NONE: ONLINE-ADS>> DEBATE: 4/15/98 - measuring impressions, digest # 03

ONLINE-ADS>> DEBATE: 4/15/98 - measuring impressions, digest # 03
Wed, 15 Apr 1998 12:28:41 -0500 (CDT)

How should our industry measure an impression?
4/15/98, digest # 03

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:
moderator: Richard Hoy
posting address:

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


"The quality of ICONOCAST, the brevity and frequency is absolutely
spot on!" - Marc Phillips, Director,
E-mail with "subscribe iconocast" in body


More Position Statements:

1.) Laura Mitrovich - Thunder House Online Marketing Communications
2.) Mark Scheele - IMGIS, Inc.
3.) Sean Psfister - C|net

Panelists' Responses:

4.) Tom Shields - NetGravity, Inc.
5.) David Zinman - AdKnowledge, Inc.

Online Ads List Responses:

6.) Cliff Kurtzman - Number of page views closest to reality
7.) Paul Hart - Technology isn't consistent enough
8.) Mike Scanlin - MatchLogic's method is not technically possible


Laura Mitrovich
Interactive Media Supervisor
Thunder House Online Marketing Communications

Here's the bottom line. Don't let anyone tell you anything

"An ad impression should be defined as the completed
delivery of a specific graphical/text element to an active

-----------------------------POST NUMBER 2


We have an alternate for IMGIS, Mark Scheele. -- richard


Mark Scheele
Director of Software

Mark was one of the original members of the AdForce
development team. He currently oversees the design and
development of the AdForce product. Mark has over 18 years
of experience in software design and development with
special emphasis in large database design and development.

IMGIS is the provider of AdForce, the leading centralized
Internet advertising management solution. IMGIS serves more
than 750 million ads monthly on behalf of advertisers,
networks and Web sites. Founded in 1994, the company
headquarters are in Cupertino, California. Major clients
include 24/7 Media Network, GeoCities, FortuneCity, and Hachette Filipacchi New Media.

IMGIS's position statement:

IMGIS supports the IAB/CASIE standard in counting Ad
Requests. Given the large number of ad delivery systems,
the products currently available and the diversity in
product features, it becomes clear that the lowest common
denominator is the only reasonable standard to which the
industry can be held accountable.

The primary purpose of providing an "Impression Count" is to
establish a simple, baseline definition that ad serving
companies can use to communicate with Advertisers and
Publishers. For this terminology to be meaningful it must be
understood by all involved parties. If a complicated
method of counting is adopted, it may eliminate many of the
participants from accurately recording counts… or worse,
result in a comparison of apples and oranges. I've had the
luxury of reading many of the initial postings and a single
theme seems to run through many of the responses: everybody
wants accurate comparable data for Advertisers and
Publishers. In this regard, I believe simplicity is the
best policy.

This does not mean that this is the only important number.
An Advertiser may request additional data points to help
decide how best to spend ad dollars. For example, our
auditors, ABC Interactive, recently sent out a questionnaire
to assess the techniques their clients use in counting
"impressions". They have identified four categories
(similar to Tom Shields "families" of ad impression
measurement): insertions, requests, downloads, and
displays/views. This information, techniques used in
counting "impressions," along with information such as click
rates, requests/download ratios, "cache busting," and "cache
counting" techniques are all important data points. In many
cases these data points are essential for the Advertiser and
Publisher to understand what they are buying and selling.
For example, "cache busting" techniques can cause wide
swings in the "Ad Requests" numbers, independent of how they
are counted. It may make sense for a "standard" to
incorporate several of these data points when discussing
"Impression Counts."

-----------------------------POST NUMBER 3

Sean Psfister
Director of Research and Analysis

An impression is an idealized term that cuts across media
types. Generally, it's construed as "one person exposed to
one ad, one time." Just like in every other medium, we
have to agree on an exposure measure that best approaches
this ideal concept--one person viewing an advertisement on a
computer screen.

It's problematic because"best" is different for advertisers
than it is for publishers. Advertisers don't want to pay for
any uncertainty in delivery while publishers want
compensation for all the product (pages) that they produce.
Determining the appropriate measure means finding a point
where advertisers accept some inefficiency in the buy, and
where publishers bonus some inventory. It's fluid and
negotiable, not necessarily "best."

That said, any exposure measurement system should be
accurate and precise. Accuracy means that it correctly
classifies an impression as such. Precision means that it
counts correctly. For example, client-side java counters are
precise but not accurate--they can't classify because they
miss all the impressions of non-Java browsers, but they can
count actual exposures (ie the ad loaded fully). I'm not
entirely certain, but I think this means that a lowest
common denominator approach will "best" serve all

Commonality of measurement across all browser and computer
types will still not deliver comparability with other media,
however. That's because we're using the same words to
describe different concepts. CPM in print refers to cost per
thousand copies but buyers are actually purchasing an
average issue audience with an assumed exposure. The Web is
infinitely more accurate, even with caching, proxy servers,
and graphics off. And more precise.

So, while we need to arrive at some industry agreement, I
don't think it will enable true comparability with other

-----------------------------POST NUMBER 4

Date: Tue, 14 Apr 1998 10:57:33 -0700
From: "Tom Shields" <>

I think in general too many people are undervaluing the
"comparability" requirement. I'd like to explore this in
this post, and post some potential alternatives for

The current state of affairs is that basically everyone
measures impressions a different way, but still calls them
impressions. Advertisers are just beginning to catch on to
this, but they aren't willing to take the time to try to
differentiate between the different measurement types. And
they shouldn't have to - internet advertising is hard
enough, without asking if the impressions you're getting are
downloads or insertions.

One possible consequence of this is that advertisers invest
the time and energy to understand the different measurement
types, and develop their own conversion factors to enable
comparison across sites. In my opinion, this is not a
likely outcome, because of the investment required on top of
the already complicated internet advertising medium.

A more likely consequence is that advertisers begin using
advertising networks for all of their buys, because they
know that at least the measurement is consistent. However,
often advertisers want more reach than any of the current
networks can provide, so they are still forced to advertise
across networks, and therefore still need to compare results
across networks. Also, many sites don't want to belong to
networks, so they don't like this outcome. The long-term
result of this could possibly be one super-network that
measures all of the advertising for the internet. I think
this, too, is fairly unlikely.

A similar outcome would be one measuring firm that enforces
the measurement method. Every site and network would have
to contract with this firm to generate comparable reports
for their advertisers. The technology here is not really
important - it could reside on the site, or be centralized
measurement, or even be similar to an audit. My
understanding is that this is how most other media, such as
TV and radio, work. In my opinion, this concentrates too
much power in the hands of just one firm, and is detrimental
to the industry as a whole.

The best outcome, in my opinion, is if we as an industry
agree on a single standard method of counting. Then
everyone is free to implement it themselves, purchase
software that implements it for them, or contract with a
service provider to run it for them. This provides the most
options for everyone in the industry, and allows the
industry to grow in parallel.

Another possibility is to define several standards, based on
the several different current counting methods. This
amounts to letting the advertisers decide which one they
prefer. I believe that advertisers will (and should) insist
on one method for all of their buys, and the risk is that a
method will "win" that is not necessarily the best one for
the industry as a whole.

As the industry matures, I believe that new standards will
be proposed, and may supplant the one we pick (for example,
when clients can do real ad view measurement), but that
should be a gradual process, and advertisers should not be
confused by the same words meaning different things. In my
opinion, we should get rid of the term "ad impression" and
invent some new terms, like "standard ad download" or
"client ad view" that are pretty strictly defined.

In every other medium I am aware of, advertisers have
converged on one measurement method, even at the expense of
accuracy, cost, and convenience. Nielsen is almost
universally derided in the TV industry, and generally known
to be inaccurate, but everybody uses their numbers to buy
on. I think internet advertising will turn out the same
way, but we still have an opportunity to influence which
method will win. And it is in our best interest to define
the standard method ourselves, so we can ensure that it is
as practical and accurate as possible.

More on what I think this method should be, tomorrow...

-ts- Voice:(650) 655-4774
Tom Shields, VP, CTO Fax:(650) 655-4776
NetGravity, Inc 1700 S Amphlett Blvd, Ste 350, San Mateo, CA 94402-2715

-----------------------------POST NUMBER 5

Date: Tue, 14 Apr 1998 23:59:20 -0700
From: David Zinman <>

Are We Arriving at a Standard? We have now heard from three
agency media professionals on the question at hand, and they
all agree! Brad Aronson (i-frontier), Tom Hespos (K2), and
Jim Meskauskas (Hawk Media), all want the industry to count
as close to actual impressions as possible. Since our
company supports the web media placement efforts of nearly
100 agencies, I can state with certainty that very few
agencies and marketers would disagree with this position.

In previous posts, I argued for measuring as close to actual
impressions as possible, because it would differentiate the
web from other media. I also described how to do so: by
counting the very last action taken by the server when
delivering an ad. This post goes one level deeper to
describe some of the technical issues involved.

Some Background: There are two separate requests a browser
makes to display a typical page with an ad on it. First, the
browser requests an html file, which contains the text to be
displayed on the page. This file also contains a tag which
tells the browser where to go to get the ad graphic. AFTER
receiving and reading the html file, the browser then sends
a request out over the Internet for the ad graphic. The
server then responds by sending the ad graphic back to the

HTML Files vs. Ad Graphics: Most of the discussion so far,
and the controversy surrounding the IAB's recommendation for
measuring impressions, has so far centered around whether to
count the html file delivery or the ad graphic delivery as
an impression. In almost every case, html deliveries will
exceed ad deliveries by anywhere from 10-20%. Here are some
common reasons why:

(1) The viewer could hit stop or terminate the browser
session after the page has been delivered, but before the ad
has been delivered. (2) The viewer could have turned off
graphics display on their browser. This would cause the
browser to request the html file, but not the ad graphic. In
this case, the browser would instead display a short line of
alt text. (3) The ad transfer could have been interrupted
due to internet connectivity issues.

Is counting ad graphic deliveries possible? Of course.
Dozens of clients are doing it with our SmartBanner service
today, and there are other providers offering similar

Brad Aronson clearly stated yesterday that he didn't want to
pay for an impression when the ad is not displayed, and I
don't blame him. If the medium has the ability to count ad
deliveries, then why not do so? Is expanding a site's
inventory by 10-20% really worth jeopardizing the
credibility and competitiveness of the Web as an advertising

Next time, I will cover some of the caching issues
associated with counting impressions.

-----------------------------POST NUMBER 6

Date: Tue, 14 Apr 1998 13:38:37 -0500 (CDT)
From: Cliff Kurtzman <>

Here's some numbers from a single page on one of our sites
that ran a sponsor's ad banner during the entire month of

Total page views recorded in log file: 67311
Total banner graphic requests recorded in log file: 28020

After looking at what is going on in the log file, it is
clear to us that page views is far and away the more
accurate representation of exposure to give the advertiser.

The reason for the large difference noted above between page
requests and banner graphic requests is NOT because some
people have graphics turned off or they have double clicked
the link to the page. That is a trivial amount (probably
about 5%). The reason for the huge difference is that it is
easier to negate the effects of caching on the web page than
on the graphic.

I'd also assert that the number of actual ad views we are
still missing in our count due to caching outnumbers the
number of false ad views we might show because someone
didn't wait for the page to load.

We also always include a text ad link in conjunction with
the ad banner on our pages, so that an ad is served even if
the person visiting is not loading graphics. This takes
away claims for adjustment from an advertiser that asserts
that the number of non-graphic enabled visitors is not
trivial. We also find that about 30% of the click-throughs
come from the text link rather than the graphic link, so the
text link provides additional benefit to our sponsor.

If a third party ad serving company were to serve the ad
banner onto our pages and report to the advertiser just the
graphic banner requests instead of the number of page views,
it would materially misrepresent the exposure that the
advertiser is receiving.

Every method of counting impressions is going to be
imperfect, but as far as I can tell doing what you can to
prevent caching and then measuring the number of page views
which host the banner is going to give you results that are
closest to representing reality.

--Cliff Kurtzman
The Tenagra Corporation

-----------------------------POST NUMBER 7

Date: Tue, 14 Apr 1998 22:01:30 -0400
From: Paul Hart <>

Let me start by saying what a great idea for a discussion,
and follow by saying what a challenging topic for a

I have been involved in the metrics discussion for a long
time now (~ 2 years :-), having been part of the original
IAB proposal and still actively involved in discussions to
bring the IAB proposal to the next level.

IMHO several issues loom large over the ability to find a
"easy" solution.

o The technology (especially browser) aren't consistent
between vendors and in many cases buggy

o People representing a wide range of interest are involved
in the process.

o The media is moving/changing extremely fast

For the sake of keeping my emails under the size limit, I
will address the issues one at a time. I will start with the
one near and dear to my heart, technology.

I will make a blunt statement that I believe is true. No
"methodology" we choose today will work in 100% of the
cases. The original IAB proposal was done as a lowest
common denomiator, but was never intended to be a final
proposal. Everyone on the committee realized we would be
upgrading and changing as the industry moved forward.

The technology in the browsers and even proxy servers just
isn't consistent enough. Lets give ourselves a break and
realize that the technology is young and will need time to

Part of the work that I and other members of the IAB have
been doing is trying to better understand the limitation we
are facing. This hopefully will enable us to find reasonable
and acceptable solution(s).

I guess my goal is to find a (one or more) "stepping" stone
solution(s), as we influence the direction towards our final
goals*. This means working with vendors such as the browser
and proxy manufactures and letting them understand our
concerns (and theirs).


Paul Hart <>

* I will discuss this further, but I believe one of the
issues we are faced with is a lack of a clear set of goals.

-----------------------------POST NUMBER 8

Date: Wed, 15 Apr 1998 07:32:07 -0700
From: Mike Scanlin <>

History has shown that on 4/14/98 10:09 AM, Keith Pieper

>MatchLogic believes that an ad impression should be
measured >when a user's browser resolves and displays an
advertisement >on a user's screen.

It's not technically possible to know if an ad was displayed
on a user's screen. For example:

(1) After the ad transfer, but before the browser has drawn
the ad, the user clicks away or suffers a power failure.

(2) The ad could load off-screen and if the user never
scrolls, he'll never see it.

(3) You won't know if a user viewed ad came from a cache or

Because "advertisements on a user's screen" can't be
measured precisely, I prefer "completed ad transfers" as a
measurement metric. The server can measure exactly how many
times it has sent the ads.

Mike Scanlin Four Corners Effective Banners

Increase your traffic: Subscribe to our Banner Tips e-mail list:   It's Free!

======================================================================== ------------------------------------------------------------------------ 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


"The quality of ICONOCAST, the brevity and frequency is absolutely spot on!" - Marc Phillips, Director, E-mail with "subscribe iconocast" in body ------------------------------------------------------------------------ ========================================================================

SUBSCRIBE INSTRUCTIONS To subscribe to the Online Advertising Discussion List, Fill out the form at:

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:

if you get the digest version, and

if you get the non-digest version.

Online Advertising Discussion List

a publication of The Tenagra Corporation


With an archive of more than 14,000 postings, since 1996 the Online Advertising Discussion List has been the Internet's leading forum focused on professional discussion of online advertising and online media buying and selling strategies, results, studies, tools, and media coverage. If you wish to join the discussion list, please use this link to sign up on the home page of the Online Advertising Discussion List.


Online Advertising Industry Leaders:

Local SEO with Video
Houston SEO
Austin Web Design

Add your company...

Local SEO with Video


Online Advertising Discussion List Archives: 2003 - Present
Online Advertising Discussion List Archives: 2001 - 2002
Online Advertising Discussion List Archives: 1999 - 2000
Online Advertising Discussion List Archives: 1996 - 1998

Online Advertising Home | Guidelines | Conferences | Testimonials | Contact Us | Sponsorship | Resources
Site Access and Use Policy | Privacy Policy

2323 Clear Lake City Blvd., Suite 180-139, Houston, TX 77062-8120
Phone: 281-480-6300
Copyright 1996-2007 The Online Advertising Discussion List, a division of ADASTRO Incorporated.
All Rights Reserved.