Databases Reference
In-Depth Information
It would be hard to imagine sponsored search not gaining some check-in type
capability.
Sponsored-search analytics.
With the increased use of check-in and mobile apps,
one would expect to see geo-location-based metrics to measure the increase in foot
traffic to brick-and-mortar stores based on sponsored-search advertisements, similar
to click-to-call metrics now.
Certainly, given the increased availability of consumer data, the future will hold
sponsored-search metrics beyond impressions, clicks, and conversions.
For example, the increasingly social aspects of Web sites, such as reviews and
consumer comments, will likely lead to sentiment-analysis metrics that measure the
tone of consumer comments about a brand or ad. This data can potentially affect how
quality score is calculated. Already, sponsored-search platforms are offering search-
ers and consumers the ability to rate ads, so integration of reviews from other sites
cannot be far behind.
With the increase in tracking devices and use of the Web via many devices such
as mobile phones, televisions, and navigation systems, advertisers will have simpler
ways to measure the combined reach of television, Web, radio, and mobile advertis-
ing in an integrated marketing communication (IMC) approach.
One can also foresee sponsored-search metrics that relate to user intentions and
mental models in an effort to understand user goals and actions. From this type of
data, one can better align advertisements or automatically adjust bids based on a
determination of commercial intent.
Development of such techniques will require complex data mining and analysis
of user interaction data to identify the strength of associations among user clicks and
actions. Current logging, analysis, and mining software can track these interactions.
However, relating them to underlying user models to understand goals and intents,
and doing so within an acceptable degree of accuracy, is not currently feasible within
the sponsored-search domain. However, in other areas, such as blogging and microb-
logging [
5
], great strides have been made in this endeavor.
It is also reasonable for planning tools to improve. For example, we can expect
keyword tools that do not rely on just historical data but gather possible keyphrases
from forum, blogs, review site comments, product reviews, customer e-mail mes-
sages, social media sites, and thesauruses. These will permit advertisers to increas-
ingly speak the language of their customers.
Based on their successes elsewhere, one would anticipate increased use of these
techniques within sponsored search to extract the intentions of users and predict
Web-searching behaviors.
Click fraud control.
Click fraud counterefforts must continue to be improved. As new
venues enter the sponsored-search domains, new potentials for click fraud also enter.
Search engines must maintain a critical point of trust. Advertisers will stop using
the sponsored system if click fraud worsens beyond some decisive point, just as mer-
chants will flee a geographical area if shoplifting becomes too bad, making profitable
operations impossible or transitioning to looting.
