Sunday, February 13, 2005

Some guy makes a blog just for AdSense

So, some guy makes this news-aggregating blog about asbestos, just to take advantage of AdSense prices on asbestos-related keywords that supposedly reach $100-per-click (due to lawyers bidding up the prices in hopes of attracting asbestos litigation). The funny thing is, the novelty of the story may itself have contributed to this guy's success. It made Slashdot, where he must have gotten at least 30,000 hits. Even if only 10% of the people clicked on the Google ads for the hell of it, that's 3000 x $100 = $300,000. Quite a load of change. I wonder how Google is dealing with this.

Ebay complains of increasing ad prices

There's a CNN/Money article about Ebay's online ad spending.
    Bill Cobb, the president of North America at eBay Inc., said the company has seen "bubble-like" price increases for the key words that drive Web search advertising popularized by Google Inc. and Yahoo Inc.

Key phrase in the article:
    Cobb said eBay will keep 2005 marketing spending in line with prior levels

If ad prices continue their upward trend, that means Ebay will reduce their level of online advertising. I suspect that ad prices are not being driven by higher ROI, but by higher traffic levels. If this assumption is true, then ad prices are dependent on the perception that traffic is valuable. The click-fraud issue makes this perception highly questionable.

Friday, February 11, 2005

Click fraud getting more press

Michael Liedtke at the AP writes up a good overview (via Search Engine Watch) of click fraud, but the disturbing conclusion at the end of the article, in the words of Clicklab's founder, is that "advertisers are going to have to accept a certain level of click fraud as a cost of doing business." The question that never seems to get answered in these articles is: what is the limiting factor of click fraud? The answer no one wants to face is that there is none. The pay-per-click model will collapse eventually; trusted-partner networks and affiliate programs will fill the void. Lisa Wehr at OneUpWeb says it best: "Click fraud is like a big elephant standing in the middle of the living room. Everyone sees it and knows it's there, but no one is quite sure what to do about it."

Thursday, February 10, 2005

Overoptimism on GOOG

David Forrest writes up a steaming pile of you-know-what on Google, forecasting a market cap of $1 trillion, eventually. His cheerleading commentary rests on the questionable-at-best assumption that the pay-per-click model is fundamentally sound. It also represents the worst of the dot-com era hypemachine, spitting out impressive irrelevancies that make the pop investor all googly eyed (excuse the pun).

Forrest makes all sorts of unsupported short-term to long-term extrapolations like "Google creates value for so many people that failure simply isn't an option." Yes, Google creates value now, but unless it can use that value to diversify its business, it has some major problems. Google's success ultimately rests on its ability to use its current dominant position to turn itself into a diversified portal that kicks the living daylights out of MSN and Yahoo. Anything short of that and it will get eaten alive. Don't buy the hype.

Tuesday, February 08, 2005

Some suggestions for Google Maps

1. Put a travel time estimate at each step in the directions.
2. When a user zooms in, the state names should appear on the state borders, regardless of where the user is looking on the map.
3. Have a "Slide Show" option for driving directions that recenters the map every second or so, sequentially for each step of the directions.
4. Allow the user to enable informational popups that appear when they mouse over points of interest on the map.
5. Allow the user to enable icons on the map showing points of interest or services (fast food, gas stations, etc.).

Google Maps!

Google Maps beta puts all other map services to shame. You really feel like you have the whole country in the palm of your hand. There's no point in describing it, just check it out.

Sunday, February 06, 2005

Slashdot discusses Google trademark issues

The Google trademark disputes were the subject of a Slashdot discussion yesterday. Most of the discussion revolves around the trademark issue itself, rather than its implications, with some input from a French lawyer and others more familiar with the details of French/European IP law.

Denial of keyword attacks from China

From Silicon Valley Watcher...

What happened: A botnet of open proxies in China was used to conduct automated searches for a specific set of keywords, which caused click-through rates for those keywords to drop (i.e., the number of advertisement clicks per page view). Google automatically removed the ads attached to those keywords because of the low click-through. The attack was discovered by ClickRisk.

So what is the incentive to doing something like this? If you are a company with competitors who advertise on Google, you may want to deprive them of this advertising medium. I suspect these bots are created in-house since I can't imagine an independent company openly advertising such services. Google has some work to do here. First step: disregard activity coming from open proxies. No legitimate user should be using them. This policy would render useless bots such as the one mentioned here.

Saturday, February 05, 2005

French court slams Google, twice

From Search Engine Watch:
    News.com is reporting that a Paris District Court today ruled against Google in a October 2003 lawsuit filed by high-end fashion designer Louis Vuitton. Google has been ordered to pay $257,430 (200,000 euros) for trademark counterfeiting, unfair competition and misleading advertising.

The issue before the French court was that vendors of knockoff Louis Vuitton products were advertising on Google's search ads.

I wonder how much money Google is raking in from those ads and how it compares to the court-ordered payout. I suspect they're still turning a profit, for two reasons:
1. This is a highly competitive market, so ad prices will be higher than average.
2. The advertisers in question have limited advertising options, by their nature, so they must rely more heavily on Google/online ads than other advertisers.
This is the second recent case of Google losing in French court:
    In a blow to Google's keyword-bidding engine, the French court ordered the company to stop linking ads to Le Meridien-trademarked terms by Monday or face a daily fine of $194 (150 euros). The company must also cease linking ads related to Le Meridien brands within 72 hours of whenever Le Meridien notifies it of listings in violation, or face a daily fine of 150 euros. Finally, Google must pay all court fees and a fine of $2,592 (2,000 euros). Google will appeal the decision.

These cases set a worrisome precedent for Google. The French courts, at least, have decided that Google is liable for trademark violations of third parties advertising on its network. A substantial portion of their revenue must come from trademarked keywords, via paid search arbitrage with affiliate programs such as Ebay's and unauthorized vendors like the Louis Vuitton knockoffs. The legal costs do not seem to be very high so far, but if they get out of hand, perhaps Google will require AdWords member to indemnify it against lawsuits arising out of trademark disputes.

The picture is not all gloomy for Google, however. So far, the US courts seem to be a bit more forgiving, as in the Geico case. But Google must prepare itself for the legal uncertainties that come with managing such a large and diverse data set.

Friday, February 04, 2005

Spin off blog

Since I've gotten more interested recently in Google's activities, I've decided to keep a separate blog on this subject. My main blog is at Danny Taggart's Blogarama (I have transferred my Google-related posts here). My main interests are in Google's revenue model, click fraud issues, and security issues (comment spam and such). This will also include discussion of Google's competitors, Yahoo and MSN. Enjoy!

Thursday, February 03, 2005

Another problem with Google's model

Tom Foremski at Silicon Valley Watcher finds a potential problem with Google's AdWords model.

Google reports quarterly profits

GOOG gapped up early in the day but sold off considerably as the day wore on. An intraday chart shows two high-volume time periods which pushed the stock down, one early in the day and one later:



I was reading through Google's 10-Q and made some observations:
  • Google's rate of revenue growth is declining as it matures: "However, although our revenue growth rate increased in the third quarter of 2004 compared to the second quarter of 2004, our revenue growth rate has generally declined, and we expect it will continue to do so as a result of increasing competition and the inevitable decline in growth rates as our revenues increase to higher levels."
  • AdSense accounts for a greater portion of ad revenue in nine months ended 09/30/04 (50%) than in nine months ended 09/30/03 (40%).
  • Google expects AdSense revenue growth to be outpaced by AdWords revenue growth: "Although we entered into a significant new AdSense for search agreement in October 2004, the growth in advertising revenues from our Google Network members' web sites is expected to be less than the growth in revenues from our web sites for the foreseeable future."
The key question to how click-fraud will impact Google is where AdSense revenue is heading compared to total revenue. As it stands now, AdSense accounts for an increasing share of total revenue that is levelling off. We'll have to wait and see if this trend continues.

Wednesday, February 02, 2005

Good click fraud article

Mike Grehan writes (via Battelle):
    It has to be said, there are many people in the industry who hold a very cynical view about the amount of time and technology search engines are likely to employ for this type of monitoring.

    Let's face it, click fraud is bad news for the advertiser, but it’s still pouring millions into search engine bank accounts.

    So, is it feasible that they really would want to put such a concerted effort into something, which effectively, helps them make less money?

    In Jake’s case, he presented them with refined data which he’d pulled together himself. Once Google had analysed the data, they got back to him (within five days) and agreed that it was fraudulent activity and agreed to a refund.

    I have to say, in my further conversations with both Jessie and Jake, we were all agreed that, as Jake put it "there’s an insane number of PPC advertisers who don’t bother tracking. They don’t bother using unique URLs for monitoring and ROI purposes."