Google's privacy data minefield

Google's data minefield is a long piece discussing the absurdity of the DoJ subpoena for completely untargeted search data for a supposedly targeted purpose - protecting children from naughty pictures online. The broad demand for unstructured data from Google is absurd. This opinion piece from Security Focus explains in detail why.

Search Engines Double Conversion Rate

Search Engines Have More Than Twice the Conversion Rate of Other Acquisition Sources, According to WebSideStory, owner of HBX web analytics service and software. This "news" from WebSideStory should surprise no one. This report essentially says, "When people are searching for something, they are more likely to buy it".

No! You've got to be kidding!

But the very bizarre shortcoming in this conversion stats "news" is that it lumps PPC advertising (SEM) in with SEO - organic ranking, thus confusing the issue of the value of either source - or separating out which is more effective.

My take on the issue is that ALL companies should do organic SEO and test PPC in conjunction with PPC landing page split testing - to find out how effective each source is for them.

A callout quote to the right side of this press release emphasizes

"Direct navigation Still Ranks No. 1, However, Highlighting the Importance of Branding and Customer Loyalty in Converting Visitors".

More dramatic news there! If someone knows your web address or has bookmarked it, they are more likely to buy than if they have never purchased from you or visited your site.

I'm afraid that they are only speaking to giant corporate entities there, since small business cannot do significant "branding" and is excluded from this avenue of endless web traffic - as pointed out near the bottom of this release.

The select sites used in this study generate more than $3 billion in annual sales online in five categories, including apparel, toys, computers and electronics, sports and leisure, and other.

Of course that size of companies are the only ones who would be willing to pay for such obvious conclusions drawn in this report. Brilliant deductions here.

The headline of the release emphasizes that "Search Engines Have More Than Twice the Conversion Rate of Other Acquisition Sources" Then lists those "other" sources as "Internet Links", which:

Includes banner ads, affiliate marketing links, shopping comparison engines, and other referring links.

I think this release was timed to increase HBX sales for WebSideStory at a slow time of year.

Where Google may be headed: Google is dead - Jan. 25, 2006

Where Google may be headed: Google is dead in this spoof article from Business 2.0 discussing one possible future of Google. The disturbing thing is the way SEO's are all seen as a "nuisance" and an enemy to be outsmarted and that we eventually lead to the death of Google. This is exactly the type of negative press we have suffered for years and all SEO's are perceived as "Black Hat SEO's" rather than "White Hat SEO''s" regardless of the substantial number of us working, not to trick and manipulate results, but to remove roadblocks and detours to ranking well.

As I discussed in a previous post referencing a Newsweek article called "Hotwiring your Search Engine Rankings

I have also happy to see Google's chief software engineer, Matt Cutts, constantly referring to the differences between white and black hat SEO's - albeit jokingly most times.

It will be nice when we finally overcome the negative media image of optimization and optimizers so we're no longer all seen as tricksters and nuisances and recognized for the valuable business service we offer and that we provide ethical and honest functions in most cases.

The Black Hat tricksters mostly spend their time creating their own little network of high ranking splogs full of Adsense ads or writing and hawking the software that does it automatically.

It's good to know that at least we are seeing an awareness of good guys and bad guys working within the SEO industry. Those of us doing honest and useful work for clients are wrankled each time we are lumped in to the bad guy group due to negative major media portrayals like the CNN spoof.

SEO Ranking Results & Google Data Centers

Running ranking reports for clients is a standard part of an SEO's job. This week I created a position report for a client - one for which we'd made significant gains in ranking for their targeted search phrase - and proudly sent off the report to them before a scheduled conference call to discuss our progress and status.

The client sent an email upon receiving the report saying "There is something wrong with your report - we rank higher than this report claims." I went back to Google and typed in the search phrase to find it ranking exactly where the report showed them the previous day.

I explained to that client that Google has (at last count) nine data centers which serve up search results and that they were getting results from a data center in the Eastern US which showed differing results from results shown to us here in California.

The difference was substantial enough to move the client from page two to page one in the search results and therefore made a dramatic difference in their satisfaction with our work. Differences are rarely that substantial in previously observed ranking reports, so it prompted me to dig a bit deeper into the issue and I sent a note to the client.

Take a look at this forum where datacenter IP addrresses are listed in detail.

Here is an overview of a coming update to all Google datacenters expected in February or March of 2006.

So you ARE ranking better from your area of the country and that particular data center which returns results to you. Things usually update to match in all data centers, but sometimes you just do better in one area of the country than in others. If you search from each individual IP address in that list discussed in the forum linked above, you'll see different rankings and may find which datacenter shows you ranking at the bottom of page two of results.

You might also search from that new "Big Daddy 1" data center and at "Big Daddy 2"

Where I'm seeing you ranked at #17 and #18 respectively. That is a good measure of where you might expect to be when Google moves to that new algorithm for all data centers in February or March. (Of course we continue to work to achieve better results before then.)

This upcoming change in algorithm and the interestingly named servers "Big Daddy" were publicly posted on Matt Cutts blog for beta testing by SEO's (and other Google Watchers) who read him regularly.

Of course this news was a bit much for the client to digest in one chunk and he had little time to read the articles I referenced in my note above, but it was enough to assure him that I knew what I was talking about and explain the differences in my report and his own keyword searches at his end of the country. It's a bit odd to try to explain that there are "Different Googles" to a client.

The same issue cropped up later in the day when I was doing further research for a different client and found, while we were speaking on the phone, that his results differed from my own on specific query operator searches. We were using the "site:businessdomain.com" query and the "allinurl:pick-your-own-URL" query operators to limit search results and got vastly different numbers of results and rankings.

The first stunning thing in this example was that we are less than 25 miles apart in Southern California. The second shocker was that I tried simply hitting the "Search" button a second time after getting the first results page and things changed again! All of this happening in a single day makes me believe that some percolating of results is going on as Google eases into an algorithm change.

Perhaps this is not all that unusual, but in seven years of this work, I've not seen the volatility noted on January 26,2006. Are we about to have a major SearchQuake? Is Google about to split the earth and spew volcanic new results? Stand by. ;-)

SEO 101 Video from ZDNet

SEO 101 - At The Whiteboard - ZDNet The video linked above is a great place to start (for visually oriented people) to begin to get a grasp on the basics of SEO. As a search engine optimization specialist, I constantly explain to prospective clients what is involved in SEO, but have referred them to many of my SEO articles to read about the basics. I suspect I'll be sending many to the C|Net video anytime a phone call is the first contact (rather than my web contact form) I get from a prospecitve new client. I'm a word guy, and found myself preferring to the SEO 101 textual transcript instead, especially after the video presentation started with an advertisement. Hmmmm.

I imagine C|Net needs to pay for the bandwidth somehow.

Google Fights Goverment Demands for User Data

The demand for a week worth of random searches from Google (and all other top tier search engines) is getting the attention it deserves from the major media as the linked headline above at ABC News shows. All major networks and most internet news sites, are incredulous at the request by the US Justice Department for a million "random" sites from each of the top search engines databases, along with a week worth of search queries.

The furor is focusing on the wrong area though. The government claims that their investigation into user queries is intended to lead them to porn sites from innocuous search phrases in order to help them build a case for reversing a Supreme Court ruling which invalidated the Child Online Protection Act (COPA).

The fear should be that the DOJ or any other government or law enforcement agency can demand information purportedly for one purpose and use it for many other purposes. The biggest concern is that vast amounts of data could then be used by the feds as a part of their longing for "Total Information Awareness". That crazy program, headed by John Poindexter through the Defense Advanced Research Projects Agency (DARPA), was dropped by Congress like a hot potato due to public outcry opposing another overreaching hunger for data, ostensibly for Homeland Security, but clearly meant to feed Bush Administration Big Brother programs.

Search behavior, which is retained by Google and their competitors Yahoo and MSN, holds different interest depending on who holds that information. The search engines inevitably use it to increase profits by helping them to understand how users interact with their search engine and to entice users to click on ads that make them money (by clicking on pay per click ads), or in the case of Yahoo and MSN - to visit their shopping sections and navigate to more profitable parts of their portals (Google's Froogle is free and earns them no income outside of the Adwords ads.)

Of course, privacy advocates, such as Sherwin Siy of the Electronic Privacy Information Center say, "If they didn't keep and store this data they wouldn't be in this bind," ... "It highlights the potential for misuse, whether it's unreasonable search and seizure by the government or sale of the information to private companies."

Of course the commercial value of the information kept by each of the search engines is vast, but the value changes dramatically depending on who looks at that same information and what they are attempting to mine from that data. The request for a weeks worth of search queries represents a vast amount of data considering the huge numbers of searches done at Google in seven days. By my admittedly rough estimate, based on information available from online sources, it appears that requested data could amount to upward of 15 billion search queries being requested from Google.

Google made a statement to the press which characterized the DOJ request as "over-reaching", which is a huge understatement. The data may have been requested for one purpose but can then be used for literally endless other purposes. And, as Danny Sullivan of SearchEngineWatch.com has said - this information is available from other sources and can be obtained (purchased) without warrants.

That simple fact makes it appear that DOJ is simply attempting to avoid paying for the information and supports the Google claim that the request is "burdensome" to fulfill, in addition to the "over-reaching" issue. If law enforcement or government agencies were to request those records on a regular basis, the cost in retrieving, sorting, saving, storing and delivering the data could be gigantic to the search engines, more expensive to Google because they serve up more than half of all searches online - more than any other search engine.

The DOJ should at least cover the costs involved in this request and if Google eventually caves and provides the information to Justice, then they should not only get some sort of payment to cover costs, including all resources and employees required to fulfill the demands for information, but future related expenses from all requests that will come due to the precedent set in this case.

Most are looking at and talking about user privacy as the reason to deny DOJ this expensive request. Privacy is not at issue here, as Danny Sullivan pointed out repeatedly - no user information is either requested or given by MSN or by Yahoo when they responded. More attention should be paid to the "overreaching" issue claimed by Google and the "random searches" issue - which makes the entire demand for data laughable - and makes the claim of COPA enforcement doubtful.

The appearance is of the Bush administration seeking vast troves of information for use in any purpose they wish (and no doubt specific reasons having little to do with protecting children from porn.) Filtering software is more effective and sensible in protecting kids from seeing things they shouldn't see. The appearance is of a demand for data ostensibly for a laudable purpose, when it is difficult to believe that information would be of any use whatsoever to investigators seeking to protect kids from naughty photos.

This news comes as the Bush administration is being buffeted by criticism from Congress of warrantless phone taps and, unless some more targeted search data were requested, appears to be simply another part of an overreaching data mining request from an information hungry monster.