PDF version of this report
You must have Adobe Acrobat reader to view, save, or print PDF files. The reader
is available for free
download.
ARCHIVED REPORT:
Semantic
Search Engines
by Geoff Keston
Copyright
2013, Faulkner Information Services. All Rights Reserved.
Docid: 00021180
Publication Date: 1312
Report Type: TUTORIAL
Preview
The idea behind semantic searching sounds
radical: to fundamentally transform Web queries from today’s
keyword-driven technology to a system that detects meaning in a user’s
words and a Web page’s content. The change is already underway,
with live semantic search engines now on the Web, but the changes
coming over the next few years are in fact manageable and in many ways
predictable, providing business opportunities to organizations that
stay on top of developments and make good decisions.
Report Contents:
Executive Summary
[return to top
of this
report]
Semantic technology, an increasingly popular
alternative to
traditional keyword-driven Web search, attempts to provide results
based on
the meaning of a query.
Related Faulkner Reports |
Semantic Web Tutorial |
Web Ontology Language & SPARQL Query Language Tutorial |
Simple Knowledge Organization System Tutorial |
The course that
semantic search will take and the pace at which it will
progress depend on both technological and business factors. From a
technology perspective, semantics are difficult – the concept of
“meaning” is tough to capture in ones and zeros. The concept operates
at the intersection of fields as diverse as linguistics and
neuroscience, and it raises many of the most vexing problems in the
fields.
Organizations
looking to better make money from searching have less lofty concerns, however, and it is these concerns that might drive the search market over the
next few years. Specifically, semantic search can do the following:
- Attract
more visitors – For some search
sites, success is measured solely
in eyeballs, or the number of views on a site. For these sites,
advertising dollars are earned solely based on the number of visitors. - Improve
the quality of results – For
many search engines,
more accurate results lead to better ad targeting and sponsored search. - Carve
out a niche in specialized markets
– Not all search
engines strive to knock Google from its perch and serve as a tool for
all purposes. Instead, some search engines focus on establishing
themselves in certain fields, or focus on certain types of searching.
Description
[return to top
of this report]
Google dominates the
search
market, handling about two-thirds of US Internet searches.1
Even
experts and ordinary consumers alike use “Google” as a verb synonymous
with performing an Internet search.
Google’s rise was as much
about its skillful commoditization of
search as about delivering better results. And the growing semantic
search field faces the same combination of technological and
business concerns.
There are a handful
of semantic search engines now on the Web, including the
technology that leaders Google and Bing have included in their
services. But semantic search engines are in the early developmental
stages and have limited numbers of users, and some of the sites are not
being aggressively developed or closely maintained. A sense of
the
available
sites and the way they work can be gotten from the following list:
- DuckDuckGo
– This site divides results
into categories and offers assistance
in disambiguating search terms that can refer to more than one thing. - Hakia
– Hakia organizes search results in ways that are
meant to be more helpful than standard results lists. For instance, it
groups together results into categories, including a “credible”
category containing pages that have been reviewed by librarians. - Kngine – Provides a sort of
synopsis, which it labels a
“concept,” about the topic of the query before providing a list of
links. - SenseBot
– SenseBot provides short summaries of each item
in a results list. - Swoogle
– Swoogle’s search queries are limited only to
semantic Web pages.
Current View
[return to top
of this report]
An illustration of
the capabilities of semantic search and the work that remains to be
done can be seen in Table 1, which depicts search results from Exalead,
an engine that allows users to upload an image file and then search for
similar pictures. Table 2 provides an example of the displays that some
semantic search engines offer, which are more complex than
traditional lists of related Web sites.
Table
1. Select Semantic
Search Results (Exalead)
Image Searched |
Select Search Results | Analysis | |
---|---|---|---|
|
|
The shape of the Note, however, that in |
|
![]() |
![]()
|
In this case, the original image is simple and clear, a very basic and unambiguous depiction of a guitar. As such, the results were all very similar, including many guitars but also some ukuleles and guitar-like toys. | |
|
![]() |
The test results returned many football helmets, but also returned the three images in the column to the left, each of which has a circular shape with a protrusion in a lower corner. |
Table
2. Select Semantic
Search Results (Kngine)
Search Query |
Select Search Results | Analysis |
---|---|---|
“Guitar” |
|
Above the traditional list of top hits that might appear on any search engine, Kngine provides an entry that summarizes that topic. For instance, in response to a search for “guitar,” it provided a paragraph-long definition, brief descriptions of its place in the musical instrument family and of the various types of guitars that exist, and a very long list of “instrumentalists” that included some people who, like the drummer Lars Ulrich, are not guitarists. |
Outlook
[return to top
of this report]
Technology Outlook
Semantic search
technology is imperfect, and there are many hurdles that must be
overcome before it is good enough to attain popularity. The problems
that researchers and product designers in the field must overcome are
not all in the field of traditional information technology. Instead,
they also belong to fields such as linguistics or neuroscience.
Considering the
number of technology problems that must yet be solved and the extreme
difficultly of many of these problems, it would appear that semantic
search engines will remain in a developmental phase for the next few
years, even with companies of the magnitude of Google and Microsoft
working on the problems.
And there
are other
new search technologies that are different from both traditional
approaches and from semantic search. One example comes from Wolfram
Alpha. As described by one of the company’s representatives in a
PCQuest interview, “Wolfram Alpha is not searching the Semantic Web per
se. It takes search queries and maps them to an exact semantic
understanding of the query, which is then processed against its curated
knowledge base.”2 New
technologies such as this could steer
the semantic market in different directions, or they could undermine
the potential of semantic searching altogether.
In late 2013, Google gave
semantic search a major boost by releasing
a new search algorithm that includes semantic-like features. Google
says of the algorithm, nicknamed “Hummingbird,” that it is the
company’s most dramatic algorithm change since 2001.3
Importantly,
Hummingbird
introduces “form-based queries.”4
When a user types something like “Who is,” the search engine will start
to suggest ways to fill in the other parts of the presumed question, as
if a form were being completed. “Who is” elicits suggestions like “Who
is my congressman,” and then “who is my” elicits further suggestions,
like “who is my senator” or “who is my patron saint.” With this
feature, Google aims to provide not search results so much as answers
to questions.
Semantic technology is changing the way search results are displayed.5
Users once saw a simple, linear list of Web sites related to their
search. But now, they see complex displays with various links and
images. These displays – such as Google’s Rich Snippets and Knowledge
Graph – aim to make information easier to find and to help provide
users with more functional search results.
Semantic search will
also change search engine optimization (SEO), which is a collection of
practices used to move Web sites higher up on search result
lists. Reporting
on her conversations with several experts on SEO, Amy
Gesenhues explains that semantic searching will move the focus of SEO
from keywords to “customer engagement.”
In this new environment,
information from sources like social networks will be
considered to capture true meaning and relevance beyond simply
matching keywords. But
this process will be long and gradual. Gesenhues quotes search expert
David Amerland as saying that “[t]his is not something that can or will
happen at
the drop of a hat…It requires time and
commitment to building a relationship with influencers and sharing with
them content that is of real value to their network.”
Business Outlook
Even when (or if)
the technological hurdles facing semantic searching are overcome, the
question will remain whether the approach is preferable to today’s
keyword-driven approaches. After all, by most standards, today’s Web
users are already able to quickly find the information they need almost
every time they look.
Semantic search
engines may carve
out niches in specialized markets. A search engine can make a great
deal of money without making a dent in Google’s or Bing’s businesses.
Even a small portion of the search engine market offers a large
customer base. Web sites that attract a fair number of eyeballs can
still succeed. And attracting eyeballs is not the only way to turn a
profit. Many search engines aim to provide more accurate results for
better ad targeting and sponsored search. Thus the way for these
companies to improve profits is to improve the quality of their search
results.
Online retailers are
also using semantic search. Most notably, Walmart created Polaris,
which lets users search the company’s online store with semantic
technology. Polaris was developed after Walmart acquired semantic
search company Kosmix.7
Standardization Efforts
In 2011, Google,
Yahoo!, and Bing announced
the launch of schema.org, in which
they agreed on a standardized language for semantic markup on a Web
site: microdata.8 In
the Google announcement, Kavi
Goel and Pravir
Gupta, of Google’s search team, said, “Historically,
we’ve
supported three different standards for structured data markup:
microdata, microformats, and RDFa. We’ve decided to focus on
just
one format for schema.org to create a simpler story for webmasters and
to improve consistency across search engines relying on the data.”9
The schema.org
site “provides a collection of schemas, i.e., tags, that
webmasters can use to markup their pages in ways recognized by major
search providers. Search engines…rely on this markup to improve the
display of search results, making it easier for people to find the
right web pages.”
Recommendations
[return to top
of this report]
Assess the Potential
Impact of Semantic Searching
The amount
of time and money an organization invests in semantic search should be
proportional to the technology’s potential benefits.
The effects of semantic searching will be felt in different ways, and
to different degrees, from one industry to the next. Certain
applications of the technology may improve dramatically, while others
will not develop beyond the experimental stage. So some
organizations could be forced to make significant changes while others
will not need to adjust at all.
Fortunately
for organizations with an interest in semantic search, there are easy
and reasonably effective ways to monitor the progress of the market and
technology. The simplest is to go to the many sites that are now on the
Web, many of which are in beta, and periodically test the results they
yield on query terms relevant to the industry.
Stick with the Big
Players
It is
unlikely that a new player will emerge from nowhere to rule the
semantic search market, like the familiar tale of a Web empire built in
a dorm room. The search market is already well-formed and dominated by
well-funded players, like Google and Microsoft, who are at the
forefront of new developments.10
And these companies can
easily swallow up any small competitors that threaten their positions
or that have useful technology, such as Microsoft did when it bought
semantic search company Powerset for $100 million, a move that formed
the foundation of its Bing launch.
Consider Displays
Today, savvy
Web marketers tune their sites and online databases to the conventions
of Google and other popular search sites. In the future, they can tweak
this tuning to achieve favored positions or more effective
presentations on semantic search engines. Organizations can begin this
tweaking process by searching for their own company names and brands on
the semantic search engines now available.
At this
time, however, it would be premature to make any significant changes
based on these results. Instead, this data will be useful for putting
certain issues on the table; for instance, an organization might
consider the ways that semantic search engine results might differ from
Google results, and consider how to craft a design that works
effectively on both. As the market shifts toward semantic searching, if
this change indeed occurs, organizations can gradually orient their Web
designs toward these new search sites.
References
1
comScore qSearch. November 2013.
2 Ahmed, M. Today’s users want
answers, not innumerable
results. PCQuest. Apr 12.
3 Benci,
R. How
semantic search is killing the keyword. iMedia Connection. Dec
13.
4 Starr,
B. 5 ways to unlock the benefits of semantic
search. Search Engine Land. Nov 13.
5
Starr, B. 10 reasons why search is in vogue: Hot trends in semantic
search. Search Engine Land. Sep 13.
6
Gesenhues, A. Google’s Hummingbird takes flight: SEOs give insight on
Google’s new algorithm. Search Engine Land. Sep 13.
7 Perez,
S. In battle with Amazon, Walmart unveils Polaris,
a semantic search engine for products. TechCrunch. Aug 12.
8 Charles
D. “Semantic Search and the
Future
of Search Engines.” Catalyst. August 8, 2011.
9 Eric
Franzon. “Google, Yahoo!
and Bing Announce Schema.org.”
semanticweb.com. June 2, 2011.
10
See:
P. Krill.
“Microsoft talks up semantic search ambitions.” NetworkWorld. January
2010.
Boyle, R.
What is Google’s semantic search. POPSCI. May 12.
Web Links
[return to top
of this report]
Bing: http://www.bing.com/
Exalead Image Search:
http://www.exalead.com/search/
Google: http://www.google.com/
Google Rich Snippets: https://support.google.com/webmasters/answer/99170?hl=en
Google Knowledge Graph: http://www.google.com/insidesearch/features/search/knowledge.html
Hakia:
http://www.hakia.com/
Kngine: http://www.kngine.com/
LibConf.com: http://www.libconf.com/
Schema.org: http://www.schema.org/
SenseBot: http://www.sensebot.net/
Swoogle: http://swoogle.umbc.edu/
Wolfram Alpha:
http://www.wolframalpha.com/
About the Author
[return to top
of this report]
Geoff
Keston is the author of more
than 250 articles that help
organizations find opportunities in business trends and technology. He
also works directly with clients to develop communications strategies
that improve processes and customer relationships. Mr. Keston has
worked as a project manager for a major technology consulting and
services company and is a Microsoft Certified Systems Engineer and a
Certified Novell Administrator.
[return to top
of this report]