Wednesday, December 3, 2008

Showstopper Google bug?

All search engines are based on the concept of Authority of Web pages and Websites. The pages and sites that are deemed to have the highest authority are retried in SERPs (Search Engine Result Pages) at the very top. Google based its success on having the best measure of Website authority - the Google PageRank algorithm. The rationale for this intellectual and technical feat and the mechanism are descibed here: The PageRank Citation Ranking: Bringing Order to the Web.
 
Let's see how good it is. The best authority on the Web for the what it says in the Bible is a copy of the bible, no? Anyone who quotes from the Bible might make a mistake, but the Bible is infallible about the Bible, I would think.
 
Here is a quote from the King James Bible:
 
If then God so clothe the grass, which is to day in the field, and to morrow is
cast into the oven; how much more will he clothe you, O ye of little faith?
(Luke 12:28)
 
The phrase "O ye of little faith" appears in the book of Matthew as well. I searched for this phrase, in quotes in Google. The first references that were quotes from the Bible appeared in the eighth page of results! Items like Time magazine articles, song lyrics and an article in one of my Web sites accounted for the first 70 or so results. Google claims it has about 36,000 results for this phrase.
 
"To be, or not to be" is the overly famous quote from Shakespeare's Hamlet. When this is searched in Google, a lone result appears in the third place in my part of the world, behind two Wikipedia articles. It is not from a Web site with the whole play, just a fragment with the soliloquy. The next result that is really from Shakespeare's play appears around position 75 again. Is there a 70 penalty for having the original text (like the 30 Penalty)? Google has about 2.5 million pages with this phrase, or so it claims.
 
Part of the problem is that we used phrases that are extremely popular. I looked up "which is to day in the field"  in Google and indeed, the very first page retrieved was from the Bible. But it was the only one actually from the Bible on that page! There were about 25,000 such pages in Google.
 
I also tried a different phrase from the same Shakespeare soliloquy, "But that the dread of something after death" - not all that famous. Not a single one of the first 10 results was a link to the original Shakespeare text. The text of the "Tragedie of Hamlet" was first listed as result number 26! There were 23,700 results for this phrase.
 
If we cannot rely on "authority" to get search engines to deliver the authentic and authoritative origins of quotes at the top of search results, then it doesn't seem to be worth that much.
 
I had better luck with this line "somewhere i have never travelled,gladly" - it is the title of a poem by e.e. cummings, who is apparently not quite as famous as the Bible or Shakespeare, and therefore he is allowed to be more of an authority on his own work than William Shakespeare or the King James Bible. For this quote too, there were about 9,000 pages claimed by Google. The entire first page of results and more were filled with links to the poem itself.
 
For "How do I love thee? Let me count the ways," the start of Sonnet 43 from Elizabeth Barrett Browning's "Sonnets from the Portuguese" the first three entries retrieved by Google were the actual poem. That's fair enough. More than that would be useless. Someone might be searching for a different page. 
 
  
 
 
 
 
 
 

Phishing - Black Black Hat

Something a bit different this time. Phishing scams and the like are probably the ultimate in  Black Hat SEO and they are a real and costly danger to e-banking and any Web sites that allow financial transactions. (see Danger! Your identity is not secure). Phishing uses a variety of techniques to direct victims from a legitimate Web site to the scam Web site, where they enter their identification information, account numbers, social security, credit card numbers etc. that are then used by crooks to steal their money.  Other schemes download appropriate  Scumware to the victim's computer. This captures their identification details and sends them to the crooks who operate these schemes.
 
Scientific American has a useful article on How to foil Phishing Scams, However, it is only useful up to a point. Most people refuse to become technically educated, and if they do, the scammers will just find new techniques to beat the system.
 
Banking and other institutions have adopted various security measures such as token devices, but these are difficult to manage for various reasons. Identiwall has a promising system based on multiple authentication using cell-phones. It can be applied for Secure ebanking, online stock brokerages and any other type of Web operation, financial or otherwise, that requires confidentiality.
 
The ultimate phishing scheme is one that can steal a legitimate Web site and redirect visitors through Cloaking, Shadow domains and similar techniques. Visitors think they are giving their identification information to the bank or stock broker, but the crook is at the other end. In principle, it is not morally or technically different from Black Hat SEO.
 
 

Monday, December 1, 2008

The nemesis of search engines

The logic of dialectics dictates that everything carries within itself the seeds of its own destruction. The basis of Web search engines is  Website Authority and Web page authority, determined by age and incoming links. That is also the basis for  The killer search engine bug. That's the short version. If you haven't figured it out: Older is not necessary better. An older description (say 1960) of how a computer works and what it can do, is not better than a new one. But the authority algorithm favors older pages. Bigger is not necessarily better either. Apple vs IBM mainframes. But authority algorithms favor the bigger Web sites. See The killer search engine bug for some of the gory details.
 
 

Sunday, November 30, 2008

How to choose and not choose a Search Engine Optimization Consultant

I saw a Website that raised the issue of how to choose a Search Engine Optimization firm. It looked very promising, but in fact, I didn't find it particularly helpful. So I wrote an article of my own. The meat of that article is 11 things you must know before choosing an SEO firm. Why 11? Because it wouldn't fit in 10.  The most important things:
1- Understand when a Search Engine Optimization Firm can't help
2- Understand what things you need to do yourself.
3- Make sure you know exactly what SEO is
4. Make sure you know exactly what this SEO consultant is going to do
 
The rest and some comments on that other site are here Selecting a Search Engine Optimization Firm
 
 

Conversion and Landing Pages

In writing about Landing Pages, I studied several articles that Google found for this search phrase. Remarkably, almost all of them were really about Conversion Rate! A Landing page, for those who do not know, is the page that is meant to be an an entry way into the Web site. It is the "bait" in commercial Web sites. Conversion rate is the measure of how many suckers visitors go from this landing page to the page where they sign up, buy something or do whatever it is you want them to do.  In this antiquated Website model, there are no search engines out there. All your traffic comes from paid advertisements or emailing, and therefore you know where the visitor is going to land and how they got there. So all the advice about landing pages, not surprisingly, is about conversion rate, which, when you think about it, is about how to get people off the landing pages and n to other pages.
 
At the other extreme from the believers in the Landing Page model, are those who believe that "every page is a landing page." Theoretically it might be true in the age of search engines, since there are less "inner" pages - the old model of the hierarchical Web site with a single front page entry point is long dead. But literally it is absurd. Not every page is a landing page. A custom 404 error page is not meant to be a landing page. A "Thank you for filling out the form" page is not a landing page, a print article page is not intended as a landing page.
 
But...
 
But even in "organic SEO" it is impossible and a waste of effort to make every page in the Web site a top draw. Some pages are more equal than others and should get more effort and thought. They may be portals such as an entry to a product list or list of links, or they may be very well written and researched articles or pages with stunnng or interesting graphics. Sometimes a page becomes unexpectedly popular. There is generally a reason that is obvious after the fact (not always).
 
But in analyzing the statistics of any website you can see that there are "landing pages" in the sense that some pages get a disproportionate about of the entry traffic to the Web site, and these are not necessarily the ones you expected to be top pages. In a site with several thousand pages, only about 550 were entry pages in a given period. Among those, the top 20 accounted for about 58% of the traffic! That doesn't mean the site would have 57% of its current traffic if it had only those pages.. The less popular pages support the more popular ones by linking to them, adding to the pagerank of the site etc. But those 550 pages are the ones that are delivering the message to visitors of that site, and the first 20 remain about the same month after month. So you need to check site statistics and exploit the opportunities. If visitors are getting to a page that isn't really what you want them to see, you have to figure out how to exploit that and lead them to a page you do what them to see.
 
How to get people to a landing page is the topic not covered in articles about landing pages. The answer is the same answer as for all search engine optimization, because SEO is optimization of landing pages:
 

On Page OptimizationKeywords and coding and content on a page.

Off Page Optimization - This usually refers to Links from other Websites.

Website SEO design - Link structure in your site, Type of software used to create your Web pages, size of your site...

 
As for improving Conversion Rate, I put most of the important tips I found plus some of my own in the Conversion Rate article, along with a lot of useful links. There are some important recommendations that can help you, some obvious bending of corners, and some very honest advice that says, "These factors can influence search rate. However, they work differently for different products and situations. Therefore, you have to try different ideas and see what works (A/B or multivariate testing).
 
A big problem you always have to deal with is that the things that make for good conversion rate often make for lousy Web pages from the standpoint of search engine optimization and user navigation. The ideal landing page has a fairly ugly and prominent message, like
 
"Buy superwidgets and gain immortality,
 riches and unlimited sex - Guaranteed or your money back"
.
And it has a single button or repeated button or link:
 
Don't Wait!
Click here to Buy Superwidgets NOW!
Hurry while they last!
 
It doesn't have a lot of Long tail  content for search engines to glom on to and it doesn't have navigation links and fancy doodads.
 
I trust you will find a lot of things to think about and use in those two articles about  Conversion Rate, and Landing Page and some ironic laughs, like the SEO "expert" who had to hire someone else to fix their landing page, and the breakthrough conversion recommendation that resulted in a huge increast in conversion rate but not a single new sale. How could that be? Think about it...
 
Meanwhile...
 
Don't Wait!
 
 
Ami isseroff