NONE: ONLINE-ADS>> DEBATE: 4/13/98 - measuring impressions, digest # 01
ONLINE-ADS>> DEBATE: 4/13/98 - measuring impressions, digest # 01
richard_at_tenagra.com
Mon, 13 Apr 1998 13:21:20 -0500 (CDT)
How should our industry measure an impression?
4/13/98 , digest # 01
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is special "debate" mailing from the Online Advertising
Discussion List. Anyone on the list is free to send in a response
to be considered for inclusion, however the core debate will be
among 16 industry experts on this subject.
You will receive this mailing in addition to your normal
Online Ads posts/digests/reports.
web site supporting list: http://www.o-a.com/
archive: http://www.o-a.com/archive/
moderator: Richard Hoy
posting address: online-ads_at_o-a.com
subscribe/unsubscribe instructions are at the end of this message
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsors:
ValueClick and ICONOCAST
Got unsold ad Inventory??? We've got advertisers!!!
ValueClick is your hands-off solution
http://www.valueclick.com/host
---
"The quality of ICONOCAST, the brevity and frequency is absolutely
spot on!" - Marc Phillips, Director, www.aptstrategies.com.au
E-mail majordomo_at_iconocast.com with "subscribe iconocast" in body
------------------------------------------------------------------------
========================================================================
Contents
1. Kate Everett-Thorp - Lot21 Interactive Advertising Group, Inc./ IAB
2. Tom Hespos - K2 Design, Inc.
3. Jim Meskauskas - Hawk Media
4. Tom Shields - NetGravity, Inc.
5. Dave Barlin - I/PRO
6. David Zinman - AdKnowledge, Inc.
7. Mark Grimes - eyescream interactive
------------------------------------------------------------------------
------------------------------------------------------------------------
<MODERATOR's NOTE:>
Hey all, only seven statements came in by the noon deadline,
so we'll just do the rest as they come in. Lets get this
thing rolling! - richard
</MODERATOR's NOTE:>
Kate Everett-Thorp
President & CEO
Lot21 Interactive Advertising Group, Inc.
Chair, IAB Media Measurement Task Force
kate_at_lot21.com
As you may expect this is a very difficult question because
what we currently want does not measure up to what we CAN do
when it comes to the measurement of ad impressions. I guess
we would first need to agree on what an ad impression is.
My definition is that an ad impression is equal to what I
(advertiser) have agreed to pay for (CPM model). In today's
world you have a variety of measures that equal an ad
impression. With that said, I am of the opinion that the
industry 'should' measure the most accurate and reliable way
possible. It is my firm belief that the industry should
promote accuracy and comparability as the benchmarks that
determine how we measure. Numerous methodologies will exist
and do. If we utilize these benchmarks, we can adapt to new
technologies and capabilities while achieving a consistency
that the advertising community will require to continue to
invest, mature and ultimately succeed online.
-----------------------------POST NUMBER 2
Tom Hespos
Media Manager
K2 Design, Inc.
thespos_at_k2design.com
I think that an ad impression should be defined as:
- The completed transfer of an ad from a publisher's
ad server, or
- The completed transfer of an ad from a third party
ad server, should such a solution be employed.
I am not aware of any technology in widespread use that
allows ad servers to determine whether an ad was actually
displayed by a browser. In the absence of such technology, I
think we should try to get as close as possible to
determining if the ad was actually seen by a user. I think
the above definition is as close as we can currently get.
In my experience, the two main arguments that usually erupt
between publishers and online advertisers are:
1) Completed transfer vs. ad request, and
2) Cached vs. non-cached activity
Personally, I am of the opinion that all activity should be
counted, regardless of whether it is cached or not.
However, before publishers can charge for cached activity,
they must acknowledge the following.
1) Cached impressions are not worth as much as non-cached
because many of the clicks on cached impressions already
show up in "standard" reporting. 2) Average click rates
across the publisher's site will dive. If everyone in the
industry charges for cached activity, the advertiser's
established click rate benchmarks will have to be adjusted
downward. 3) If there is no "cache busting standard" across
the web, advertisers lose the ability to compare the
performance of two sites on an apples-to-apples basis. The
degree to which a publisher makes attempts to defeat proxy
caching will affect the number of cached ad impressions that
are counted.
Summing up my opinion, I believe that online advertisers and
publishers need to agree on a definition of "ad impression"
that will address two issues:
1) Do we count completed transfers or ad requests?
2) Do we count cached activity?
-----------------------------POST NUMBER 3
Jim Meskauskas
Media Planner
Hawk Media
Jim_at_HAWKMEDIA.com
As a media planner/buyer, I have a less technical outlook on
this issue. I look at it the same way I do traditional
media. When reviewing traditional media and trying to make
a recommendation to a client on the "best" medium to use in
order to maximize reach over a target, I try to break all
forms of media down to impressions. The reason for this is
obvious, aside from the "apples to apples" thing. The
definition of a impression is fairly simple: one ad
exposure. You could say it is the 'atom' of the media
universe.
In traditional media one always deals with the given that
one's numbers are "about x," never "exactly x." But as
planners, we say "x" as though it were absolute. With
online, however, we are presented with the possibility of
knowing "exactly x." Or so people think.
Technological applications and their potential for precision
have created the illusion that absolute measurement is
absolutely possible. That continues to be a major promise of
the Internet and the industry itself has promoted this
image. The reality is that there is still a ways to go
before there is absolute measurement of impressions. With
caching, 3rd party ad servers, and buggy back-end software,
I have to accept a certain amount of imprecision in the
measurement of impressions.
As a planner, my ideal of the measured impression is one ad
exposure. Period. If I can marry that ad exposure to a
unique IP address, and in turn correlate that to a call to
action, wonderful! My dreams have come true. As a buyer,
just like in traditional media, I need to be sure I get a
certain amount of weight over a certain amount of time so as
to satisfy a communications goal. The medium can't be held
responsible if I leave the room during 'Seinfeld' when an
advertiser's spot runs. Or a reader skips the page my
insertion is on. Or a driver turns around to yell at the
kids when they pass my billboard. But neither can a site
expect to sell me on something like cached impressions.
Magazines don't try to sell on pass-along readership, after
all.
When it comes to online ad impressions, I'm confident with
the current definition and how it is applied as a metric.
It's easy, it's quantifiable on a basic level, and it can
make this otherwise esoteric field comprehensible to clients
and the rest of the ad industry alike.
-----------------------------POST NUMBER 4
Tom Shields
Vice President and Chief Technology Officer
NetGravity, Inc.
tshields_at_netgravity.com
One way to answer this question is to identify the
stakeholders, list their requirements, generate a list of
possible measurement methods, and then compare them with
respect to the requirements. This then should produce a
method or set of methods that represent the "best" way to
measure ad impressions for all of the given stakeholders.
In my opinion, the two major stakeholders are publishers and
advertisers. I'd venture to guess that nearly all other
interested parties, such as technology vendors, service
providers, rep firms, and auditors, will find their
interests aligned with either or both of these sides.
cost-effectiveness, maximization of revenue, and consistency
across time. Clearly a publisher wants to minimize the cost
of doing the measurement, and provide the greatest revenue
opportunity, in order to maximize profits. Also, a
publisher wants consistency in order to do meaningful site
analysis over time.
Advertisers want comparability, accuracy, and the ability to
measure multple ad media types. They need comparability to
enable buys across multiple publishing sites, so they can
compare ROI. They want accuracy in order to help determine
their actual value of their purchase, and to compare to
other media. And, they need to run banner ads, Java ads,
form ads, and other rich media types, and measure them
reasonably. In my experience, advertisers find
comparability to be the most important of the three
requirements.
In my analysis, there are three major "families" of ad
impression measurement methodologies in use today. The
first I will call "ad insertions" and refers to measuring at
the time the page is downloaded by the browser. Often a
page is dynamically constructed, and an ad inserted at this
time. The second I will call "ad downloads" and refers to
measuring at the time the browser requests a piece of media
to be displayed in the page. I believe that this family
includes methods that count before redirecting to an actual
URL, sometimes known as "302" methods. The third I will
call "ad views" and refers to the browser reporting when an
ad has been displayed successfully. This method is not
commonly in use, mostly because current browsers do not
support it for standard banner ads. Each "family" has a
multitude of measurement variations; for example, some "ad
download" methods involve cache-busting, while others do
not.
In my opinion, the above requirements mandate a method that
comes from the "ad download" family, because the "ad
insertion" family is not possible for network providers, and
the "ad view" family is not practical at present. I also
believe that the requirements mandate a very specific
exposition of the methodology that is used, because
seemingly trivial differences in measurement methodology can
result in startlingly different numbers. Finally, I believe
that one methodology should be insisted upon by the
advertisers, so that they are able to compare numbers in a
meaningful fashion across buys.
-----------------------------POST NUMBER 5
Dave Barlin
Stategic Partnerships Manager
I/PRO
barlin_at_ipro.com
In trying to reach a common understanding of what the
industry should be counting when we try to measure
impressions on the Web, we need to keep in mind that
advertisers have two goals:
1. Comparable measures of impressions across sites
2. Accurate counting of impressions across sites
I've ordered them in this order because - assuming at least
some level of basic accuracy -- comparability is the more
important goal. While there may be inaccuracies in the
Nielsen or Arbitron ratings for television and radio (as
there always are for non-census measurement), they are still
valuable resources because these measures are comparable
across broadcasters.
Given the current state of the Web, the measurement that is
most common is the ad request (as defined by the IAB).
Given that advertisers and their agencies need comparable
numbers, it is probably to the industry's advantage to
continue to move towards the ad request, despite its many
shortcomings (which I'm sure many others have described in
some depth).
However, having said this, there are huge advantages that
can be gained by reducing or eliminating the large (though
somewhat consistent) inaccuracies of ad requests. What
advertisers really want to know with their impressions
measurement - and what they've never been able to get in any
medium - is "how many times was my ad actually seen by the
end user?" The only way that we'll be able to get that is
by moving to client based measurement. Client based
measurement will allow the industry to deliver the best
possible solution: Comparable statistics that are 100%
accurate, and are unaffected by cacheing or other
distortions.
In summary, I/PRO believes we should use ad requests in the
short term to achieve comparability, but should move to
client based measurement asap.
-----------------------------POST NUMBER 6
David Zinman
Founder, Dir of Product Mgmt
AdKnowledge, Inc.
dzinman_at_adknowledge.com
Before I begin, many thanks should go to Richard Hoy for
spearheading this critical debate (applause here).
To answer "How should the industry measure an impression?",
we must first agree on objectives. For my opening post, I
will focus on trying to set a goal for the debate. In future
posts, I will address the mechanics of exactly what event
should be counted as an impression.
When I co-founded Focalink (now AdKnowledge) 3 years ago,
our goal was to make the placement of ads on the web
efficient for marketers. Ultimately, we want to make the web
MORE effective than other media. One of the strongest
differentiators of the Web versus traditional media should
be it's ACCOUNTABILITY. Therefore, any standard for
measurement should set as it's goal to measure as close to
the true number of times that an ad shows up on a viewer's
screen as possible. This demonstrates to marketers that they
will get more accurate information when they buy web instead
of other media.
Unfortunately, the IAB's recommended standard overstates the
true number of adviews. Their standard, an ad insertion,
boils down to counting an adview whenever an html page
request from a browser is processed by a server. As we all
know, however, sometimes users turn off their graphics, hit
the stop button before graphics are downloaded, or never
receive the graphics due to many of the vagaries of the
Internet. In all of these cases, users would not see the ad.
This makes impression counts for the Web little better than
those for other media, and in some ways worse! In
television, for example, if a commercial ran without the
audio, or was cut off halfway through, the advertiser would
receive a make good. Shouldn't the Web be MORE accurate than
TV?
So, the challenge before us is clear: can we muster the
resolve to set a higher standard for web measurement?
Nothing less than the competitiveness and viability of the
medium is what is at stake.
-----------------------------POST NUMBER 7
Mark Grimes
President
eyescream interactive
grimes_at_eyescream.com
What is an online ad impression?
Simple isn't it? Every time a graphic is loaded.
Oops, I forgot 15-20% visitors to web sites have their
graphics turned off. How about anytime the graphic, and/or
the text message is loaded? Oops, forgot text links are ad
impressions too...gads so are ASCII ads in email discussion
lists. What about caching? Is an ad impression the same
thing to a publisher, as it is to a advertiser?
I think an ad impression online is a vague term that must be
pre-defined and agreed upon by the publisher, advertiser and
agency.
When the unit of measurement is different for multiple
campaigns what is the standard measurement? When there are
too many variables, changing too quickly is it possible to
standardize?
I haven't any idea.
========================================================================
------------------------------------------------------------------------
This week's Online Advertising Discussion List sponsors:
ValueClick and ICONOCAST
Got unsold ad Inventory??? We've got advertisers!!!
ValueClick is your hands-off solution
http://www.valueclick.com/host
---
"The quality of ICONOCAST, the brevity and frequency is absolutely
spot on!" - Marc Phillips, Director, www.aptstrategies.com.au
E-mail majordomo_at_iconocast.com with "subscribe iconocast" in body
------------------------------------------------------------------------
========================================================================
SUBSCRIBE INSTRUCTIONS
To subscribe to the Online Advertising Discussion List,
Fill out the form at: http://www.o-a.com/
UNSUBSCRIBE INSTRUCTIONS
You may leave the Online Advertising Discussion List at
any time by sending an email message with the word UNSUBSCRIBE
(just the word) in the body or content of your message to:
online-ads-digest-request_at_o-a.com
if you get the digest version, and
online-ads-request_at_o-a.com
if you get the non-digest version.
Online Advertising Discussion List
http://www.o-a.com/
a publication of The Tenagra Corporation
http://www.tenagra.com/
