The Google riddle: 'organized', 'useful' but impossible to comprehend

The Google riddle: 'organized', 'useful' but impossible to comprehend

Summary: Like the company he represents, Matt Cutts wants to be helpful. His post “More info on PageRank” at “Gadgets, Google & SEO” offers:"Every few months we update the PageRank data that we show in the toolbar, and every few months I see a few repeated questions, so let me take a pass at some of them.

SHARE:
TOPICS: Google
1

Like the company he represents, Matt Cutts wants to be helpful. His post “More info on PageRank” at “Gadgets, Google & SEO” offers:

"Every few months we update the PageRank data that we show in the toolbar, and every few months I see a few repeated questions, so let me take a pass at some of them."

While Google puts forth that it strives to “organize the world’s information” so that it is “useful” and “accessible,” even Cutts becomes exasperated by the Pandora’s box that his company embodies. 

His post, predictably, brought on a deluge of comments seeking more clarification, more knowledge on the black box that is Google.

JonMu said: 

- Do you re-scale pagerank values from time to time? (too many PR9s or PR10s)
- How does link-devaluation play into the PR algorithm?
- Are penalties reflected in PR? How do sites with penalties pass PR?
- What is the reason you’re exporting PR values at all?

Jan said:

I’d love to see PageRank to be more accurate on highly dynamic websites. ATM it seems PR inheritance is based mostly on URL-depth (delimited by / ) instead of true link-depth. Our mainsite has PR4 while links to dynamic pages from that page usually have PR0 at best. The target-page content is always static we’re just using mod rewrite to hide highly parameterized URLs.

Cutts chimes in to the information thirsty chorus:

Oy, enough PageRank questions! My advice is not to obsess about PageRank too much; it is one of more than 100 different factors in how we score documents.

Cutts fans push back.

Bob said:

So how about letting us know what the other 99 factors are so we can stop obsessing about it? Haha

Harith said:

You keep talking about PageRank her on your blog, on forums and webmaster conferences. And now you ask us not to be obsessed about PageRank?

C’mon Matt. Give us a break.

Tedd Sullivan said:

glad you explained PR and PR Obsession. But the one thing that seems obvious when you look at the datacenters in a short time is that they are not all even similar in the SERP they return for some searches. So there has to be huge difference not just in PR, SITE: but in the actual indexes.

Google CEO Eric Schmidt acknowledged at the Search Engine Strategies conference last August:

the precise formula of ranking is so proprietary only a small number of people actually in the company know it. And I've chosen not to even know it.

Google’s paying AdWords customers are also left in the Google black box dark.

Google’s cheery “Welcome to AdWords” marketing speak FAQ “Concerned about costs? Don't worry—AdWords puts you in complete control of your spending” does not put forth a realistic portrayal of the uncertain nature of the Google dynamic, competitive auction pricing system.

The marketing FAQ says:

Set your budget
There's no minimum spending requirement--the amount you pay for AdWords is completely up to you. For example, you can set a daily budget of five dollars and choose to pay five cents each time your ad is clicked.

 

Avoid guesswork
We provide keyword traffic and cost estimates so you can make informed decisions about choosing keywords and maximizing your budget.

 

Pay only for results
You're charged only if someone clicks your ad, not when your ad is displayed.

 

Avoid guesswork? Google’s AdWords Help Center, suggests the (expensive) Pandora’s box that is Google search advertising: 

Can I see what my competitors are bidding?

No. We don't reveal individual data for any of our users. Remember that your bid alone doesn't dictate whether or not your ads will perform. We also take into account the quality of your keyword-targeted ads. If your keyword requires a higher maximum cost-per-click (CPC) in order to trigger ads, we'll let you know by moving your keyword to an 'inactive for search' state. We'll then list the minimum bid needed to activate your keyword.

How do I choose my maximum CPC?

To help you decide a maximum CPC amount, click 'Want to purchase the most clicks possible?' Our system will then estimate, based on your current keyword list, the maximum CPC needed to place your ads at or near the top position on a search result page. If you want your ads to appear among the top positions for your current keyword list, fill in the recommended amount in the appropriate field and click Continue.

If the recommended maximum CPC is too high for your budget, fill in a maximum CPC you're comfortable with and click 'View Traffic Estimator.'…

Remember, the Traffic Estimator is only intended to give you a general forecast of your campaign performance. Due to various reasons, including the dynamic nature of the search market, your ads may perform differently than the Traffic Estimator predicted. Once your ads are active, you can adjust your maximum CPC…

ALSO SEE: "Google AdWords: soon over priced with poor ROI"

Topic: Google

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

1 comment
Log in or register to join the discussion
  • Neural Network; Federated Search; Death of The American Legal

    There are two search strategies to discuss. Neural Networks are used to fly airplanes and they work by learning. Boeing will have a pilot take a plane up, turn on monitors and let the computer learn from what the pilot does. Eventually the computer can fly the plane steadier than the pilot. But it is not possible to answer why the computer is able to do so. There is no algorithm created that can be studied with this form of AI.

    In business, neural networks are not often used because the ability to answer why a conclusion was reached or a recomendation made is desired. It is not just quality but the documentation of how that quality was reached that is wanted. In otherwords the ability to audit the result is just as important as the result itself. A different AI technology called Expert Systems might be used instead.

    If you look at the work being done in Seattle on Federated Search or at the meta search engines like dogpile.com you might be able to see the Future of Software Design.

    Here not only is Google data used but so are other search engines and the data is grouped in ways that a human expert (a librarian) may have.

    The application of something like Dogpile.com in a government agency or business is awsome.

    Say the organization subscribes to several private data banks. Perhaps Lexis/Nexis, perhaps WSJ, perhaps its own bank of documents. Now not only is freely diseminated Google data searched but also data that is private or confidential and the presentation of that information can be explained. You can ask why. That makes for a satisfying search.

    Lets think about how proper societies will benefit.

    In Washington State a law was changed that might have done away with afirmative action. The Governor put out a statement that even though affirmative action was by this law revoked, there were at least 14 or more state and federal laws that maintained the status quo. And those were just the ones he had in his head!

    This did not happen by accident. Every so often a legislator proposes that all the rules and administrative codes involving say employment be organized so that employers and employees might know what is correct behavior on afirmative action.

    The regulations are scattered all over the place, organized under health care, or education, or occpational health, or tax or discrimination or privacy, whatever - you get the idea. It takes a person years and years to develop the experience to waid through all of that. It takes a lawyer.

    Unfortunatly the wrong doers like it this way because it allows unethical behaviour to go more easily unquestioned by citizen employees and customers. The lawyers like it this way becuase, unlike other professions where the pace of change requires constant relearning and no one has relevant experience beyond three to five years, they can draw from years of experience, perhaps 20 or 30, when the conflicting laws were written.

    The legal profession is an odd one in that every citizen is presumed to know the law but you are a fool to represent yourself. With Google (possibly a neural net) and Federated Search (likely Expert System) that dynamic is changed. The citizen can be as expert on the law as the seasoned lawyer. No other profession is so profoundly altered by search technology.

    Frank L. Mighetto CCP
    mighetto