I think the author doesn’t understand that Google scrubs the raw data, to come up with their annual list, as discussed in this Official Google Blog entry:

we do not simply retrieve the most frequently-searched terms for the period—the truth is, they don’t change that much from year to year. This list would be predominated by very generic searches, such as “ebay”, “dictionary”, “yellow pages,” “games,” “maps”—and of course, a number of X-rated keywords. These are constants, and although unquestionably popular, we don’t think they actually define the Zeitgeist.

Instead, we looked for those searches that were very popular in 2006 but were not as popular in 2005—the explosive queries, the topics that everyone obsessed over.

Thus, all your conclusions just don’t follow based on the above. If one wanted to research the topic more carefully, one can study the data AOL released last summer.

Absolutely correct George.  If Google published the ‘real’ most popular search terms list the preacher’s wife would blush and their stock price would tank. Having bought search engine queries in the past I can corroborate that the most popular search terms are adult/XXX in nature followed by the usual suspects: Britney Spears Paris Hilton and Pam Anderson. There are lots of mis-spellings (you didn’t expect people to spell right did you?) and surprisingly lots of people still type “Google” even though “they are at” Google .. go figure. 

This is my own rough estimate but I would say a 25-35% of all of Google’s Searches (well north of one quarter of all their traffic) is for completely unmonetizable stuff: “Frog in a blender”, “Smoking Fetish Porn”, “2asjkdhkj” style jiberish and some pretty unspeakable stuff. Any other search engine would share the same characteristics.

Thank you for your comments Mr. Kirikos and Mr. Schilling.  Regardless of whether the information is scrubbed or not, the underlying message still remains.  It is true that the constants and the gibberish will always be searched, but the “the explosive queries,” as stated by Google in its official post, were unquestionably dominated by trademarks. 

As an aside, am I the only one that finds it ironic that Google lumped “EBay” in with other “very generic terms” such as “dictionary”, “yellow pages,” “games,” and “maps”?

I raised an eyebrow at the inclusion of “ebay” in a list of generics, since it is clearly a trademark. Still, searching on the term “ebay” is not necessarily the same as searching for eBay(TM). I’d expect the first hit to be eBay itself, but what are the remainder? This might even be interesting to me if I were an eBay user. As it happens, I’m not, but many are.

What’s more interesting to me is the tint of the glasses through which you, Mr Palmer, view this data. You think of “trademarks”, I think of “proper nouns”; you think of “generic terms”, I think of “common nouns”. I don’t think of “Wikipedia” or “world cup” as trademarks. They may be that, but they’re just names first and foremost. The fact that many people think of these names as trademarks first and foremost demonstrates a certain mindset: a sadly pervasive one that devalues anything without a balance sheet. This is emphasised by the comment that “consumers” are the ones doing the searching. Can’t these just be people looking up new buzzwords, rather than consumer units seeking producer units to offer currency in exchange for goods and services?

Many of the names are, I grant you, the names of businesses, and those are full-blooded trademarks. What is interesting and not shown by this data is the ratio between those that seek the site by direct navigation (such as typing “bebo.com” in the address field) vs those that seek the site by searching on the name. That’s data I wish I knew.

Ultimately what this list shows is “current buzzwords”. It comes as no great surprise that proper nouns (trademarks or not) feature prominently among them. Perhaps the interesting thing is the matter of which entities managed to generate a buzz this year. Were there more searches for “Wikipedia” than “Linux?” The data doesn’t say, but it seems that Wikipedia’s star was still rising in 2006, whereas Linux is entrenched in the mainstream.

Lastly, for additional perspective, compare this year’s top ten with last year’s. “Myspace” fell from slot one to slot two, overtaken by “Bebo”, which is a new entrant to the list. “Wikipedia” fell from fourth to sixth. “Ares”, “Baidu”, “Orkut”, and “iTunes” featured last year, but are now gone. In fact, there were no common nouns in last year’s list, whereas this year saw “video” and “wiki” as big gainers. It’s genuinely impressive for a common noun to make the list, since they don’t come and go in the way proper nouns do.

One can use the Google Trends tool to get a better sense of the popularity of the generic terms that are scrubbed from the Zeitgeist annual report.

For example, “Bebo” has a long way to reach the popularity of “sex”. “Games” are searched much more often than “Bebo”. “Bebo” is not yet as popular as the F-word. The internet was made for porn, not “Bebo”, if this trend graph is to be believed. I’m sure one can find countless other examples that were scrubbed.

Thank you for your thoughtful comments Mr. Watson.  I would like to emphasize a consumer need not recognize the name of a business as a trademark or even know what a trademark is in order to use trademarks.  The formation of the search query that contains a trademark is evidence that the marketing affiliated with the mark is working.  If a user goes to Google, or any search engine for that matter, and queries a well known business name, such as BEBO, they are likely searching for one of two things: (1) BEBO.com itself, or (2) more information about the BEBO.  The user has associated the searched term with a source of the product or service which they are seeking.  Either way, the user is likely searching specifically for more information regarding the owner and operator of the website.  In other words, the user is searching to locate or learn about the source.  Herein lies the foundational purpose of a trademark: to operate as an indication of source.

Mr. Kirikos: Perhaps we should refocus the discussion on the queries with the highest growth rate over the year.  This would remove the issue of any scrubbed constants.

Mr. Palmer, If Google makes nearly 100% of its revenue by selling advertising on search queries and if a majority of its searches are for trademarks, does that mean Google might make a majority of its revenue by pawning off other people’s brands?

In the domain name space if you register hoovervacuuums.com (spelled deliberately with 3 u’s) you are cybersquatting and subject to potential penalties under the ACCPA. If Google takes over the address bar resolution of the unregistered domain name hoovervacuuums.com (via their error search partnership with the Firefox browser) and if Google then monetizes that type-in and redirection of this unregistered (redirected) domain-name by displaying competing advertising for Vacuum Cleaners and accessories, are they not (in a manner of speaking) breaking the same laws?

Mr. Schilling, it would be an understatement to say that your questions are great ones.  After some thought and some canned queries on Google, it appears that the Google’s in-site advertisements that appear on the right side of the screen (“Sponsored Links”)  are primarily activated when generic terms are searched.  For example, a search for “money”  or “computer”  will spit out a page-full of sponsored links that bid their way to the top of the list.  Alternatively, at the time of my search, a query that contained nothing but a well-known trademark usually resulted in an empty right side of the screen.  For example, “Amazon.com”,  “Yahoo!”,  and “Circuit City” resulted in a lack of sponsored links.  A few outliers that I stumbled upon were: “Best Buy” (sponsored link to IBM entitled “Quality Computer”); and “Buy.com”  (sponsored links by Overstock entitled “Buy at Overstock.com” and Compete.com entitled “Hot Deals”).  This is likely the result of self-governance by keyword bidders, but Google may have stepped in with some preventative measures.  This will require some more research.  Based on an extremely unscientific and incomplete jaunt on Google, I would venture a preliminary guess that a majority of Google’s advertising revenues come from searched generics,  especially the constants that were scrubbed from their top-10 list. 

I can’t say that I know much about Google’s partnering with Mozilla (Firefox) for address bar error redirection.  However, it appears that the ACPA and the UDRP only apply to registered domain names. 

Thanks Eric.  The Firefox/Google error traffic redirect arrangement functions in a similar manner to the MSN error search gateway, whereby an individual who types an unregistered (but trademark infringing) domain name into their browser’s address-bar is served with a search engine gateway page ultimately featuring assorted paid search listings. Individual and commercial domain name registrants are obligated by the Lanham Act (and good conscience) to play by the rules, avoiding the registration of trademark intent domain-names.  The double standard (and my recent pet-peeve) is that search engines engage in the practice of taking over unregistered domain names via an arrangement with the dominant browser manufacturers, Microsoft (who display their MSN paid search network) or Firefox (who display Google in exchange for a share of traffic profits) and subsequently engage in defacto catchall-typo squatting with impunity.

It would be interesting to see if typographical variants of popular searches ‘Overstok’ or ‘Circut City’ are also scrubbed, as the examples you illustrated are.

If a user goes to Google, or any search engine for that matter, and queries a well known business name, such as BEBO, they are likely searching for one of two things: (1) BEBO.com itself, or (2) more information about the BEBO.

It would be interesting to see real studies on search engine user behavior since, yes, assertions like the one above are widely stated as fact.  If someone is aware of one brand of product, but is interested generally in that type of product, they might just as well be conducting a search using a trademark term as an indication of a more general interest.  For example, more people are probably aware of the term “Tivo” instead of the term “digital video recorder”.  I might be interested in retrieving information about devices that are, in some sense, “like Tivo”, but I may lack the vocabulary to know what sort of device that is.  One could probably find a similar brand/category overlap with the terms “iPod” and “mp3 player”.

Take fast-food or casual restaurants.  If you want a cola, how many times do you really bother to check whether the establishment sells “Coke” brand cola or “Pepsi” brand cola.  Most of the time, if you order a “Coke”, you’ll get whatever brand of cola is served by the establishment and, in fact, by saying “Coke” I’d bet a lot of customers understand that to mean “whatever cola you serve”.  Without data, of course, all we have are assertions.

We do know that consumers expect a marketplace to work in certain ways, and consumers do expect trademarks to distinguish among goods in the context of that marketplace.  It remains the case that traditionally if you want to buy a certain brand of automobile, you must head over to that part of town (and every town has ‘that part’) where there is a cluster of automobile dealers.  Or, if you want to buy a certain brand of soda at the convenience store, you must confront the refrigerator where ALL of the other brands of soda are located.  If I can get my Brand X soda, in the pink can with blue letters, on the shelf at the 7-11 next to the Coca-Cola, in the red can with white letters, then I know to a certainty that some people who walked in looking for a Coca-Cola are going to see my can of soda, know that it is not Coca-Cola, but think, “Hey, this is different, maybe I’ll give it a try.”

Now my soda example is not “bait and switch” marketing, it is simply normal competitive marketing.  Trademark law becomes relevant when someone buys my Brand X soda because they thought it was really Coca-Cola.  Trademark law is not relevant when I manage to place or advertise my product in such a way that some people looking for Coca-Cola might change their mind and buy my product instead, again knowing that they have changed their mind.

We accept this in the physical world every day.  Only on the internet do we think there is something wrong with saying to searchers, “Hey, if you are looking for X, then have you thought about maybe trying Y.”  To add another level of weirdness, we only think there is something wrong with it in the context of commercial competition.  Can you imagine applying this sort of “searchers must be corralled” concept to, say, political thoughts.  What if we had laws that required search engines to respond to queries for the names of political parties, philosophies, or candidates, to only produce advertisements or other links which promoted those political parties, philosophies, or candidates?  We wouldn’t stand for it in the United States, but there are certainly countries in this world which have arranged the internet to function that way.  What we do accept, oddly, is the same sort of regulation, but only applied to commercial activities. 

To throw in one personal example of how trademarks are used by consumers… I smoke Camel cigarettes.  When I am driving down the road and I am out of smokes, what do you think goes through my mind when I see a sign in front of a gas station that says “Marlboro - $3.99”?  What that sign says, to a Camel smoker, is “This gas station sells cigarettes”.  If I pull into that gas station and find out they ONLY sell Marlboro brand cigarettes, instead of having the usual rack with all of the other brands, I’m going to be one unhappy camper.  Now the folks who make Marlboro give those red and white signs away free to gas stations, but do you think they’d get upset over the fact that the Marlboro sign has the effect of promoting ALL cigarette brand sales?  No - but on the internet, it’s an outrage.

I can’t say that I know much about Google’s partnering with Mozilla (Firefox) for address bar error redirection.  However, it appears that the ACPA and the UDRP only apply to registered domain names.

That’s certainly true, but misses the point that laws and regulations are intended to serve an underlying purpose.  In terms of “what happens when a user does X” it largely does not matter whether a domain name is registered or not, the browser behavior is the same whether a domain name is non-existent and re-directed to paid search by the browser, or whether the domain name is registered to an individual who directs the domain name to paid search.  Either way, you end up with paid search by entering a domain name.  It’s sort of like the legal distinction between “assault” or “assault with a deadly weapon”.  If I shove a banana down your throat in an attempt to choke you to death, have I assaulted you with a “deadly weapon” or not? 

So, a banana, if used properly, can certainly be a deadly weapon.  Or it can just be snack.  Search engine advertising, on the other hand, might be fair competition or it might be unfair competition.  But we probably need to make finer-grained distinctions about HOW search terms are used by advertisers.