 |
|
|
New Tune: ContextWeb Sets out to
R.O.C.K.* the Contextual Advertising World |
|
|
|
Janice McCallum |
|
|
|
20 September 2004 |
|
|
|
|
Online contextual advertising is commonly associated with
Google’s AdSense and Yahoo!/Overture ContentMatch programs
that are open to anyone who submits the required materials,
agrees to the terms and conditions, and enters their bids
in the auction-based system. However, the success of online
advertising has attracted the major players—both
advertisers and publishers—into the contextual advertising
realm. These players have a different risk/return profile,
but they are willing to pay real money to reach highly
targeted prospects. Is there an advertising technology and
services company that can deliver the level of accuracy in
contextual matching that the professional publishers and
top advertisers require? With some new venture funding,
ContextWeb is setting out to prove that they have the right
formula. |
|
Try
looking up the definition of “context” on Google. Type “define:
context” in the search box and the
first definition that is
listed comes from Princeton University’s cognitive science
laboratory: “discourse that surrounds a language unit and helps
to determine its interpretation.” This definition is followed
by 26 additional definitions. With so many meanings, no wonder
we’re all confused about “keyword matching” versus “topic or
category matching” as methodologies used by contextual
advertising networks.
For
ContextWeb, a company founded in 2000 that is launching
its ContextAd contextual ad matching service today, the
“language units” that define context are the pages of Web text
targeted for its new service. From my conversation with Anand
Subramanian, the CEO of ContextWeb, it is clear that ContextWeb
has learned from early trials and errors in the contextual ad
industry to create applications for contextual content matching
that go beyond many of today’s leading advertising services.
Rather than relying on software that "crawls" pages
periodically to determine ad matches - an ineffective approach
to examining today's dynamically generated Web pages - ContextWeb
claims its
R.O.C.K.* technology is the first to extract terms from a
Web page in real-time. R.O.C.K. matches them to terms found in
ContextAd's hierarchical taxonomy of ad topic categories,
determining the primary subject of the text by analyzing the
surrounding text. Deeper-level and more-tightly matched terms
are also extracted to provide more precise matches. For
example, R.O.C.K understands that "Boston Bruins" is a deeper
level term under the "Ice Hockey" term, which allows for both
very specific targeting and avoids matching ads to
non-contextual hierarchical categories such as "bears".
This finer control over contextualization allows for much
more efficient placement of ad inventories. Combined with an
automated bidding system designed to cater to the needs of both
premium publishers and premium advertisers seeking more
optimized value in ad placements, ContextAd is a system that
aims to help create far more value in online advertising across
the board.
Contextual advertising has been a huge hit in many
ways but it's viewed by many major advertisers largely as a
convenient means for small Web sites to earn revenue from their
editorial content. Now that the top national advertisers
and top publishers are getting into the game, it may be time
for the tools and methods used for contextual advertising to
step up to their more sophisticated needs. What are some of the
key elements that will help contextual ad networks gain a
larger share of this expanding online advertising market? A few
suggestions follow:
- Please the Publishers. For publishers with
longstanding relationships with select advertisers, the
prospect of inviting unvetted advertisers to their editorial
sites via contextual ad networks is viewed as a risky
venture. Publishers require some assurance that contextual
ads will be relevant and appropriate for their audience and
not compromise the value that they offer to their top-level
advertisers. An advertising program with an unknown return,
based on click-through rate performance with unspecified
revenue share models, is also highly questionable to the
professional publishers. If an ad network can guarantee a good contextual
match, the publisher’s risk of inappropriate and unwanted ads appearing on their site
is reduced and
the likelihood that readers will click on the relevant ads
will increase, increasing the value of return on the
advertisements for all.
- A Vertical Search Solution to Personalization. The
term “vertical search” is gaining steam as a way to describe
searches that are carried out within a network of specialized
sites, such as TechTarget for the IT community. In
advertising as well as publishing, it’s all about reaching
the right community of readers at the right time with topical
content. With trade publishing, publications are designed to
reach a community of readers who have shared interests and
have purchasing authority for a defined class of goods or
services. Put that same principle in an online environment,
where a Web site attracts a self-selected audience by virtue
of its editorial content and tone, and suddenly you don’t
need to ask each visitor if he or she is interested in dairy
farming if he or she is a visitor to DairyToday.com, for
instance. So, in essence, a contextual ad network that
operates within a group of publisher sites that serve a
community with common interests provides a pre-defined level
of personalization.
- Context + Community = $$. Combine an ad that is
relevant to the specific article in front of a qualified
reader and the results will add up to more advertising
revenue for the publisher and a better return on the ad for
the advertiser. ContextWeb hasn’t explicitly indicated that
they are pursuing a vertical search strategy as they begin
building up their ContextAd network. However, they have
targeted publishers of sites that fall into some of the most
heavily trafficked content areas, such as finance, IT and
travel. My advice to them is to stay focused on communities
of sites.
- Think about the future. In my earlier paper
How Publishers Can Profit From Contextual Advertising Models
I wrote that contextual matching technology could be applied in
new ways to extend the utility of online search and browsing.
To name just a few possibilities, contextual links on the
current page could:
- Serve as an intuitive guide to readers to help them find
additional research material;
- Locate contact information for experts on a topic;
- Retrieve a quick list of top companies that produce a
product;
- Or even notify readers of real-time activities of
interest that he or she could participate in with the click
of a button.
With mid-to-large advertisers just starting to open their
purses to fast-growing online advertising, there is still
plenty of room for new ad networks such as ContextWeb that can
provide a better experience for advertisers and publishers, and
of course, customers, too. With its multi-tiered approach
to contextual advertising ContextWeb has the opportunity to
attract many in the professional publishing industry who have
shied away from contextual advertising because of its
perception as a high risk/low return advertising program. Ready
to R.O.C.K.?
-
Janice McCallum
_________
*R.O.C.K. is ContextWeb’s acronym
that loosely stands for “Real-Time Indexing; Optimization
of eCPM; Categorization and Keyword Hierarchy”.
To top of page
 |
