Eyeblaster Analytics Bulletin

Issue 3 | Trends in Conversions

March 2009
Background
In a time of recession,
accountability of online
advertising becomes more
critical than ever. Whether you
are pursuing a direct response
initiative or looking to achieve
measurable branding objectives,
the focus is on the process of
`conversion’: that moment when
an individual exposed to an ad
performs a desired action.1
This issue of the Analytics
Bulletin focuses on the role
of conversions. Given the
accountability and measurability
of online media, what measures
can we take in order to maximize
conversions at zero or minimum
cost? Eyeblaster conversion
tags give the ability to track
user’s activity on the advertiser’s
site after viewing or clicking on
an ad, but this is only one of a
myriad of methods that can be
deployed to ensure a better yield
of success.
Pg 193 – Marketing Sherpa’s 2008
1 Pg 193 - Marketing Sherpa's 2008
Summary of findings
• Click-thru Rate vs. Conversion Rate:
Worldwide (WW) ’08 CTR is 0.14%
while WW’08 Conversion Rate stands at
0.32% - more than double.
• Correlation between Format Type and
Conversion Rate: Rich media ad formats
increase the tendency of users to convert
by 2.5x.
• Impact of Data Capture on Conversion
Rate: Lowering the number of phases the
user needs to go through by bringing the
conversion mechanism into the banner.
You are 5x more likely to convert in a
banner than you are on a site.
• CTR vs. Data Capture: Comparisons
across unique data reveals users are
nearly 8x more likely to convert in a
banner than click on it and about 34% of
unique users who interact will fill in form
details within a banner.
• Automatic Optimization and its impact
on Conversion Rate: Conversion rate
of campaigns using an optimization
algorithm is significantly higher (1.25x)
than the average conversion rate of
rich media and practically double the
global average conversion rate for all ad
formats.
• Search as a conversion metric: One can
measure an increase in search activity
as a result of display advertising that is
sustained for a sizeable amount of time
after the display campaign has finished.
• Behavioral sequencing: Consumers
exposed to targeted messages in
sequence are more likely to move
through a conversion sequence.
 Table of contents

Research Methodology ............................................................................. 4

Results in Detail ........................................................................................ 4
Post-Click vs. Post-Impression ................................................................. 5
Improving the conversion process with Data Capture ................................ 6
Automatic Optimization mechanism and its impact on conversion rate ..... 10
Search within the conversion process ....................................................... 11
Benchmarks
Appendix 1: North America .................................................................... 15
Appendix 2: Europe ................................................................................ 16
Appendix 3: Spain .................................................................................. 17
Appendix 4: APAC .................................................................................. 18
Appendix 5: Taiwan ................................................................................. 19

3 Analytics Bulletin Issue 3 | March 2009

 Research Methodology
This study is based on data of more than 2,000
campaigns representing 29 billion impressions collected
between Jan ’08 and Dec ’08 covering all sectors.
These campaigns represent advertisers using Eyeblaster
conversion tags. Conversion rates were based on
total conversions with no differentiation of the type.
All calculations were performed on daily delivery data
aggregated to the ad level.
Thinking about conversions…
It could be argued that
a true conversion would
be where a consumer
has made a monetary
transaction directly with the
advertiser, i.e. purchased
something. The focus here
is on conversions where
the conversion means
‘showing intent’ – such
as through data capture
– indicating a consumer’s
desire to purchase.
It is worth noting that a total conversion
could happen directly on an advertiser’s site
but possibly more so on a reseller’s site or in
some other place, such as offline or in-store.
The true effects of digital advertising against
this overall conversion process (included
within or in isolation from other media) are
not considered here due to current tracking
methods. For the purpose of this document,
we shall limit conversions to those indicating
intent within the ad itself or direct on an
advertiser’s site, appreciating this in itself is
only showing a slither of the actual overall
conversion rate and ROI.
4 Analytics Bulletin Issue 3 | March 2009
Results in detail
Click-thru Rate (CTR) vs. Conversion Rate
CTR is a valuable metric to measure the amount of users
who wish to visit the advertiser’s site, but this in itself is not
a true conversion. The common perception is that CTR
indicates campaign success; the following analysis shows
that low or high CTR does not necessarily correlate with
the campaign’s overall objectives. Conversion tags placed
on the advertiser’s web site on the other hand, track actual
user activity that the advertiser wishes to track.
In observing the CTR metric alone, we expect all
consumers to react immediately and overlook consumers
who converted after being exposed to – and possibly
interacting with – the ad.
In order to see the full picture we should sum both types
of conversions.
CTR vs. Conversion Rate
0.35%
0.32%
0.30%
0.25%
.5
x2
0.20%
0.14%
0.15%
0.10%
0.05%
0.00%
CTR Total Conversion Rate
Eyeblaster Jan.2009
Figure 1 Source: Eyeblaster Research 2009
In looking at the overall lift in the conversion rate, it
indicates that a higher percentage of consumers take in
the advertising message, then require time to digest the
information, possibly clarify it from other sources, before
finally being in a position to actually convert.
Conversion rate provides more accurate and precise
measurement of the campaign’s success and global
data reveals that by focusing on CTR alone, we are
missing 2.5x the actual data an advertiser requires to
correctly analyze the overall consumer journey.
 Post-Click vs. Post-Impression
Correlation between Format Type and
Conversion Rate
With pressure mounting on trying to ascertain the most
cost-effective online advertising, it would seem that high
volume standard display advertising would naturally be a
first choice. However, when we look at the true conversion
rates for the 3 most popular ad format types online, we
clearly distinguish a higher rate for rich media advertising.
Conversion Rate by Formats
Standard advertising plays an important role in enhancing
reach and frequency of the entire media mix (Rich,
Standard, Search) so while not having as strong a recall
and conversion rate, it acts as a reminder for those already
with awareness, and as a reach extender. We shall look at
ways to enhance the standard banner effectiveness later.
Another way to look at this conversion process is to try to
understand when the conversion ultimately takes place,
i.e. is the advertising message creating an immediate
or a time delayed conversion? Previously we have seen
that rich media is creating stand-out to ensure brand
recall later, but is this a trend we should expect when we
consider all display advertising?
Let’s compare the following two scenarios:

0.60% • See-Click-Convert (post click)

0.50%
• See-Delay-Convert (post impression)
0.50% 0.47%

Conversion Rate
0.40%

0.30% 0.16%

0.20% 0.14%
0.14%
0.20%

0.12%
0.10%

5
.7
0.10% x1
0.00%
0.08%
STD Banner Polite Banner Expandable 0.08%
Banner
Eyeblaster Jan.2009 0.06%
Figure 2 Source: Eyeblaster Research 2009
0.04%

0.02%
Both popular rich media
formats are showing 2.5x
0.00%

Post - Click Post - Impression

uplift in the ability to generate
conversions over standard Figure 3 Source: Eyeblaster Research 2009

display inventory.
The results indeed show that there is a sizeable uplift in
The results reveal that rich media is doing two things the number of consumers who do not necessarily respond
here. First, it is attracting attention of the viewer against and convert immediately to an advertising message,
site content, and thereby ensuring greater stand out and this factor needs to be explored further in trying to
from site editorial. Second, by allowing a consumer to understand the consumer’s mindset.
tangibly touch and interact with the advertising message,
the consumer experience is enhanced through physical
involvement, with a resulting positive effect for the
advertiser, shown in actual conversions.
1.75x conversions happen
after a time delay
5 Analytics Bulletin Issue 3 | March 2009
 Improving the conversion process with Data
Capture
We all know that in the “viewer to converter” funnel there
is a drop off at every step. Therefore, we should aim to
Whether you go for standard
minimize the funnel steps. In general, data capture is the
ability to collect information submitted by the user. By
or rich display advertising,
placing data capture functionality into the advert itself;
we are addressing a type of conversion that occurs on
not including data capture
the web site where the advert is placed, instead of the
advertiser’s web site. This feature can be implemented in
within any DR campaign is at
various ways: one of the most common is allowing users
to request additional advertiser information from within the
least halving your overall ROI
ad. It can also be used for polling users’ opinion, collecting
email addresses, cell phone numbers, or any other details,
effectiveness.
including full transactional details.

“Viewer to Converter” Funnel

On-Site conversion in-Banner Conversion Rich

Regular banner flow (Rich Media banner flow
or Standard)

Viewing and/or interacting

with the ad Viewing and/or interacting
with the ad
Clicking on the advert

Arriving at the Details Submission

web page

Sign in-start
page Details
Confirmation
Sign in -
Confirmation
page

Figure 4 Source: Eyeblaster Research 2009

6 Analytics Bulletin Issue 3 | March 2009

 Conversion Rate (on-Site vs. in-Banner)
You are 5x more likely to
convert in a banner than
1.20%
1.06%
convert on a site as a result of
1.00% clicking on a standard banner
0.80%
Impact drives Conversions x5 Case Study
0.60%
0.51%
Health and Beauty campaign
0.40%
Let’s try to clarify the power of data capture by
0.20% referring to the following case study.
0.20%
In a Health and Beauty campaign we monitored
0.00% the amount of email addresses submitted on
Standard Rich Banner Data Capture a daily basis during the campaign’s lifetime, 22
Banner days. (Over 3 million impressions were served on
Yahoo, nine.MSN and news.com.)
Figure 5 Source: Eyeblaster Research 2009
in-Banner vs. on-Site Conversions

When comparing the data capture within a banner

60
(1.06%) with that of on-site (0.20% and 0.51%) we
Conversion or Form Submission

see that by removing a number of required steps, the 50

conversion rate is 5x that of standard display and
40
double that of rich media as a result!
30
By simplifying the process that the consumer is required
to follow, the advertiser achieves a more desirable result. 20

However, it is important to note that the level of interactivity 10

within Rich Media is NOT a bearing on the conversion
0
result here; it is purely about making things easier for the 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
consumer to connect and convert in general. Users want Days from campaign’s start date
to connect with a brand where they are. Form Submissions Confirmation Conversions

When we consider that standard advertising does not Eyeblaster Jan.2009

have the ability to measure anything beyond a click,

Figure 6 Source: Eyeblaster Research 2009
the need for inclusion of a data capture facility across
networks is startlingly obvious. Not only is this challenging
the notions of cost-effectiveness of standard media buys
at driving conversions, but also what we have come to The amount of emails submitted
define as Direct Response advertising online. It would from the banner itself (data
serve all standard advertising well to move towards the
inclusion of simple response based functionality.
capture) was significantly higher
throughout the campaign’s lifetime
in comparison to emails submitted
from the advertiser’s site.

7 Analytics Bulletin Issue 3 | March 2009

 Further challenging the notion of CTR as an
effective conversion metric
With CTR being the de-facto measurement of online DR
campaigns, how does this need to be addressed given the
wealth of data on consumer behaviour discovered over the
last 10 years about rich media advertising?
Unique Clicking Users vs. Form Submission
1.20%
1.06%
1.00%
0.80%
Personalizing the “Viewer to Converter”
Funnel
Impressions and CTR are non-personal attempts at trying
to understand what consumers actually do online. In
breaking down the “viewer to converter” funnel for those
who convert in banner, we need to try and understand how
many times does a consumer get exposed to an advert,
how many touch the advert, for how long do they interact
with it and what is the resulting conversion pattern?
Therefore we see a funnel of total impressions ➔ unique
impressions ➔ unique interacting users ➔ number of
conversions (which we will determine as form submission
and assuming that person will not submit more than one
form and thus can be related to as a unique number.)
• See-Interact-Convert (in-Banner)

0.60%
77
x7.

0.40%

0.20%
0.14%

0.00%

Click Through Rate In - Banner

Conversion Rate
Figure 7 Source: Eyeblaster Research 2009

Using global data, we can see that the inclusion of data
capture within a banner is just shy of 8x more effective at
addressing conversions than CTR. It begs the question,
why don’t all DR campaigns have basic data capture
functionality built-in as standard?
Consumers are nearly 8x more
likely to fill in a form in a banner
then click on it
• Unique viewers’ rate – unique out of total impressions
• Unique Interacting Users - unique interacting users out
of unique impressions
• Average Dwell Time – number of seconds a viewer
actively is engaged with an advert
• Converting users – form submissions out of unique
interacting users
34% of unique users who
interact will fill in form details
within a banner
What we see from the data is the viewer is being exposed
to the ad at least twice (42.79%).
7.40% of these then go on to interact with the ad, and do
so for well over a minute (78.08 secs).
33.56% of those who interact go on to convert within the
banner.

8 Analytics Bulletin Issue 3 | March 2009


Case Study
e-Commerce transactional
banner
How far can we push conversions
in-Banner?
A leading European mobile operator wanted to
attract new subscribers to their network. The
concept of data capture was taken to the next
stage as they wanted to get much more personal
attributes from the interested user as part of
entering into a contract with the operator. In
this case not just address details, but additional
qualifiers such as passport or ID card number
as well as bank account details. The incentive
to do so was an offer for a free mobile handset.
Communication and conversion
People love to talk. Whether it is product feedback or
recommending to a friend, there are a variety of methods
to start to penetrate the consumer conversation to ensure
higher conversion rates. Widgets within social networks
are one such method. But what is more interesting is
that due to the poised typing position people have within
communication channels, such as email and Messenger,
there is often more of a willingness to fill in data capture
forms within these channels.
Health Campaign
CTR vs. Data Capture
0.70%
0.60%
0.60%
0.50%
0.40%

.06
x4
The placement of a creative on a leading 0.30%

portal seemed to offer assurance to the 0.20%

interested consumer and combined with the 0.10%
0.10%
incentive was enough to inspire confidence
of 500 consumers who went on to give full 0.00%
personal details within the advert! This level Click Through Rate In - Banner
of success was well-beyond the expectations of Submission
the advertiser, but it goes to prove the willingness
of modern consumers – and the need to not Figure 8 Source: Eyeblaster Research 2009
overlook data capture within creative executions.

Taking standard display advertising to the

next level:
A Health and Beauty Campaign that utilized a standard
display advert within a communication channel found that
the CTR was under performing. However, the inclusion of
a simple data-capture form that opened in an expansion
panel upon clicking on a button within the creative saw
a quadruple rise in in-banner submissions over the CTR,
incentivized by the chance to win a pair of sunglasses.

Inclusion of a simple data-

capture form… saw a
quadruple rise in in-banner
submissions over the CTR
This is a trend that is seen in Automotive, Travel and
Health who often utilize data capture as part of their
campaign objectives.
9 Analytics Bulletin Issue 3 | March 2009

Case Study
Data Capture as the
desired call to action
Masterfoods hoped to compel interested
consumers to sample a new type of chocolate
without the need to click through to a mini-site.
They created a contest to drive interest and
engagement and a series of ads was developed
around the idea of “Mix It Up”. The prize was the
chance to win a 3 month’s supply of chocolates
in exchange for personal contact details. The
executions were adapted to suit the publisher
environment; e.g. high impact on the homepage,
with the ability to play a game prior to entry
into the competition – but data-capture alone
within communication channels. All creatives
were designed to remove steps to conversion
by immediately presenting the competition entry
form on roll-over which was the focus, instead
of muddying the call to action with click-thru.
It worked. In a single day 35,000 people
entered details indicating they wanted to
taste Mars chocolates! An incredible indication
to a brand as to audience favorability and a true
demonstration of online DR at work.
The inclusion of an incentive,
such as a giveaway or access to
exclusive content, will positively
impact the data capture rates.
10 Analytics Bulletin Issue 3 | March 2009
Automatic Optimization mechanism and its
impact on conversion rate
Not only can you minimize the number of steps in the
“viewer to converter” funnel, there is an ability to monitor in
real-time the effects of one creative against another during
the campaign.
Eyeblaster’s Automatic Optimization mechanism uses
an algorithm to find the ads with the highest success
rate based on the metric defined, which can often vary
upon placement of media by section or publisher. The
case study below demonstrates the results of such
a methodology deployed and the overall campaign
performance based on conversions.
Comparisons against similar
campaigns - one with
optimization and one without,
reveals a considerable change
in the ability to control a
positive shift in ROI.
From data collected across one Automotive advertiser
over a six month period, the conversion rate of campaigns
using an optimization algorithm (0.63%) is significantly
higher (1.25x) than the average conversion rate of rich
media ads (0.51%). In fact, it is practically double the
average conversion rate for WW ’08 – 0.32%.
To understand this more fully, let’s delve inside a campaign
that implemented real-time optimization.
0.70%
0.60%
Automotive
Advertiser (0.63%)
0.50%
x1.25
0.40% Rich Media
(0.51%)
0.30% x1.97
0.20%
WW ’08
0.10% (0.32%)
0.00%
Conversion Rate
Figure 9 Source: Eyeblaster Research 2009
 Search within the conversion process
Case Study As we have seen, not all display advertising messages can
be tracked either by an immediate response within the
A campaign using the banner, but neither can they be determined by a post-view
arrival at an advertiser’s site.
Automatic Optimization
mechanism Influence of Online Display on Clicks
of Search Campaigns
Creative A is tested against Creative B in the
same automotive campaign. All factors such as
placement, ad format and size remain consistent, 2000 600

just the creative messaging was different – there 1800

were 5 messages per placement. Looking at the 1600

500

final conversion rate as the ultimate point to weigh

against – irrespective of click-thru or interaction

Impressions (in 1000s)

1400
400
rate – we are able to control a significant uplift in 1200

conversions.

Clicks
1000 300

800
Conversion Rate with Automatic 200
Optimization 600

400
100

200
25.00%
Weeks 0 0
1 2 3 4 5 6 7 8
20.00%

Impressions of display campaigns Clicks of search campaigns

15.00%
Figure 11 Source: Eyeblaster Research 2009

10.00%
5.00%
0.00%
1 3 5 7 9 11 13 15 17 19 21 23 25 27
Conv Rate of a Regular f light Conv Rate of an Optimized f light
Figure 10 Source: Eyeblaster Research 2009
The non-optimized flight was performing better
initially, but once the technology automatically
adjusted the weighting of the creative in the
optimized flight, you can see the uplift in the trend.
When looking at the final data, we find that the
optimized creative over time lifts the overall
conversion rate by 0.82% for the non-optimized
message to create impressive spikes of 23.42% –
and totally smashes the average global conversion
rate of 0.32%.
In some scenarios, after a consumer is exposed to a
message, they wish to perform an independent evaluation,
which ultimately may even lead to a purchase offline or at
a reseller’s site. Think about your recent purchase behavior
– did you review information on a comparison site first?
And was that purchase ultimately made directly on the
advertisers site – if online at all? Therefore we need to look
at cross-channel activity in order to ascertain a deeper
understanding of overall conversion effectiveness beyond
in-banner or on-site conversions.
In this example, we find that initially search activity is fairly
dormant. Once the display activity has begun, there is a
followed lift in the search activity. What is important to note
is that often there is a delay between the display exposure
and the performing of the search, as a consumer finds
their own time to do their research.
Optimizing creatives lifts
the overall conversion rate

11 Analytics Bulletin Issue 3 | March 2009

 4500 4500

4000 4000

Thomas McIlheran, Senior Media Manager

3500 3500
says: “The Eyeblaster Channel Connect
Impressions (in 1000s)

3000 3000
Reporting system is the most progressive
2500 2500 reporting tool for monitoring the joint impact of

Clicks
2000 2000
search and display advertising.  CC4S allows
advertisers and marketers to put numbers
1500 1500
around exactly how well these online tactics
1000 1000 work together as opposed to how they work
500 500
against or in spite of one another.”
0 0

03/10/ 17/10/ 31/10/ 14/11/ 28/11/ 12/12/ 12/12/ 09/01/

2008 2008 2008 2008
Impressions of display campaigns
Figure 12 Source: Eyeblaster Research 2009
The tailing-off effect of the search activity is maintained
at a much slower decline than the display advertising, as
the emotional connections a consumer has made as a
result of the ad exposure continue to play on their minds
for perhaps the next couple of weeks. It could also be fair
to say that the multiple exposures to display messaging
during this time are having a positive impact in the desire
to search, and this higher frequency ensures a greatly
drawn-out time span of searches – long after the effects
can be seen within traditional display advertising metrics,
i.e. clicks have declined, but searches are maintained.
This inherently challenges traditional notions of how
to optimize display advertising frequency, based
upon display advert CTR.
There is a delay between the
display exposure and the
performance of the search -
potentially by as much as 3
weeks!
From campaigns we have looked at already, we have
found a positive shift towards brand/product name
searches as a result of the display activity, proving recall
effect of display advertising. We have also seen uplift in
search activity as a result of display impressions. We shall
look forward to sharing such results in future bulletins.
2008 2008 2008 2009
“Using Channel Connect for Search and
Clicks of search campaigns
Display, Mindshare was able to pinpoint crucial
campaign data and draw important insights
about the interaction of our search and display
ads. In the end, it provided us with a more
comprehensive overview of user behavior”,
Harry Case, Director of Media Analytics
and Technology, Mindshare.
Targeted messaging with behavioral
sequencing
The ability to adaptively sequence creative messaging can
be done using various tracking methodologies utilizing
cookies and Flash Shared Objects. This has distinct uses
to ensure relevance to the viewer.
First, you can ascertain if a user has previously been to the
advertiser’s site, and begin to delve where on the site and
the profile of the person based upon established criteria.
This enables knowledge of the person exposed to the
creative to be ascertained prior to being shown a creative
and is known in Eyeblaster as DéjàVu.
Second, there is the ability to learn what creatives have
been previously shown and/or interacted with and adapt
the creative messaging accordingly. This is known as
behavioral sequencing.
Finally, there is also the ability to read in host-site
content of where the ad is sitting and adapt the creative
messaging to the surrounding page content. Called
Mash-up technology, this is another way to increase
advertising relevance.

12 Analytics Bulletin Issue 3 | March 2009

 Case Study
Sony Ericsson k550i campaign
Sony Ericsson ran a campaign across CEEMEA, which adapted differently depending on how a user
had interacted – and where – on site or within a banner. The creative was also changed each week to
try and minimize ad fatigue through over exposure.

http://creativezone.eyeblaster.com/#ItemName=SE%20K550

Total
Data Impressions Clicks CTR Stage 1 Stage 2 Upload Send
Custom
Week 1 20,827,665 38,825 0.19% 1,704 249 25 1,978
Week 2 33,928,483 32,264 0.10% 12,416 85 1,058 200 13,759
Week 3 35,917,203 32,233 0.09% 20,245 423 1,600 280 22,548
Week 4 36,036,434 33,369 0.09% 17,978 111 1,827 357 20,273
Total 126,709,785 136,691 0.12% 52,343 619 4,734 862 58,558

Figure 13 Source: Eyeblaster Research 2009

The results demonstrated that as the creative Those exposed to a more

fatigue happened with each week the campaign
ran, even with changing the creative per week, targeted message… were
CTR steadily declined due to over-saturation. more likely to move through
Stages 1 and 2 represent a targeted message the conversion sequence
showing a personal message for this week’s
competition (stage 1) or if the user had already
participated, a message advertising the coming
week’s competition (stage 2) up to the final week
4. The data revealed that those exposed to a
more targeted message through prior involvement,
either on site or within the banner, were more likely
to move through the conversion sequence; in this
case uploading photographs alongside personal
data and as a result, actively recommend the
campaign to their friends with a forward-to-friend
mechanism built into the ad unit.
This is one of the most exciting findings of how
activity on-site and in-banner can work together
to engage consumers and lead them through
a conversion cycle, as well as engage brand
advocacy to their friends.

13 Analytics Bulletin Issue 3 | March 2009

 Conclusion
The advent of being able to link data across-channels is a
revolutionary breakthrough for the advertising community The potential to deliver better
and heralds the start of a completely new way of being
able to effectively and accurately strategize media ROI for clients is easily within
and reduce overall marketing dollar wastage. We look
forward, in future Analytics Bulletins, to show how this is grasp for agencies and their
progressing.
clients
We have shown how there are multiple ways to utilize
conversions and gain higher ROI for advertisers. As with
most software, between 5-10% of possible functionality is
currently being utilized within most campaigns and thereby
the potential to develop more cost-effective advertising is
being lost through lack of knowledge of what is possible
with modern day ad serving. As this functionality is
generally included within the ad serving costs, coupled
with a little creative strategy, the potential to deliver better
ROI for clients is easily within the grasp for most agencies
and their clients.
Ask your Eyeblaster representative about how
you can harness the potential of Eyeblaster’s Ad
Campaign Manager to deliver greater ROI for your
next client objective.

14 Analytics Bulletin Issue 3 | March 2009

 Appendix 1
Benchmark for North America Q4 ’07 - Q3 ’08
Performance Metrics (Format and Verticals)

Basic Metrics Video Metrics Expandable Metrics

Interaction Impressions Avg.
Avg. Video 50% Fully Total
Average Started with any Panel Expansion
IR CTR Duration Played Played Expansion
Duration Rate Expansion Duration
(Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Standard Banner — — 0.09% — — — — — — —
Rich Media
Commercial Break 13% 2.73 2.61% 5.46 72% 10% 3% — — —
Expandable Banner 11% 3.87 0.30% 23.62 15% 79% 69% 11% 7% 51.46
Floating Ad 3% 2.78 3.15% 6.64 48% 50% 34% — — —
Formats

Floating Ad With
5% 3.00 2.80% — — — — — — —
Reminder
Floating Expandable 10% 3.26 1.52% 5.26 62% 83% 70% 72% 5% 23.70
In Game 13% — 4.67% 16.60 94% 71% 47% — — —
Polite Banner 5% 3.36 0.17% 19.76 53% 68% 52% — — —
Push Down Banner 4% 2.29 0.28% — — — — 35% 3% 18.80
Video Strip 17% 4.02 0.17% 17.26 48% 59% 31% 46% 6% 70.23
Interaction Impressions Avg.
CTR CTR Avg. Video 50% Fully Total
Average Started with any Panel Expansion
IR (Rich (Standard Duration Played Played Expansion
Duration Rate Expansion Duration
Media) Media) (Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Apparel 12% 3.33 0.30% 0.24% 15.04 47% 74% 36% 20% 7% 30.41
Auto 9% 3.35 0.23% 0.11% 19.63 16% 76% 64% 17% 9% 31.20
B2B 6% 3.11 0.48% 0.17% — — — — 12% 9% 65.23
Careers 10% 3.45 0.30% 0.09% — — — — 24% 16% 53.43
Consumer Packaged
7% 3.63 0.35% 0.11% 24.90 25% 54% 33% 9% 6% 48.21
Goods
Corporate 1% 3.09 0.09% 0.05% 22.05 7% 47% 31% 12% 8% 78.68
Electronics 8% 3.71 0.24% 0.12% 19.63 20% 66% 71% 15% 11% 42.02
Entertainment 11% 3.60 0.26% 0.13% 15.39 13% 61% 46% 9% 5% 49.49
Financial 5% 3.35 0.15% 0.07% 13.08 56% 76% 60% 6% 4% 46.69
Verticals

Gaming 7% 4.35 0.63% 0.08% 59.46 17% 49% 35% 22% 8% 61.49
Government/Utilities 5% 3.18 0.17% 0.07% 24.48 62% 78% 66% 16% 4% 26.07
Health/Beauty 13% 3.29 0.24% 0.17% 30.80 8% 63% 41% 19% 13% 61.72
Medical 6% 3.07 0.14% 0.10% 21.88 38% 72% 58% 7% 5% 36.49
News/Media 10% 2.64 0.27% 0.07% — — — — 14% 9% 22.58
Restaurant 1% 2.94 0.13% 0.09% 16.04 50% 39% 27% 3% 2% 33.43
Retail 11% 3.09 0.29% 0.07% 23.48 11% 40% 25% 16% 9% 28.08
Services 8% 3.08 0.38% 0.05% 23.23 56% 77% 63% 33% 21% 70.24
Sports 7% 2.98 0.28% 0.08% 11.37 55% 68% 52% 7% 4% 44.50
Tech/Internet 4% 3.59 0.17% 0.14% 14.82 51% 77% 65% 6% 4% 88.84
Telecom 7% 3.28 0.18% 0.09% 23.55 34% 81% 62% 15% 10% 48.35
Travel 4% 3.80 0.21% 0.08% 40.20 59% 78% 61% 13% 5% 31.28
North America: including US and Canada. The benchmark is for a rolling year, last updated Q4 2008.

Interaction Rate (IR) User Initiated Interactions' divided by 'Served Impressions'

Click-through Rate (CTR) Frequency of Click-throughs as a percentage of served impressions. 'Clicks' divided by 'Impressions'.
Avg. Video Duration The average duration the video was played, including user and auto initiated videos (in seconds).
Started Rate The number of times the video started out of video's served Impressions
50% Played Rate The number of videos that played over 50% of their total video length divided by 'Video Started'.
Fully Played Rate The number of videos that were fully played divided by video that started
Total Expansion Rate Total expansions divided by served impressions. Including Auto-initiated expansions
Impressions with any Panel Expansion Rate The number of impressions with at least one panel expansion.
Avg. Expansion Duration The average time a panel was expanded, including user and auto initiated expansions
Rich Media Ads All Eyeblaster's format excluding standard banner, wallpaper and window ads
— Not Available - either this measurement is not applicable to the format or there was not enough data to be statistically relevant

15 Analytics Bulletin Issue 3 | March 2009

 Appendix 2
Benchmark for Europe Q4 ’07 - Q3 ’08
Performance Metrics (Format and Verticals)

Basic Metrics Video Metrics Expandable Metrics

Interaction Impressions Avg.
Avg. Video 50% Fully Total
Average Started with any Panel Expansion
IR CTR Duration Played Played Expansion
Duration Rate Expansion Duration
(Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Standard Banner — — 0.14% — — — — — —
Rich Media —
Commercial Break 4% 2.30 3.96% 3.97 72% 48% 29% — — —
Expandable Banner 35% 3.27 0.66% 44.12 25% 62% 48% 41% 17% 52.60
Floating Ad 13% 2.68 4.16% 8.90 65% 51% 30% — — —
Formats

Floating Ad With
6% 2.38 4.66% — — — — — — —
Reminder
Floating Expandable 20% 2.58 0.48% — — — — 56% 14% 41.95
In Game 19% — 5.80% 17.17 99% 93% 80% — — —
Polite Banner 7% 2.95 0.21% 44.42 59% 67% 51% — — —
Push Down Banner 16% 3.46 0.15% 44.72 55% 84% 78% 22% 12% 57.37
Video Strip 23% 4.87 0.18% 63.20 10% 74% 60% 43% 8% 72.39
Interaction Impressions Avg.
CTR CTR Avg. Video 50% Fully Total
Average Started with any Panel Expansion
IR (Rich (Standard Duration Played Played Expansion
Duration Rate Expansion Duration
Media) Media) (Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Apparel 15% 3.05 0.38% 0.17% 51.92 35% 57% 44% 47% 16% 47.41
Auto 14% 2.85 0.41% 0.15% 49.68 51% 68% 51% 42% 16% 55.11
B2B 9% 3.01 0.44% 0.34% — — — — 12% 6% 46.55
Careers 11% 3.00 0.54% 0.15% — — — — 25% 11% 51.31
Consumer
25% 3.58 0.57% 0.12% 39.68 31% 56% 42% 47% 18% 63.97
Packaged Goods
Corporate 18% 3.11 0.32% 0.08% — — — — 59% 26% 74.74
Electronics 21% 2.90 0.35% 0.13% 46.15 36% 72% 59% 42% 17% 48.40
Entertainment 28% 3.35 0.57% 0.18% 54.69 30% 62% 46% 40% 18% 39.44
Financial 17% 3.03 0.30% 0.14% 33.11 45% 62% 43% 36% 17% 64.06
Verticals

Gaming 19% 3.48 0.68% 0.21% 74.16 39% 61% 41% 41% 17% 51.12
Government/Utilities 20% 3.33 0.47% 0.08% 49.45 41% 64% 42% 41% 16% 51.84
Health/Beauty 18% 2.95 0.42% 0.10% 40.92 38% 62% 45% 30% 15% 47.21
Medical 12% 2.92 0.34% 0.28% 52.04 32% 50% 32% 54% 18% 57.12
News/Media 8% 3.04 0.35% 0.15% 93.30 45% 60% 42% 42% 14% 30.76
Restaurant 23% 3.13 0.38% 0.07% 25.33 53% 46% 37% 44% 15% 57.49
Retail 21% 2.83 0.47% 0.15% 42.97 57% 63% 44% 54% 22% 57.30
Services 21% 3.13 0.44% 0.08% 56.89 29% 62% 42% 42% 18% 55.09
Sports 8% 3.05 0.31% 0.10% 32.58 42% 75% 53% 38% 24% 28.36
Tech/Internet 15% 3.05 0.38% 0.14% 32.89 43% 82% 72% 37% 14% 46.39
Telecom 22% 2.98 0.43% 0.13% 49.42 48% 56% 41% 42% 17% 51.14
Travel 22% 3.15 0.49% 0.10% 42.38 59% 69% 57% 40% 14% 46.39
Europe: including Austria, Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Netherlands, Poland, Portugal, Spain, Sweden, Switzerland, UK.
The benchmark is for a rolling year, last updated Q4 2008.

Interaction Rate (IR) User Initiated Interactions' divided by 'Served Impressions'

Click-through Rate (CTR) Frequency of Click-throughs as a percentage of served impressions. 'Clicks' divided by 'Impressions'.
Avg. Video Duration The average duration the video was played, including user and auto initiated videos (in seconds).
Started Rate The number of times the video started out of video's served Impressions
50% Played Rate The number of videos that played over 50% of their total video length divided by 'Video Started'.
Fully Played Rate The number of videos that were fully played divided by video that started
Total Expansion Rate Total expansions divided by served impressions. Including Auto-initiated expansions
Impressions with any Panel Expansion Rate The number of impressions with at least one panel expansion.
Avg. Expansion Duration The average time a panel was expanded, including user and auto initiated expansions
Rich Media Ads All Eyeblaster's format excluding standard banner, wallpaper and window ads
— Not Available - either this measurement is not applicable to the format or there was not enough data to be statistically relevant

16 Analytics Bulletin Issue 3 | March 2009

 Appendix 3
Benchmark for Spain Q4 ’07 - Q3 ’08
Performance Metrics (Format and Verticals)

Basic Metrics Video Metrics Expandable Metrics

Avg.
Avg. Video 50% Fully Total Impressions
Started Expansion
IR CTR Duration Played Played Expansion with any Panel
Rate Duration
(Seconds) Rate Rate Rate Expansion Rate
(Seconds)
Standard Banner — 0.18% — — — — — — —
Rich Media
Commercial Break 4% 3.87% — — — — — — —
Formats

Expandable Banner 53% 1.02% 73.23 20% 56% 44% 64% 23% 67.38
Floating Ad 8% 2.08% — — — — — — —
Polite Banner 6% 0.25% 60.29 47% 65% 50% — — —
Push Down Banner 12% 0.21% — — — — — — —
Avg.
CTR CTR Avg. Video 50% Fully Total Impressions
Started Expansion
IR (Rich (Standard Duration Played Played Expansion with any Panel
Rate Duration
Media) Media) (Seconds) Rate Rate Rate Expansion Rate
(Seconds)
Apparel 21% 0.62% 0.45% 26.86 41% 55% 40% — — —
Auto 19% 0.44% 0.18% 21.57 54% 54% 40% 76% 30% 58.70
Consumer Packaged
42% 0.95% 0.15% 31.62 25% 62% 38% — — —
Goods
Electronics 17% 0.45% 0.16% 40.81 61% 74% 60% 47% 22% 57.02
Entertainment 51% 1.06% 0.22% 35.28 44% 66% 51% 53% 10% 51.19
Financial 14% 0.34% 0.12% 87.28 60% 84% 52% — — —
Verticals

Gaming 40% 0.94% 0.58% 38.05 44% 62% 45% — — —

Government/Utilities 13% 0.30% 0.08% — — — — 46% 7% 73.58
Health/Beauty 38% 0.81% 0.22% 54.38 35% 62% 46% 43% 13% 51.90
Retail 18% 0.65% —
Services 41% 0.54% 0.14% — — — — 79% 23% 63.60
Tech/Internet 9% 0.48% 0.14% 44.00 31% 69% 40% 68% 24% 57.03
Telecom 34% 0.50% 0.39% — — — — — — —
Travel 22% 0.68% — 35.64 60% 43% 29% 50% 8% 38.69

The benchmark is for a rolling year, last updated Q4 2008.

Interaction Rate (IR) User Initiated Interactions' divided by 'Served Impressions'

Click-through Rate (CTR) Frequency of Click-throughs as a percentage of served impressions. 'Clicks' divided by 'Impressions'.
Avg. Video Duration The average duration the video was played, including user and auto initiated videos (in seconds).
Started Rate The number of times the video started out of video's served Impressions
50% Played Rate The number of videos that played over 50% of their total video length divided by 'Video Started'.
Fully Played Rate The number of videos that were fully played divided by video that started
Total Expansion Rate Total expansions divided by served impressions. Including Auto-initiated expansions
Impressions with any Panel Expansion Rate The number of impressions with at least one panel expansion.
Avg. Expansion Duration The average time a panel was expanded, including user and auto initiated expansions
Rich Media Ads All Eyeblaster's format excluding standard banner, wallpaper and window ads
— Not Available - either this measurement is not applicable to the format or there was not enough data to be statistically relevant

17 Analytics Bulletin Issue 3 | March 2009

 Appendix 4
Benchmark for APAC Q4 ’07 - Q3 ’08
Performance Metrics (Format and Verticals)

Basic Metrics Video Metrics Expandable Metrics

Interaction Impressions Avg.
Avg. Video 50% Fully Total
Average Started with any Panel Expansion
IR CTR Duration Played Played Expansion
Duration Rate Expansion Duration
(Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Standard Banner — — 0.11% — — — — — — —
Rich Media
Commercial
3% 2.70 3.14% — — — — — — —
Break
Expandable
21% 3.32 0.34% 22.96 6% 47% 36% 24% 11% 57.25
Banner
Formats

Floating Ad 8% 3.00 2.28% 50.89 76% 84% 78% — — —

Floating Ad With
5% 2.64 1.60% 11.04 63% 64% 21% — — —
Reminder
Floating
32% 2.93 0.43% — — — — 78% 32% 151.76
Expandable
Polite Banner 2% 2.82 0.11% 40.54 40% 66% 50% — — —
Push Down
26% 2.89 0.32% 17.32 23% 63% 47% 45% 11% 24.57
Banner
Interaction Impressions Avg.
CTR Avg. Video 50% Fully Total
Average CTR (Rich Started with any Panel Expansion
IR (Standard Duration Played Played Expansion
Duration Media) Rate Expansion Duration
Media) (Seconds) Rate Rate Rate
(Seconds) Rate (Seconds)
Apparel 10% 3.11 0.36% 0.18% 17.65 17% 49% 34% 26% 12% 67.31
Auto 7% 2.92 0.23% 0.22% 46.30 27% 78% 69% 17% 10% 52.91
Careers 24% 3.45 0.46% 0.04% — — — — 62% 22% 36.10
Consumer
12% 3.27 0.33% 0.12% 24.46 25% 70% 56% 26% 13% 50.61
Packaged Goods
Corporate 5% 2.55 0.12% 0.05% — — — — 10% 5% 50.69
Electronics 16% 2.88 0.32% 0.11% 34.20 34% 64% 50% 29% 15% 56.57
Entertainment 11% 3.09 0.35% 0.16% 29.23 16% 68% 55% 21% 10% 56.96
Financial 8% 2.76 0.15% 0.05% 19.87 22% 64% 52% 30% 8% 74.25
Gaming 6% 4.60 0.30% 0.10% 33.03 20% 54% 41% 6% 3% 46.98
Verticals

Government/
18% 3.77 0.31% 0.18% 69.27 20% 38% 24% 25% 13% 55.22
Utilities
Health/Beauty 12% 2.86 0.45% 0.11% 26.73 18% 47% 34% 23% 13% 62.54
Medical 15% 2.75 0.41% 0.09% — — — — 32% 16% 55.26
News/Media 11% 3.51 0.21% 0.18% — — — — 19% 11% 68.97
Restaurant 27% 4.16 0.67% 0.17% — — — — 37% 17% 32.42
Retail 10% 3.70 0.22% 0.11% 35.47 21% 50% 37% 21% 11% 36.07
Services 9% 2.77 0.14% 0.06% 50.42 13% 51% 35% 15% 8% 53.70
Sports 3% 2.36 0.40% 0.06% — — — — — — —
Tech/Internet 4% 2.61 0.12% 0.10% 27.48 27% 58% 38% 23% 11% 67.82
Telecom 13% 3.23 0.31% 0.09% 32.26 11% 52% 40% 34% 15% 49.93
Travel 15% 2.95 0.32% 0.15% 12.99 17% 63% 51% 25% 10% 48.51
APAC: including Australia, China, Hong Kong, India, Japan, Korea, Malaysia, New Zealand, Pakistan, Singapore. The benchmark is for a rolling year, last updated Q4 2008.

Interaction Rate (IR) User Initiated Interactions' divided by 'Served Impressions'

Click-through Rate (CTR) Frequency of Click-throughs as a percentage of served impressions. 'Clicks' divided by 'Impressions'.
Avg. Video Duration The average duration the video was played, including user and auto initiated videos (in seconds).
Started Rate The number of times the video started out of video's served Impressions
50% Played Rate The number of videos that played over 50% of their total video length divided by 'Video Started'.
Fully Played Rate The number of videos that were fully played divided by video that started
Total Expansion Rate Total expansions divided by served impressions. Including Auto-initiated expansions
Impressions with any Panel Expansion Rate The number of impressions with at least one panel expansion.
Avg. Expansion Duration The average time a panel was expanded, including user and auto initiated expansions
Rich Media Ads All Eyeblaster's format excluding standard banner, wallpaper and window ads
— Not Available - either this measurement is not applicable to the format or there was not enough data to be statistically relevant

18 Analytics Bulletin Issue 3 | March 2009

 Appendix 5
Benchmark for Taiwan Q4 ’07 - Q3 ’08
Performance Metrics (Format and Verticals)

Basic Metrics Video Metrics Expandable Metrics

Avg.
Avg. Video 50% Fully Total Impressions
Started Expansion
IR CTR Duration Played Played Expansion with any Panel
Rate Duration
(Seconds) Rate Rate Rate Expansion Rate
(Seconds)
Standard Banner — 0.04% — — — — — — —
Rich Media
Expandable Banner 45% 0.58% 307.36 40% 72% 63% 33% 16% 73.48
Formats

Floating Ad With
4% 1.97% — — — — — — —
Reminder
Polite Banner 2% 0.13% 69.28 61% 53% 38% — — —
Push Down Banner 2% 0.20% 17.32 23% 63% 47% 29% 0.14% 18.58
Avg.
CTR CTR Avg. Video 50% Fully Total Impressions
Started Expansion
IR (Rich (Standard Duration Played Played Expansion with any Panel
Rate Duration
Media) Media) (Seconds) Rate Rate Rate Expansion Rate
(Seconds)
Apparel 6% 0.25% — 58.71 30% 67% 51%
Auto 29% 0.54% — 20.32 19% 61% 47% 36% 11% 14.64
Consumer Packaged
5% 0.20% — 72.24 39% 60% 47% — — —
Goods
Verticals

Entertainment 14% 0.35% — 229.87 45% 69% 58% 28% 13% 80.67
Financial 52% 0.74% — — — — — — — —
Health/Beauty 28% 0.38% — 296.51 42% 71% 62% 26% 17% 85.24
Tech/Internet 2% 0.20% 0.04% — — — — — — —
Telecom 36% 0.47% — 93.79 29% 59% 43% 43% 28% 84.96
Travel 18% 0.42% — — — — — 28% 10% 27.70

The benchmark is for a rolling year, last updated Q4 2008.

Interaction Rate (IR) User Initiated Interactions' divided by 'Served Impressions'

Click-through Rate (CTR) Frequency of Click-throughs as a percentage of served impressions. 'Clicks' divided by 'Impressions'.
Avg. Video Duration The average duration the video was played, including user and auto initiated videos (in seconds).
Started Rate The number of times the video started out of video's served Impressions
50% Played Rate The number of videos that played over 50% of their total video length divided by 'Video Started'.
Fully Played Rate The number of videos that were fully played divided by video that started
Total Expansion Rate Total expansions divided by served impressions. Including Auto-initiated expansions
Impressions with any Panel Expansion Rate The number of impressions with at least one panel expansion.
Avg. Expansion Duration The average time a panel was expanded, including user and auto initiated expansions
Rich Media Ads All Eyeblaster's format excluding standard banner, wallpaper and window ads
— Not Available - either this measurement is not applicable to the format or there was not enough data to be statistically relevant

19 Analytics Bulletin Issue 3 | March 2009

www.eyeblaster.com