A contradiction

Back in April, Google announced a set of new advertising policies, among which was allowing advertisers to choose which sites display their ads. However, this new capability potentially shuts out a "long tail" of publishers which may be:

1. Sites with objectionable content.

2. New sites without a track record, or unfairly suspected of generating fraudulent clicks.

3. Sites with low nominal conversion rates, either due to low real conversion rates, or to low enough traffic resulting in zero conversions over some time periods.

This is part of a more general contradiction: in the quest for "quality" traffic, how much of Google's revenues will actually remain intact? If you take all the junk traffic out of the equation, will we see the same revenue growth rates? Add this to increasing competition coming from Yahoo and AskJeeves and it's an open question whether GOOG's high multiple is really justified.

Google Current

Apparently, Google is in a content partnership with Al Gore's Current TV, which just debuted last night. The unstated premise of this TV network is to promote an extreme liberal agenda to young people in a "hip" MTV-style format (at which Al Gore is an expert). The buzz so far is that this project is a complete flop. I haven't watched any of the programming, but I'm wondering what kind of content Google Current is coming up with. There will surely be some egg on Google's face as the overall crappiness of the whole enterprise becomes widely appreciated. A couple of facts about Current TV:

1. Their stated purpose is to provide 3 minute bite-sized segments (no shows), because young people are so stupid they don't have an attention span longer than 3 minutes.

2. The TV listings information on their web site requires personal indentifying information - and they don't give you the listings! (The database they're building here surely won't be used in some 2008 election campaign fundraising, right?)

3. Their web site sucks ass - SLOW, NO content, NONE of the promised "interactivity."

Run GOOG Run

You know why GOOG isn't going down anytime soon? Because everybody's down on GOOG right now. People are worrying out loud that this is another dot com bubble. That's the surest indication. Also, check out the long/short ratio on ClearStation.

Google ads are dumb

As everyone knows, Google ads are pretty dumb. Keyword-driven ads are too simplistic, especially for third party, niche sites. They don't care about context or propriety. This is a huge inefficiency. And even with this inefficiency, Google's revenues are what they are, and growing at the rates they are growing. Now think about this: Google's engineers are experts in artificial intelligence. They are innovative. They will eventually figure out how to get context into the picture. Think about getting 50% click-through ratio instead of 5%. It is a fundamental change in the scale and scope of Google's operation. One reason to be bullish.

(I'm not necessarily bullish on GOOG. Click fraud remains a big threat to Google's revenues from small ad network partners.)

Cool Google game

Play Guess-the-Google. The game generates a montage of images resulting from a Google search. You have to guess the keyword that was searched. You have 20 seconds per montage (extra points for unused time), 10 montages per game. The maximum score is 400 (high scores are already at 399).

It's a pretty fun and innovative game. After playing a few times, I got a 345. My main beef with the game is that the keyword database is pretty small. Keywords start repeating after only a couple of games. So, it just turns into a game of memory (no wonder people got 399). Also, the reaction time of the flash program is pretty slow. So, if you type in a guess (or many guesses), the timer counts down while you're waiting for a response. I also noticed a bug where a montage repeats, but is associated with a different keyword. Good potential here, though.

UPDATE: Ok, I just got a 398.

New advertising strategy

Google is declaring war on internet ad servers by offering image ads and CPM (cost-per-impression), in addition to CPC (cost-per-click), bidding model. Site targeting will ensure that click fraud stays at a minimum (even though it may significantly impact revenues). Regional and local targeting gives advertisers control on a fine geographic granularity. This is especially important to the CPM model (which does not use keyword targeting).

Google is trying to take over a shrinking industry (image and banner ads), but the costs of doing so are low due to their leadership in search advertising. DoubleClick has fallen from its height in the dot com era doing exactly what Google wants to do. It will be interesting to see if Hellman & Friedman get their money's worth for their acquisition, now that Google is moving in on their turf.

This move may tarnish Google's "good guy" mystique. It will probably increase demand for ad blockers like AdBlock. Also, the image/CPM ads may not be as compelling or relevant to the user as keyword-targeted ads. I sincerely hope Google avoids the pop-up business.

Google AdSense allows site-specific targeting

From Search Engine Watch:

Google will now allow its advertisers to decide which AdSense sites will display their ads. They can make either an "include" or "exclude" list. This is an excellent move by Google to counter click fraud. However, it may shut smaller and newer publishers out of the advertising game, since they are most prone to click fraud activity. The real question is whether the revenue hit that results from allowing exclusions will be made up for by increased advertiser confidence.