Editor’s Note: This is the second installment of a search-engine-optimization case study from contributor Jill Kocher on the SEO travails of The Motor Bookstore, a retailer of automotive manuals.
Like most ecommerce businesses, The Motorbook Store relies heavily on traffic from search engines. But the site’s traffic was greatly reduced when Google recently released its so-called “Farmer” algorithm update. The company’s owner, Luis Hernandez, Jr., agreed to share his analytics and sales numbers with us, and from that Kocher — an SEO expert — is filing her installments. The first of these, “SEO Case Study: One Store’s Struggle with Google Updates,” we published last week. We thank Mr. Hernandez for exposing his business to us, and to you.
Designing an ecommerce site’s architecture is a lot like drawing a metropolitan roadmap. The biggest roads need to radiate from and connect to the most valuable pages on the site, except that the roads carry link popularity and human visitors, rather than cars. In the same way that humans can look at a map and understand which cities are more or less important, the search engines can digest a crawlable site’s architecture to identify which pages are more or less important.
In a roadmap, a major metropolis like Dallas sits in the center of a web-like structure, with large roads carrying lots of traffic flowing to and around it. Similarly, the home page sits at the center of the site’s map. Links radiating from that home page carry link popularity to the less popular pages deeper in the site. The less important a surrounding city — or page — is, the fewer and smaller roads —or links — lead to it.
For example, Richardson, Texas, as a large suburb of Dallas, enjoys broad access to Dallas and other major suburbs via Highway 75. But Parker, Texas sits off the beaten path on Farm to Market Road 2514, a local divided highway that can handle only a fraction of the auto traffic. A category page can be compared to a major suburb like Richardson, subcategory pages to minor suburbs, and product pages to outlying communities like Parker.
What Does Your Map Show?
Site architecture serves three purposes in search engine optimization:
- Provide a crawl path for spiders;
- Pass link popularity via links;
- Pass relevance via anchor text.
How individual pages are linked determines how much link popularity will flow through a site from the home page, and which pages will receive the most visibility. The more relevant navigation paths a site’s architecture offers to a destination page, the more strongly and deeply link popularity will pass through the site.
When an e-tailer sells a diverse set of products with searchable attributes, offering a single path to each product creates a limited number of pages and a limited number of internal links. In effect, the metropolitan roadmap for a site like this would look more like Amarillo, Texas, than Dallas.
Amarillo is one of the metropolitan areas in Texas, but the roads leading to and from Amarillo are significantly less complex than those surrounding Dallas. Fewer roads funnel cars to the outlying communities, just as fewer links funnel link popularity more shallowly through a site. The suburb of Canyon, Texas, similar to a category page, gets some measure of link popularity from the home page, but there are no outlying cities to be found — the product pages may not receive enough link popularity to rank or drive organic visits for valuable searches.
Navigational Path at The Motor Bookstore
Like Amarillo’s roadmap, The Motor Bookstore, a DeBary, Fla.-based retailer of automotive manuals, offers few paths through the site from the home page to a product page. Starting at the home page, the only navigational path to the Bentley repair manual for a 2001 BMW Z3 Roadster is:
- Home page, to
- “Automobile” category page, to
- “BMW Repair Manuals” sub-category page, to
- Product page, for “BMW Z3 Repair Manual: Z3 Roadster, Z3 Coupe, M Roadster, M Coupe 1996-2002.”
From the Z3 manual, if the customer also wants to find a manual for a BMW motorcycle, the only path is through the “Motorcycle & ATV” category link because vehicle make is not a navigational option. The “make” path has been omitted from the site architecture, making it harder to win searches for vehicle make keyword phrases. Similarly, landing pages for model, year and publisher are not included in the architecture. Potential customers search in large numbers for manuals by make, model and year as well as vehicle type. Whether a site’s architecture creates the types of pages that are naturally targeted toward those phrases plays an important role in whether that site will be able to win rankings and visits for those phrases.
What Are Consumers Searching For?
Using keyword research to influence the architecture of a site ensures that the pages will exist to optimize for the likely combination of parameters that consumers are searching for, and builds internal linking naturally.
For example, thousands of potential customers search organically for BMW Z3 manuals each month, but The Motor Bookstore doesn’t win those searches to convert potential customers to paying customers. Specifically, Google Keyword Tool reports that 210 people on average search Google each month for the exact phrase “BMW Z3 manual.” Google has indexed 44 pages on The Motor Bookstore that mention BMW and Z3 and manual, but none that target “BMW Z3 manual” specifically.
True, 210 searches are probably not worth optimizing for manually, but with an architecturally-optimized site, searches like these could be targeted and optimized across the site. Multiply the example of one keyword phrase worth 210 searches to thousands of keyword phrases made possible in a rearchitected site, and now the opportunity is much more interesting.
In a rearchitected site that creates landing pages for model, year and publisher, The Motor Bookstore would link to the same Z3 manual from category pages for 1996, 1997, 1998, 1999, 2000, 2001, 2002, BMW, Z3, auto and Bentley Publishers. In addition, the manual would be linked to from all of the relevant combinations of these 11 pages, such as 1999 + BMW + Z3 + Bentley; or Z3 + 2001. These page combinations provide unique pages to optimize automatically for the different combinations of search parameters, and also create stronger networks of internal links in the process.
SEO Architecture Is Difficult
To put it plainly, rearchitecture is hard and encompasses some SEO perils of its own. (See “Five SEO Mantras for Website Redesign,” my previous article on the topic.) A new platform will likely be required, or at least an internal search and faceted navigation solution such as Endeca and other companies offer. It’s not enough to create a matrix of sitemap pages for SEO and slap them up on the site. Long-term SEO and business success depend on creating architecture and content that is valuable to customers, and also communicates value to search engines.