Proper SEO techniques will allow humans and robots to see your site

There is always a conflict between how accessible your website data is to Humans and to Robots.  The ability to “convert” a human to finalize a purchase is paramount so keyword spammy webpages that reduce conversions are simply not worth it.  However you also can’t convert humans unless the #1 lead source to your website is being catered to as well, whether overtly or behind the scenes. 

This method of targing both the human conversion and the robotic discovery is accomplished by implementing proper SEO techniques.  Many people ask me what the “trick” to Google is.  I can summarize it very succinctly.

TELL THE TRUTH

Google can spot a fake and if you are going to rely on black hat tricks and schemes, you’re simply going to see a short-term boost in ranking which will wither on the vine.

Humans and Robots have different needs

The example on the right demonstrates a clone avoidance technique using the NOFOLLOW rel parameter on anchor text (<a href> hyperlinks).  In a traditional website we tend to let Google see EVERYTHING which is not effective.  Think of a typical brick and mortar store.  We have a nice front entrance with customer-oriented displays that are less organized but are beautiful and pleasing.  We also have a back door that opens to highly organized inventory warehouse with bare cement floors and barcoded shelving units. 

Humans should enter our website through the front door and see things like the customer service counter and the privacy policy and featured items… and the checkout aisle.

Robots don’t need to see any of this.  They aren’t going to buy anything, they don’t need to see our investor information, and they don’t need unorganized but pretty FLASH movies or glamorous pictures.  Not only can they not see them… they simply don’t care.  The diagram above illustrates how we set NOFOLLOW on portions of our website that may be visible to humans but we want the search engines to ignore them. 

Avoid Cloning through NOFOLLOW

We also want to ensure that Google indexes our website in the proper order and we channel the “juice” as concentrated as possible to our “money pages” and the hierarchies that go with that.  Take a product where the customer can navigate there in two separate paths.  They may come to my Nike yellow tank top through /Nike/Tank-Top/Yellow or through /Tank-Top/Yellow/Nike.  This creates two separate URL signatures that land on the same, exact product… effectively a clone.

To avoid this, we set a “weight” on each parameter as to its importance.  In this case we believe that more conversions will be determined by Brand and then Type and then Color.  Any other “path” to this item is “NOFOLLOW” enabled so Google will only see the one path… however the humans will see both.

Protecting your paths will ensure SEO dominance and conversions.

Ranking #1 on Google is a great objective, but dominating the web so exclusively that your competitors get starved out is even better.  This method of Exclusionary Dominance™ is the secret to many online success including our case study today of Football Fanatics that has managed to dominate the JU Dolphins T-Shirts search result.  We will delve into how they managed to accomplish this and see if we can learn from their success for our own website projects.

First we need to Google for "JU Dolphins T-Shirts".  We have chosen a very niche product name so we can see this exclusionary effect on the competiton.  This is a keyword phrase that is broad enough to have competition but specific enough to predict that a buyer typed it in and he's looking to purchase a JU Dolphins T-Shirt.

See how they dominate search results #1-#10

• Football Fanatics (FF) is the central money portal
• College Football Store is the pay per click venue
• Football Fanatics is the comparison shop channel
• eSportsMania has 2 competitive placements
• YahooSports is an FF private label store
• ShopNCAASports is another FF store
• JaxFanShop is an FF hyper targeting domain
• Note the other ShopNCAASports double tap
• Shopping.com is an FF comparison shop channel

13 out of 16 listings exclude competitors

All channels are being targeted including:

• Military Procurement (AAFES)
• Bizrate/Nextag comparison shop
• Froogle/Vast shopping directories
• Amazon/Yahoo/MSN marketplaces
• eBay auctions
• Affiliate websites (de-ranked in Google on purpose)
• Hyper targeting domains (super-focused on keywords)
• Private label (Yahoo Sports, NCAA) with Google ranking

Other techniques ensure dominance:

• Double tap stacking (two listings per natural result)
• URL rewrite ensures keyword relevance
• Verbose and repetitive descriptions (title, META)
• High density and unique keywords (META + content)
• Keyword Domain matching (JaxFanShop targets Jacksonville)

Traditional advertisers are spending in the millions to target these products and consumers.  The natural consequence of a television ad 10 years ago was to “remember” the brand or to write down a response PO box and send a letter.  Now the customer simply remembers the brand and product (not the domain name necessarily) and “googles” for it.

This causes search engine “piracy” where the traditional advertiser motivates the customer to purchase, but when they go to purchase, the top ranked websites covet the “conversion.”  This means that whomever ranks substantially #1-#10 have the highest chance of converting the sales that were funded by the other advertisers…  Effectively the top ranked sites get the majority of the benefit of the entire industry’s advertising in that topic.

In physical commercial real estate there are thousands of good “street intersections” to sell JU Dolphins T-Shirts.  On the internet, this single search result page is the ONLY PAGE ON PLANET EARTH (statistically speaking) where the competition can compete for online conversions.  This makes the value of being listed on this page high and #1-#10 dominance very exclusionary.

Atomic Data makes search engine dominance possible

Online retail is not the same as brick and mortar retail.  When a brick and mortar store launches online they fall into this biggest trap.  Take an apparel shop… when you first walk in you find a men’s department and a ladies department.  The store is physically trying to demographically segment you.

If you create a data model that matches this, you will end up with the first <xml> node being <gender> which is a highly limiting path to follow for a search engine even though it may make the most sense for a human being.  You would then add data for teams, sports, colors, sizes, variants, materials of manufacture, and many other “parameters” for this data.  To avoid 3rd normal database limitation, you would start to peel this data out into separate tables… one for colors… one for teams…one for sports.  Then you would need to create many-to-many crosslink tables.  Over time, your table count just gets larger and larger as new needs arise.

The Root Object Classification

There is certain data that “hangs” off each sub-classification.  In this example the Item class stores who the manufacturer is (because most items have manufacturers).  The Apparel class contains the style information (because style is global to all apparel objects), whereas the Shirt class contains collar styles, sleeve variants, etc.

By localizing this information to class levels, once I define a “field” for the Apparel class, all future objects that inherit from that class will inherit that field.  Any objects that do not inherit from the Apparel class will not have the field at all.

Note how different this is from a traditional 3rd normal representation of data where we would have fields like “color1” and “color2” and “color3” simply to leave enough fields available just in case we might need them for a particular product application.

Maximum Flexibility for Customer Paths

Now that our data is structured with infinite flexibility while still retaining a core hierarchy (for default navigation purposes), when a customer walks into our store, we can simply ask Google “how they sent them” to us… and what keywords they used.  Now when the customer enters our “store” we can toss all of the inventory up into the air and literally rebuild our store to match the words they used in the order they used them.  Now they can enter as “ladies yellow tank top” and we structure our product data in terms of gender first, color next and product class third… but we also can welcome customers that ask for “white womens Nike shirt” which we do by scanning for aliases of class nodes, parent classes, and other permutations of the item for maximum comfort to the customer and higher conversion rates on sales.

Know a business that would benefit from our whitepaper on how Atomic Data Modeling can make search engine optimization possible?  Download it now:

02-Atomic-Data-Enables-Search-Engine-Dominance-by-FUZION.pdf (369.99 kb)

I'm pleased to be speaking to the Miramar group of the Florida Dot Net group at www.FlaDotNet.com.  You can register for this event at the following website:  Click here to register.  I will be discussing how proper search engine capabilities start at the database level using atomic data modeling practices.  The samples of the atomic data model will include how to layer in object inheritance at the SQL Server level, utilizing some new features in SQL Server 2008 including the intrinsic Hierarcy data type and a nice overview of search engine techniques that can benefit from a highly optimized and atomic database.  I hope to see you there!

You can get a head start by reading my blog series on the topic at:

www.NielsenData.com - Atomic Data - Best Business Practices for Product Catalog Data

There are other resources that ascribe to the Atomic Data Modeling concept which you can find at:

Zimbio.com - The Atomic Data Warehouse

Wikipedia.org - Data Warehousing and the use of Atomic Data within the Data Mart

Other announcements of this event include:

• CommunityMegaphone.com
• DeveloperFusion.com
• Upcoming.Yahoo.com
• TechVibes.com

I'm pleased to be speaking to the Miramar group of the Florida Dot Net group at www.FlaDotNet.com.  You can register for this event at the following website:  Click here to register.  I will be discussing how proper search engine capabilities start at the database level using atomic data modeling practices.  The samples of the atomic data model will include how to layer in object inheritance at the SQL Server level, utilizing some new features in SQL Server 2008 including the intrinsic Hierarcy data type and a nice overview of search engine techniques that can benefit from a highly optimized and atomic database.  I hope to see you there!

You can get a head start by reading my blog series on the topic at:

www.NielsenData.com - Atomic Data - Best Business Practices for Product Catalog Data

There are other resources that ascribe to the Atomic Data Modeling concept which you can find at:

Zimbio.com - The Atomic Data Warehouse

Wikipedia.org - Data Warehousing and the use of Atomic Data within the Data Mart

Other announcements of this event include:

• CommunityMegaphone.com
• DeveloperFusion.com
• Upcoming.Yahoo.com
• TechVibes.com

It's about that time again so I'm going to be speaking once again.  Please join me at the Houston Tech Fest in Houston (naturally) Texas for my seminar on finding your Search Engine and Data "Superman" amid your "Clark Kent" business.  Being able to identify as a coder the business methods needed to get proper search engine (SEO) rankings while satisfying good design criteria an reusability is important.  This seminar will walk you through such advanced topics as:

• Atomic Data Modeling
• Fast Page Load with Highly Normalized Data
• Content Distribution Networks and Edge Caching
• SEO and SEM Techniques in Code
• Funneling "Juice" with your Web Traffic
• Comparison Shopping Syndication
• Expanding Marketing Channels through Code

Join my Houston Tech Fest Group on Facebook!