NONE: Re: ONLINE-ADS>> Ratings: What is Optimal?
Re: ONLINE-ADS>> Ratings: What is Optimal?
TLee_at_webcmo.com
Thu, 15 Oct 1998 10:20:34 -0500 (CDT)
Whether a measurement is optimal depends on whether it
provides optimal information to its users.
All these measurements were designed to provide information
for advertisers (or ad agents) to select the right web
site(s) to optimize their banner ad results.
Suppose I am an advertiser and I am looking a web site to
host my banner. Can these measurements provide the
information I need to make my decision?
1. Traffic rating: Should I put my banner on the web site
that has the highest traffic? The answer is: NO! As long as
a web site is able to deliver my banner impressions during a
certain time window, I really don't care whether the web
site ranks NO. 1 or No. 1000, because I believe all these
web sites are able to deliver enough impressions I need.
(Please remember that most of such giant web sites have more
than 50% unsold banner inventories.)
As a matter of fact, one should not have had moved such
rating method to the Internet simply because it works for
the traditional media. The rating system in the traditional
media is used mainly for measuring impressions of a network
or a program in a network. Here on the Internet, we are
proud of being able to deliver exactly the impressions an
advertiser wants, why are we bothered to pay someone to
provide much less accurate web site impression capability
that is irrelevant to my banner purchase decisions?
Here we see a good example of how people misuse statistics
to provide information that is totally meaningless to our
marketing decisions. (Another good example of such misuse of
statistics is online forecasting. Please read the article
"Lets forecast!" to see how lousy such online forecasts are
at www.webcmo.com.)
2. Demographic information: Of course, these measurements
also provide demographic information. Is this information
useful? Our answer is: yes. It at least satisfies our
curiosity about numbers. However, our concern is: Is this
information useful to our banner purchase decisions? Or more
importantly, does the demographic information provide the
OPTIMAL information for our decisions?
First off, are their panels representative to the web
population? If their panels are not representative, their
demographic information is useless. I don't want to discuss
this issue because even one has little knowledge about
statistics, he can get the conclusion by himself.
The second question is: are their panels representative to
the visitors of the web site I am going to put my banner on
it? Even we assume their panels are representative to the
online population, under MOST circumstances, they are not
representative to the visitors of a certain web site.
Therefore, if these measurements tell you the demographic
information of the people who visit a web site, that
demographic information is quite different from the people
who really see your banner ad.
The third question is: Is there any other information that
provides more accurate targeting than such demographic
information? If I decide to put my banner on a web site, it
is not because this web site ranks No. 1 in traffic. It is
not because these measurements show that its visitors fit my
targeting customers' profile. It is because people who are
interested in the contents of the pages (with my banners on
it) are also interested in my products (or services). It is
because people who are searching certain information are
also interested in my products.
On the Internet, there is other information that is much
more powerful for our banner targeting. However, so many
people are still chasing demographic information and pay a
lot for it. Please read the article "The limitations of
online demographic information" at www.webcmo.com. It lists
the reasons and cases of how to target your customers more
efficiently and more accurately by using non-demographic
information.
Suppose I know 35% of my customers are between 18-30 and I
sell game software. Please tell me: should I put a banner on
a game related web site or a search-result-return-page of
the key word "game"? Or should I go to a web site of that
these measurements show 50% of the visitors are between
18-30?
In our recent survey: Understand Advertisers' Preferences,
web marketers show (in our preliminary analysis) their
strong preferences on content based targeting and key word
search targeting.
Whether these measurements are optimal does not depend on
the sample size. If their samples are representative (as
they claim), an 1000 panel has little difference from an
100,000 panel. And there is no way to conduct a census
tracking. (BTW, even your panel is over one million, your
are still conducting a survey instead of a census.)
You make your decision whether these measurements are
optimal.
Cheers!
Tim Lee
Director of Research
http://www.WebCMO.com
A site dedicated to web marketing research.
WebCMO Digest - free newsletter about web marketing research & strategy
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsors:
24/7 Media, Internet.com and AdKnowledge
Reach more than 1 out of 3 online users through name-brand sites
17 Content Channels including News, Business/Finance, Women's Interest
The 24/7 Network - http://www.247media.com/ad_sales/ad_sbc.htm
---
Internet.com Delivers The Internet Professionals You're Looking For!
Over a million unique users, 75% have purchased online this past year.
Target your message to them, http://www.internet.com/mediakit/find.html
---
Double the effectiveness of your next web advertising campaign.
How? With AdKnowledge. For complete details, contact us.
http://ads02.adknowledge.com/SmartBanner/page/9751.4
------------------------------------------------------------------------
========================================================================
Online Advertising Discussion List To Unsubscribe send UNSUBSCRIBE
http://www.o-a.com/ to online-ads-request_at_o-a.com
