A shift toward topic modelling for search has been researched and documented over the past year and may prove valuable as a shift in major search engines algo’s toward providing more accurate results related to common understanding and trends. These results may lead to better result page listings for users, but will this decrease ad revenue for Google with increased organic CTR?
Topical Models may be at a disadvantage as it may limit some queries within that topic. So, a query that could (or should) be cataloged in multiple topic directories will most likely lose (or gain) value in different segments. This would depend upon the distribution of that piece of content within each topics index. But, in theory, topical search should work better than traditional keyword frequency because of semantics within topics.
But here is an example from the Helsinki Institute for Information Technology abstract which shows how a linear ad mixture of different multi-nominals may work (shown to the right).
While SERPs change depending on many variables, here is the “Judas” query from earlier this year:
Not much has changed in the News section, pulling trends (Topical Based), however results related to wikipedia listings appear and can be attributed to a more Hyperlink-Induced Topic Search algorithm. The theory of linking externally to other authoritative hubs (and this having an effect on search results) has been around for over a decade and used within Ask.com at points.
So, are topical search results being influenced by external linking to hub authorities? Can a site associate itself directly with specific “authority hubs” simply by linking therein?
If so, this would affect they way on-site content and keyword analysis is researched and produced. Seo’s must consider the topic in which they want their site (or page) to be scored within. Become an expert within that topic, craft copy related to that topic and link within those topics and only those topics.
Majestic.com has taken great steps toward providing a sites Topical Trust with its Topical Trust Flow as seen below:
Search results clustering (SRC) is a challenging algorithmic problem that requires grouping together the results returned by one or more search engines in topically coherent clusters, and labeling the clusters with meaningful phrases describing the topics of the results included in them.
The problem with topical based search results are its exclusivity when considering terms in the query and ignoring similar terms.