March 2010 www.infotoday.com
Worst Practices in Internet Advertising
by PETER MARSH | CTO, Apex ads. Everybody on the planet hates banner
ads that scud across the screen without
[This column lets experts in the infor- warning, where the “X” button is hidden
mation technology industry discuss the or so small that you can’t help but click
challenges and trends their businesses instead of close. Clicking on these ads in
are facing in their special niche of the no way equates to the user’s interest in
marketplace. —Ed.] the products or services being sold.
n 1988, perennial presidential can-
didate Pat Paulsen was asked by
the Seattle Post-Intelligencer for his According to the recent comScore study,
view on wildfires and the environment. only 8% of today’s internet users account
He said, “I think the major cause of forest for 85% of all clicks. Half of these clicks
fires is trees.” come from young adults in
Paulsen may have stolen lower-income brackets. There-
that line from an old Smothers fore, placing value on CTRs not
Brothers comedy skit, but I am only encourages increasingly
obnoxious ad designs, but it More appropriate? A display ad for a heartburn medication is teamed with a feature about unusual burgers on Long Island.
reminded of it whenever I hear
media industry analysts talk also shifts the focus away from
about the disappointing per- the majority of online audience juxtapositions by reviewing the com- editorial and advertising content. The
formance of internet advertis- members, most notably those pleted editorial and advertising pages following example (which comes from an
ing and how only a small per- that many advertisers and mar- before releasing them to print. This is Atex customer site) shows how important
centage of print advertising has keters want to reach. not possible with rotating banner ads on it is to go beyond simple keywords when
migrated to online. Peter Marsh websites, especially those coming from automating the delivery of internet display
I think the major cause of Relying on keywords third-party ad networks, and keywords ads alongside related articles, listings, or
poor-performing internet advertising is alone to deliver contextual advertis- alone cannot ensure that the context will other editorial content.
that most internet ads are poor. ing—Just because I’m reading a story be relevant (or even appropriate) for the The display ad appearing on the top of
An October 2009 study from comScore about airplanes does not necessarily mean site visitor. this page is for a heartburn medication.
and Starcom reports that the number of that I am planning a holiday vacation. Contextually, it is highly relevant because
web visitors who click on internet display Worse yet, if I am reading an article about the adjacent feature story is about un-
ads has dropped 50% in the past 2 years. a car crash, do not assume that I am in usual burgers on Long Island, N.Y. The
By replacing a
Another study done last year by MORI the market for auto insurance. related stories on this page pertain to
Research found that 20% of 14–32-year- There are many sites dedicated to simple keyword guides for finding the best Long Island
olds surveyed trust print advertising, chronicling bad or inappropriate link- burgers in general.
strategy with a
while only 6% trust online advertising. A ages between online news or editorial This ad would be relevant for a similar
similar Hearst study of 18–54-year-olds content and the display ads appearing more advanced text feature on the Best Burritos in Brooklyn.
found that 21% read print ads, yet only “in context” on the same page. For ex- However, it would not be at all appropriate
7% pay attention to online ads. ample, see the figure at the bottom of the for an article about gourmet restaurants
One of the most alarming (and scien- page from badcontextualads.com, where website publishers or upscale eateries.
tific) studies conducted in April 2009 by an ad for an online meat retailer ap- By replacing a simple keyword strat-
will improve the
McPheters & Co. and Condé Nast found pears on a web page for PETA (People egy with a more advanced text mining in-
that 63% of internet banner ads were not for the Ethical Treatment of Animals). likelihood that ads tegration, website publishers will improve
seen by web users. Eye-tracking software Thousands of other, often more horri- the likelihood that ads get viewed by the
get viewed. ...
determined that users’ eyes skipped over ble examples abound online, such as the right target audience members. Just as
37% of the internet ads and stopped on display ad for a grilling competition ap- important, advertisers will be convinced
less than one-third of the ads. When pearing above an article about a couple Delivering good contextual ad- that their products and services are being
these test results were weighed against who attempted to dispose of a corpse by vertising requires the use of text-mining seen in the right context.
other “probability of exposure” factors, barbecuing it. technology (such as the Atex Text Mining
the researchers concluded that a full-page In the print world, editors and layout Engine) to properly match the relevance, Ignoring the importance of multi-
color magazine ad had 83% of the value artists guard against such inadvertent tone, sentiment, and personalization of channel ad campaigns—A couple of
of a 30-second television commercial, while