum required bid for some people names, here are the results:

Its really hard to figure out the logic behind their algorithm, if any, so for example, consider these facts based on the numbers above:

• George Bush is valued at twice Osama Bin Laden
• Arnold Schwarzenegger is valued three times Tom Cruise
• Steve Jobs is worth twice as much as Bill Gates
• Isaac Newton is worth twice as much as Albert Einstein
• Alicia Witt is the most valued among this list, she is worth five times Britney Spears
• Terry Semel, Larry Page, Mark Cuban and Bill Gates all have the same value
• Among the bloggers I checked, Ken Norton topped the list at $0.30
• I, Caterina, Chad, and JenSense were tied as the cheapest bloggers, at $0.05 each
• Poor Udi Manber is valued at $0.03, hopefully his recent move to Google will help improve that 😉

— amr

I guess its ok when media makes mistakes interpreting Comscore reports, they do not deal with numbers much, but its painful to see financial analysts making the same mistakes.

The search share that is relevant for revenue is not the Comscore total media reported qSearch share which includes all types of search like image, video, directory, groups, etc. These ancillary verticals do not contribute to revenue (not directly), hence they can double marketshare in those areas and still make no additional revenue for that quarter.

Also that specific report had a ton of methodology changes over the last year, so year-over-year comparisons are totally out of whack. For example, Comscore had a technical issue with how they count Google image search PVs which they corrected in Jan 06, that led to inflating Google’s growth in Jan and Feb compared to last year.

That said, the proper metric to look at is the Web Search PVs marketshare, and specifically within the top 5 players (Google, Yahoo, MSN, AOL, and ASK), since Comscore added a new bunch of ISP search portals to their total report in Jan 06.

Cheers,

— amr

PS: Bearn Stearns, you also had a typo in the table on page 2 of your report, the column headers for the last two columns should read Jan 06 and Feb 06 (instead of Jan 05 and Feb 05).

Sorry for the light blog posting lately, very busy with a couple of projects at Yahoo and exploring G.R.A.W. on the 360.

I grabbed this screenshot from Google’s SERP more than a week ago, I think this is the last change to make the Google SERP look identical to the Yahoo SERP 🙂 (actually Yahoo shows 4 ads in the north in very rare cases, and Google only shows up to 3, though I did see them testing 4 last year).

Ignore the result numbering (like 9. and 10. for the last two results) that is an artifact of the CustomizeGoogle module (I like numbering for results, and it has an option to add that).

Finally, don’t you guys think that this promotion for Desktop Search is horrible from a UI point of view? it overlaps heavily with the pagination module at bottom of their SERP that you can mistakenly click on it instead of the Next arrow:

Cheers,

— amr