1
Visual Noise:
The Role of Site Clutter in
Advertising Branding Effectiveness
October 29, 2001
Study Participants:
Gerard Broussard, OglivyOne
Jeffrey Graham, Dynamic Logic
Debbie Reichig, iVillage
Marc Ryan, Jupiter Media Metrix
2
Research Background
• iVillage and Dynamic Logic have partnered since 7/00 to
measure the branding effectiveness of advertising on their
Web site
• iVillage branding results have consistently exceeded
norms; we hypothesized their site redesign, based on user
feedback which reduced clutter on the site, was a factor
• Research was to explore the relationship between clutter
and branding effectiveness
– Research developed with OglivyOne and Jupiter Media Metrix
3
What is online clutter?
4
Clutter offline vs. online
• In offline media, clutter is usually defined by the number
of advertisements in a given space or time period
– Television research generally represents clutter as the total
advertising/promotional time
– Ad clutter has been proven to have the potential to detract from
ad effectiveness
• Clutter on web pages doesn’t seem to be a function of
number of advertisements only
5
Does online clutter equal ad clutter?
This page has three
ads, but doesn’t
seem very cluttered.
The ads stand out.
This page has only one
advertisement, but is
extremely cluttered.
The ad logo is lost.
6
Ad Clutter is increasing….
0
100
200
300
400
500
600
700
800
900
1000
Ads per person
Sep-00
Oct-00
Nov-00
Dec-00
Jan-01
Feb-01
Mar-01
Apr-01
May-01
Jun-01
Jul-01
Aug-01
Source: AdRelevance
7
…but the range is still small
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Travel, Maps & Local
Health & Fitness
General News
Business & Finance
Fashion, Romance & Celebrity
Automotive
Home & Garden
Telecom & Internet Telephony
Reference & Education
Community
Sports & Recreation
Personal Expression
Portal
Movies & Television
Employment
Computing & Technology
Search Engine
Society, Politics & Science
Incentive
Yellow & White Pages
B2B
Music & Streaming Media
Comics & Humor
Kids & Family
Games
Shopping & Auction
Average Page
Clutter
Travel and Health sites have the highest level of clutter, pulling the
industry average of 2.6 elements per page up. Portals are almost
exactly average.
Source: AdRelevance
8
Our study’s definition of clutter
• We defined clutter as the number of text and
image elements on a page
• For study we modified an iVillage page to High,
Medium and Low clutter
• We also included in our survey a question about
clutter perception
9
High, Medium, Low Defined Pages
230 Words
8 Colored elements
18 Interest Areas
230 Words
8 Colored elements
18 Interest Areas
141 Words
4 Colored elements
10 Interest Areas
141 Words
4 Colored elements
10 Interest Areas
119 Words
3 Colored elements
8 Interest Areas
119 Words
3 Colored elements
8 Interest Areas
High
Medium
Low
10
2) They were exposed to a Web page that included advertising (except for
control group)
Exposed to Web
Pages, with or
without ad (control)
High, Medium,
Low, Skyscraper
and Control
Cells
1) Respondents were recruited from iVillage and randomly assigned to a
High, Medium, Low, High/Skyscraper or Control Cell (8/21 – 8/25)
Recruited from
iVillage Web site
* Both groups are random samples from the same population (they are
statistically the same people)
3) They were asked a series of questions in a blind questionnaire
Given blind survey
questionnaire
Does Clutter Affect Branding?: Methodology
11
Research Matrix
N=1500
Control (low 100 100 100 300
clutter, dummy ad)
High Clutter with 100 100 100 300
Skyscraper
High Clutter 100 100 100 300
Medium Clutter 100 100 100 300
Low Clutter 100 100 100 300
Ad Banner 1 Ad Banner 2 Ad Banner 3 Total
12
Analysis Methodology
• Results were analyzed two ways to assess the impact of
exposure to advertising
– Test for difference between two population proportions
This is used to assess if the brand level of the exposed cell is
significantly higher than the control cell.
– ANOVA Test (Analysis of Variance)
This tests for significant differences.
• There are two ways that the cells were defined within the
study
– “Defined Clutter” – defined by research team based on actual
number of elements on web page that respondent was shown.
– “Perceived Clutter” – defined by respondents per their own
assessment of page clutter.
13
Findings
14
Key Findings
• Our “Defined Clutter” classifications did not hold
up; there were no significant differences among
defined clutter levels in branding metrics
• Among “Perceived Clutter” levels, however, there
were significant differences in purchase/behavior
intent and brand favorability
– Lower perceived clutter sites had higher brand
favorability and purchase intent
*Statistically significant difference between control and exposed group at a 90% confidence level 15
Defined Clutter: Aggregate Results
59%
65%*
64% 64%
61%
50%
60%
70%
Control Low Medium High Sky
Lift: 9%
Lift: 3%
Lift: 9%
Significant increases in branding metrics were only present in the
Defined Low clutter site.
Lift: 9%
n=300 n=305 n=302 n=300 n=300
Brand Value Score,
Aggregate Results
*Statistically significant difference between control and exposed group at a 90% confidence level 16
Perceived Clutter: Aggregate
Results
59%
65%*
63%*
57%
50%
60%
70%
Control Not
Cluttered
Somewhat
Cluttered
Very
Cluttered
n=300 n=531 n=597 n=79
Lift: 10%
No Lift
Lift: 7%
Only the ads on pages characterized as “Very Cluttered” had no
significant increase in brand values.
Brand Value Score,
Aggregate Results
17
Purchase/Usage Intent
18
Defined Clutter:
Purchase/Usage Intent
62%
67%* 67%*
66%
63%
50%
60%
70%
Control Low Med High Sky
n=300 n=305 n=302 n=300 n=300
Lift: 8%
Lift: 7%
Lift: 2%
Lift: 7%
Exposure to the ads on the Low and Medium clutter sites led to significant
increases in Purchase/Usage Intent.
Purchase
Intent Score
*Statistically significant difference between control and exposed group at a 90% confidence level
19
Defined Clutter:
Purchase Intent : ANOVA
2.68
2.66 2.64
2.53
2.4
2.5
2.6
2.7
2.8
Low Medium High Sky
There was not a significant differences among the “Defined Clutter”
cells.
SUMMARY
Groups Count Sum Average Variance
Low 305 816 2.68 1.89
Medium 302 804 2.66 1.75
High 300 793 2.64 1.82
Sky 300 760 2.53 2.07
ANOVA
Source of Variation SS df MS F P-value F crit
Between Groups 3.798 3 1.266 0.067 0.569 2.612
Within Groups 2265.919 1203 1.884
Total 2269.717 1206
20
Perceived Clutter:
Purchase Intent
62%
67%
66%
55%
50%
60%
70%
Control Not
Cluttered
Somewhat
Cluttered
Very
Cluttered
n=300 n=531 n=597 n=79
Lift: 8%
Lift: 6%
No Lift
There were no significant increases in Purchase Intent among the
three groups.
Purchase
Intent Score
21
Perceived Clutter:
Purchase Intent: ANOVA
2.69 2.63
2.20
2.00
2.25
2.50
2.75
3.00
Not
Cluttered
Somewhat
Cluttered
Very
Cluttered
Purchase Intent was significantly lower when the advertisement
appeared in a site that was perceived to be “Very cluttered”.
SUMMARY
Groups Count Sum Average Variance
Very 79 174 2.20 1.86
Somewhat 597 1573 2.63 1.84
Not 531 1426 2.69 1.91
ANOVA
Source of Variation SS df MS F P-value F crit
Between Groups 16.084 2 8.042 4.296 0.014 3.003
Within Groups 2253.633 1204 1.872
Total 2269.717 1206
Significantly higher
than “Very Cluttered”
22
Other Metrics
• Brand Favorability: No significant lifts, significant
variance in Perceived cells, none in Defined cells
• Aided Brand Awareness: No significant lifts, no
significant variance in Defined and Perceived cells
• Message Association: Significant lifts but no
significant variance in both Defined and Perceived
cells
• Brand Attributes: No significant lifts, no significant
variance in Defined and Perceived cells
23
Defined vs. Perceived Clutter
* Statistically significant difference between control and exposed group at a 95% confidence level
Perceived vs. Actual Clutter
4%
4%
9%
40%
49%
54%
55%
47%
37%
Low
Med
High
Very Somewhat Not
• 91% of respondents exposed to high clutter page thought it was
only “somewhat” or “not” cluttered
• Our range of clutter levels was meant to be realistic; it probably
should have been more extreme
24
Conclusions
• Our definition of clutter did not have a strong relationship
to branding effectiveness
• Clutter seems to matter, but it is in the eye of the
beholder
• User-centered design is crucial for ad effectiveness
• Future research should seek to understand factors in
clutter perception
– Including multiple page experience
– Look at different audiences/sites |
|
