2. CSS and Semantic Markup
CSS is commonly mentioned as a best practice for general web
design and development, but its principles provide some indirect SEO
benefits as well. Google used to recommend keeping pages smaller than
101 KB, and it used to be a common belief that there were benefits to
implementing pages that were small in size. Now, however, search engines
deny that code size is a factor at all, unless it is really extreme.
Still, keeping file size low means faster load times, lower abandonment
rates, and a higher probability of being fully read and more frequently
linked to.
CSS can also help with another hotly debated issue: code to text
ratio. Some SEO professionals swear that making the code to text ratio
smaller (so there’s less code and more text) can help considerably on
large websites with many thousands of pages. Your experience may vary,
but since good CSS makes it easy, there’s no reason not to make it part
of your standard operating procedure for web development. Use tableless
CSS stored in external files, keep JavaScript calls external, and
separate the content layer from the presentation layer as shown on
CSS Zen Garden, a site
that offers many user-contributed style sheets formatting the same HTML
content.
Finally, CSS provides an easy means for “semantic” markup. For a
primer, see Digital Web Magazine’s article, “Writing Semantic Markup”
(http://www.digital-web.com/articles/writing_semantic_markup/).
For SEO purposes, only a few primary tags apply, and the extent of
microformats interpretation (using tags such as <author> or <address>) is less critical (the engines
tend to sort out semantics largely on their own since so few web
publishers participate in this coding fashion, but there is evidence
that it helps with local search). Using CSS code to provide emphasis, to
quote/reference, and to reduce the use of tables and other bloated HTML
mechanisms for formatting, however, can make a positive
difference.
3. Content Uniqueness and Depth
Few can debate the value the engines place on robust, unique,
value-added content—Google in particular has had several rounds of
kicking “low-quality-content” sites out of its indexes, and the other
engines have followed suit.
The first critical designation to avoid is “thin content”—an
insider phrase that (loosely) refers to content the engines do not feel
contributes enough unique material to display a page competitively in
the search results. The criteria have never been officially listed, but
many examples/discussions from engineers and search engine
representatives would place the following on the list:
Thirty to 50 unique words, forming unique, parsable sentences
that other sites/pages do not have.
Unique HTML text content, different from other pages on the
site in more than just the replacement of key verbs and nouns (yes,
this means all those sites that build the same page and just change
the city and state names thinking it is “unique” are
mistaken).
Unique titles and meta description elements. If you can’t
write unique meta descriptions, just exclude them. Similarly,
algorithms can trip up pages and boot them from the index simply for
having near-duplicate meta tags.
Unique video/audio/image content. The engines have started
getting smarter about identifying and indexing pages for vertical
search that wouldn’t normally meet the “uniqueness” criteria.
Note:
By the way, you can often bypass these limitations if you have a
good quantity of high-value external links pointing to the page in
question (though this is very rarely scalable) or an extremely
powerful, authoritative site (note how many one-sentence Wikipedia
stub pages still rank).
The next criterion from the engines demands that websites “add
value” to the content they publish, particularly if it comes from
(wholly or partially) a secondary source.
3.1. A word of caution to affiliates
This word of caution most frequently applies to affiliate sites
whose republishing of product descriptions, images, and so forth has
come under search engine fire numerous times. In fact, it is best to
anticipate manual evaluations here even if you’ve dodged the
algorithmic sweep. The basic tenets are:
Don’t simply republish something that’s found elsewhere on
the Web unless your site adds substantive value to users, and
don’t infringe on others’ copyrights.
If you’re hosting affiliate content, expect to be judged
more harshly than others, as affiliates in the SERPs are one of
users’ top complaints about search engines.
Small things such as a few comments, a clever sorting
algorithm or automated tags, filtering, a line or two of text,
simple mashups, or advertising do not
constitute “substantive value.”
For some exemplary cases where websites fulfill these
guidelines, check out the way sites such as CNET, Urbanspoon, and Metacritic take
content/products/reviews from elsewhere, both aggregating
and “adding value” for their users.
Last but not least, Google has provided a guideline to refrain
from trying to place “search results in the search results.” For
reference, look at the post from Google’s Matt Cutts, including the
comments, at http://www.mattcutts.com/blog/search-results-in-search-results/.
Google’s stated feeling is that search results generally don’t “add
value” for users, though others have made the argument that this is
merely an anticompetitive move.
Sites can benefit from having their “search results” transformed
into “more valuable” listings and category/subcategory landing pages.
Sites that have done this have had great success recovering rankings
and gaining traffic from Google.
In essence, you want to avoid the potential for your site pages
being perceived, both by an engine’s algorithm and by human engineers
and quality raters, as search results. Refrain from:
Pages labeled in the title or headline as “search results”
or “results”
Pages that appear to offer a query-based list of links to
“relevant” pages on the site without other content (add a short
paragraph of text, an image, and formatting that make the
“results” look like detailed descriptions/links instead)
Pages whose URLs appear to carry search queries (e.g.,
?q=miami+restaurants or ?search=Miami+restaurants versus
/miami-restaurants)
Pages with text such as “Results 1 through 10”
Though it seems strange, these subtle, largely cosmetic changes
can mean the difference between inclusion and removal. Err on the side
of caution and dodge the appearance of search results.