Optimizing for Semantic Search
From Metadata to Microdata
Search queries are often imprecise, clunky and awkward in some scenarios.
This can make the challenge of finding and presenting users the right information, at the right time, difficult for the search engines (not to mention for those optimizing a website). While expert at solving large and generic classes of queries, search engines continue to struggle with queries that are more, well, complicated in nature. Just think of the variety of long tail, often ambiguous searches conducted by those users visiting your website; now multiply that by several million instances per minute and you start to see the scope of the problem from the vantage point of a search engine. Fear not, the tech world is in control.
Search engines have taken the lead in solving the problem with the introduction several years ago of semantic search. What is interesting about the ongoing digital conundrum is that many Web workers are being forced to rethink their SEO strategies entirely as semantic search matures (Google’s Knowledge Graph consequently will be the subject of WM’s Mastering Search column next month). That’s not always a bad thing — but the rate of change is picking up.
Anyone now in pursuit of natural search engine traffic will need a deeper understanding of users' information needs and expectations, insights into how those users interact with the information published and/or provided, and further how search engines process and index that data. It’s enough to make your virtual head spin! Or, it would be if you focused on optimizing for the engine rather than the customer experience.
Back in the Day…
For quite some time, SEO professionals were able to use their subject matter expertise and metadata mastery to help users locate their Web property from search engines. It worked pretty well (particularly for very specific search phrases), but today content and metadata are only part of what is required to achieve SEO success. In the age of semantic search, there is much more (data) to consider.
The aim of semantic search, at least for the search engines, is to improve the accuracy of results by understanding searcher intent and the contextual meaning of the query. The opportunities of such a powerful capability are immense; from analysis and prediction to interest-based recommendations. While there is definitely interest on the part of consumers — who receive a better search experience, semantic search is also opening doors for businesses seeking deeper levels of engagement with their users - but they’ll need a key to enter the new promised land of search. The name of that key — data optimization.
Over time there will be a shift in the type of content and correspond data that is being optimized. To master search from an optimization perspective you will need to focus your attention now on two key areas (as related to on-site optimization) — metadata and structured data (namely microdata). No, SEO isn’t dead — but it is evolving, and fast, towards data optimization.
Metadata Optimization: For the foreseeable future, metadata should remain a top priority for search optimization professionals. A page document’s title (read WM’s article “Title Tags of Top Sites” — http://wsm.co/ih08Hw) still seems to carry the most weight as it relates to on-site SEO factors. Ensuring that title tag/element is unique and descriptive (keyword rich) satisfies the search engines’ requirement of document-level data relevancy. While the keyword tag in particular is not necessarily relied on heavily by search engines any longer, the description may still play a role. WM’s research into “SEO Data Mechanics: Titles and Descriptions” (http://wsm.co/wHDgsx) reveals several important best practices and research into how these metadata elements are treated by the search engines.
Why pay attention to metadata? Despite the major technical advancements that come with semantic search (addressed momentarily), most engines currently use a combination of factors to determine where pages are placed on the search return list. And what’s still carrying most of the weight? You guessed it, metadata — particularly relevant data within the title tag.
Structured Data Optimization: Metadata optimization has provided a proven method (historically) to move up or shore up search result positions. The same holds true for links, but it is within structured data that we’re able to differentiate our listings on the search results.
Integrating and optimizing structured data provides another, very valuable data layer for search engines that, once processed, is attractive and appealing to search engine users. How and why? Providing details on the data being published within a Web document shows users exactly what they’ll find upon click-through which offers search engine users a better experience on the results pages overall.
Search engines’ support of Schema.org provides the single best opportunity enhance the listings that are returned with data and information useful to the searcher. When encountered, the click-through rate of these listings is higher and their capacity to shorten the distance to a conversion is much higher. Read WM’s February 2012 issue which featured an introduction to structured data for merchants (http://wsm.co/MthFfZ) that also explains many of the benefits and procedures necessary for implementation.
The use of structured data right now is more common sense than anything, with little in the way of evidence to suggest that it positively influences search position as it stands today. When optimization professionals focus on providing as much useful information about the nature of the document, they may just be considered more trustworthy than those that do not — by search engines and users. Often, that’s enough to warrant the time investment.