Moving on up: Search’s role in the

purchase funnel
Synopsis of Market Research Studies: 08
How Online and Offline Marketing Work Together to Drive Sales
Volume

Google Confidential and Proprietary

Background

• Mercedes-Benz has worked with MSP over the last three years to:
ƒ Quantify the effectiveness of marketing and incentives on business performance
ƒ Leverage the insights to optimise marketing resources

• This year, Mercedes-Benz challenged MSP to:

ƒ Quantify the individual roles and synergies between offline and online media
ƒ Quantify the role of Google (natural and paid search) in the purchase funnel

• Using data from Mercedes-Benz and Google:

ƒ MSP innovated algorithms and modeling techniques
ƒ Quantified key business drivers

Google Confidential and Proprietary

Paid Search Incrementally Converts the Queries to Sales
• Google and other search engines drive a significant amount of interested traffic for
“free”. The study found that using paid search to capture this interest amplified the
effectiveness of the medium.

Brand Awareness

Generic Search Direct

Indirect
TIMING

Model Search

Paid Search

Business Performance
Business Performance

Purchase MB Sales
4
Google Confidential and Proprietary
 5
Query Volume and Paid Search
double the effectiveness of MB UK Media

The impact of advertising in driving online search behavior and capturing the queries with
paid search is to double the effectiveness of brand and model-specific advertising.
Elasticity

Direct +Paid Search

Brand Model
Total advertising elasticity is the combined effect of direct and indirect channels

Google Confidential and Proprietary

 6
Paid search is more effective in driving sales than
traditional brand TV advertising

• Model TV advertising is the most effective in driving sales

• Paid search is more effective than Brand TV in driving sales
Elasticity

Google Confidential and Proprietary

 7

How It All Works Together

Both Offline Media and Online Display are strong drivers of sales – however when working together
and driving query volume and paid search they are twice as effective

Brand Awareness Brand TV & Press

Model TV, Press & Outdoor

Generic Search Direct

TIMING

Indirect

Model Search Online Display

Paid Search

Business
Purchase Business Performance
Performance
MB Sales
Google Confidential and Proprietary
The Ideal Resource Mix Combines Both
Traditional and Internet Media

• As Mercedes-Benz spending on the Internet increases, the relative ROI will adjust as it
moves up or down the efficiency curve. The graph below is an illustration of the efficiency
curve in relation to hypothetical marketing mix and profitability…
2#30'&

Ideal
Mix

!"#$%&'()!'*
435 +,-(.'(%/,-00.'(%1
6')7

“As a result of the study we have secured more marketing budget from
Head Office and are allocating more of this to online and search”
Will Adney, Director of Research Mercades Benz

8
Google Confidential and Proprietary
How do online and offline media work together and
what is the effect on brands?

Google Confidential and Proprietary

 Campaign elements

3 x web
advertorials
3 x TV ads

2 x email
newsletters

paid search

3 x social 4 x bespoke websites

media sites
phone boxes,
posters, POS

Google Confidential and Proprietary

 Ice Methodology

Qualitative Research Quantitative Research Modelling

Measurement of: Effect & efficiency of
Establish what the target segment really - consumers’ perceptions of communications & events
thinks about the brand, its products and the brand analysed by statistical
services - their experience and use modelling
of products and services
In their own words - Their exposure to
different marketing activity

PHASE 1 PHASE 2 PHASE 3

Google Confidential and Proprietary

New quantitative sample methodology

• Problem 1
• We need >5% of the sample to have been exposed to each
communication we want to evaluate
• Some communications we are interested in did not reach 1% of
the population
• So we couldn’t recruit sample from a managed panel and
evaluate everything we wanted to evaluate

¾ Solution: recruit survey sample from runcactuskidrun.com via survey

overlay

Google Confidential and Proprietary

New exposure methodology

• Problem 2

• “Do you recognise this specific piece of advertising?”

;
:
?

¾ Solution: track the prior search clicks of survey respondents using cookie
technology

Google Confidential and Proprietary

…this tailored ICE study

• Sample recruited from runcactuskidrun.com

• Recruited during campaign (mid Aug - early Sep)

• 651 gross respondents

¾ Results in terms of relative not absolute impact of communications

¾ Results on a per person exposed basis

Google Confidential and Proprietary

Modelling key factors
Oasis is available nearly everywhere
Oasis is good value for money
Convenience
Oasis is a drink to have on the go

Oasis is aimed at people like me

Affinity Oasis fits in with my lifestyle

Oasis tastes great

Taste Oasis comes in flavours I like
I prefer the taste of Oasis to mineral water
Oasis is British
Heritage Oasis has been around for a long time
Oasis is more thirst quenching than water
Oasis is thirst quenching
Thirst Quenching Oasis tastes artificial (-ve)
Oasis offers something different to other soft drinks
Oasis is healthy in comparison to other soft drinks
Stand Out Oasis bottle are a distinctive shape
I would talk about Cactus Kid advertising with my friends
Brand You see a lot of advertising around for Oasis
I like the recent Oasis ‘Cactus Kid’ advertising
Fame Oasis is quirky

Oasis does innovative advertising

Communication Oasis does interesting advertising
I like Oasis advertising

Google Confidential and Proprietary

Google Confidential and Proprietary
KPI Drivers

100%
WOM
90%
80%
70% Media
60%
50%
40%
30%
Experience
20%
10%
0%
Brand Fame

Google Confidential and Proprietary

Advertorials and Paid Search drive engagement
Positive B rand Engagem ent Index

1.56
1.5

1.00
1

0.5
0.32

0.07
0
Online Advertorials TV Google paid search Email

Google Confidential and Proprietary

Oasis media driving Brand Fame
Positive Brand Engagem ent Index
0.45 0.42
Paid search was the 4th most effective
0.4 individual piece of communication

0.35
0.31
0.3

0.25

0.2
0.15
0.15 0.14
0.12
0.09
0.1
0.04
0.05
0.01
0
Online T V - D iner O nline G o o gle paid T V - M o tel Email - Email - So cial
A dvert o rial Ad A dverto rial search click Ad H eat Wo rld Z o o T o day M edia -
- T he Sun - Z o o T o day E-news M yspace

Google Confidential and Proprietary

Summary

“By helping us understand the effectiveness of the digital part of the campaign
the ICE research will help us plan for even better digital campaigns in future”.

Linda Neville

Strategic Planner Coca Cola Great Britain

“It is clear from this study that search has an important role in branding
campaigns beyond simply generating enquiries. It has the power to deepen
brand engagement especially alongside other media”.

Mark Greenstreet

MD Aegis ævolve

Google Confidential and Proprietary

Brand Effects of ad impressions ?

22
Google Confidential and Proprietary 22
Brand Effects of Paid and Organic Search

Objective

Can search help build brands?

• To understand the relative efficacy of both

paid and organic results on a Google
SERP and to quantify the branding and
multiplier effect of paid (ppc) advertising.

Google Confidential and Proprietary 23

 Methodology
The Study was undertaken by Enquiro
Research in partnership with Ipsos Mori.
All fieldwork took place online

Respondents randomly assigned to matched test groups of n=250

each cell created was representative of the online population

Top sponsored

Respondents each given identical search tasks

Side sponsored

Exposure to search listings was controlled – Each cell

was exposed to a different combination of listings
Top sponsored
& Top organic

Duration of exposure was not limited and controlled by

user click thru on the page

Respondents were then asked a questionnaire to

ascertain their awareness, liking, purchase !"#$%!"#
consideration and take out of relevant test brands

Google Confidential and Proprietary

Search can drive brand awareness
Brand Awareness - Generic Keyword Query
Which companies from the list below are you aware of?

• Similar to results for other countries, the brand awareness for generic queries is
highest for top organic and top sponsored ad positions.
• An increase of 5% for Top sponsored vs Top organic and 8% when organic and Top
sponsored co-occur.

82%
79%

73% 74%

69%

Side Sponsored Top Organic and Side Top Organic Top Sponsored Top sponsored and
sponsored Organic

Google Confidential and Proprietary 25

Search can help reinforce brand qualities

Brand communication - Generic Keyword Query

Which of the following statements do you associate with Optivita?

• Search can be highly effective in brand communication. Being in a top sponsored

position generates a significant uplift in brand qualities
• Search in this case is most probably the key which unlocks a consumers cognitive
bank of knowledge about a brand

54%
53%
47%
45%

37%

28%
25%

16%

Low in Fat Healty Nutritious A brand I can trust

Control Top Sponsored

Google Confidential and Proprietary 26

...and most significantly also drives a increased
purchase consideration
Purchase consideration - Generic Keyword Query
From the list below, indicate which of these brands you would consider if you were looking to buy ..[product]

• Even at the latter end of the purchase • Limited advantage • Significant gains
funnel search still has both very strong gained by being top of though when a brand
absolute and relative effects on the organic listings is both top sponsored
purchase consideration. and top organic

68% 68%

63% 64%

59% 59% 60%

89,)":
54% 55% 54%

48%
47%

Control Side Sponsored Top Organic Side Sponsored Top Sponsored Top Sponsored
and Top Organic and Top Organic

Optivita Other

Google Confidential and Proprietary 27

Summary

1 Search can drive traditional brand metrics across all stages of the
purchase funnel

2 The greater the equity on the search engine results page the
greater potential lift in brand metrics

3 The adaptability of Paid search lets you communicate and

reinforce brand qualities throughout the purchase process

Beyond the absolute gains - relative gains against competitors

4 should be the most compelling argument for advertisers to adopt
search as a brand channel

Google Confidential and Proprietary 28

Understanding the Consumer:The Online Path to Purchase

Google Confidential and Proprietary

Study Objective: Understand the online journey

• How much do people research online?

• How long does the research process take?
• How many searches and site visits?
• What sorts of search / site visit?
• To what extent do consumers shop around?

Methodology

• Data mining of comScore’s Mediametrix online user panel

• Selected a competitive set of websites in a category (apparel, mobiles,
computers)
• Identified people who have made an online purchase or used a store
locator between 15 November and 31st December 2007
• Analyzed activity up to 12 weeks prior to the last conversion (online
sale/locator use) observed
30
Conversion Types Per Category
For each category, we examined 10-12 leading sites (e.g. for computer purchases:
tesco.com, amazon.co.uk, currys.co.uk, comet.co.uk, pcworld.com, apple.com)

-(.'(%2D#H7"B% I&3#%43H"&3#

@3C:D&%#B =, ;?,

!3A'.%273(%B <=, ??,

@.3&7'() ;9, >9,

9, >9, 89, ?9, E9, F9, <9, =9, G9, ;9, >99,

Source: comScore custom analysis- UK Population, 15+ Home and Work 31

Q4-2007 activity
 Computer Case Study – “Coloured or cheap?”

From “PC World” to purchase at ebuyer in 66 days

First Search Purchase

PC World eBuyer

currys
Searches

Dell laptops
PC World
Tablet laptops Coloured
Pink PC World laptops PC World
laptops laptops
laptop toshiba
Cheap
Coloured Pink
laptops portege
laptop laptops
m400-11h
comet

PCWorld.co.uk Comet.co.uk eBay

Pixmania Misco
PCWorld.co.uk Pixmania Currys.co.uk computers
Dell Currys PCWorld.co.uk PCWorld.co.uk
PCWorld.co.uk
Visits

Dell Amazon Shopping.com Kelkoo Bizrate Amazon ebuyer Sony Style

Dixons.co.uk PriceRunner
eBay Bizrate Dell Nextag
DealTime.co.uk computers Dell Dell ebuyer
eBay MicroDirect
Lenovo
Comet.co.uk computers Pixmania
Toshiba

Day 0 Day 15 Day 22 Day 26 Day 30 Day 37 Day 43 Day 47 Day 48 Day 62 Day 66

Source: comScore custom analysis- UK Population, 15+ Home and Work 32

Q4-2007 activity
 47% of computer conversions occur after 30 days or more

Share of Conversions by Time Period from First Search

Conversions beyond 30
>G, days from first search

><,

>E,

>8, @3C:D&%#B
!3A'.%:73(%B
>9,
O::"#%.
G,

<,

E,

8,

9,
Time from First J'#B& M> M8 M? ME MF M< M= MG M; M>9 M>> M>8
Search K"L N%%$ N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B N%%$B

Purchases & research tracked at a leading set of computer retailers’ websites – Apple.com, Amazon.co.uk, Comet.co.uk, Currys.co.uk, Dabs.com,
Dixons.co.uk, Dell.com, Ebuyer.com, PCWorld.co.uk, Tesco.com
Source: comScore custom analysis- UK Population, Q4 2007; 33
Use of generic terms varies dramatically by category

Proportion of conversions with generic search terms

@.3&7'()
!3A'.%273(%B
=>,
@3C:D&%#B

2#3:3#&'3(@3(P%#B'3(B5'&7
EE,
"&.%"B&3(%Q%(%#'HI%"#H7

>8,

Source: comScore custom analysis- UK Population, 15+ Home and Work 34

Q4-2007 activity
Summary

The online journey is long and complicated, search is integral to the process
1 and consumers shop around a lot.

Research behaviour is highly category specific, particularly in the use of

2 generic search terms.

Current ROI tracking on paid search needs improving. Recommendation:

3 track leads to conversions for longer, don’t just measure conversions on the
last click.

35
Exploring engagement and branding on
YouTube and TV
Study Methodology and Objectives

YouTube partnered with two advertisers, Motorola (with Mindshare) & GM to:

1. Quantify user engagement on YouTube and television

2. Measure the branding value of ads on YouTube and TV
3. Understand the combined impact of YouTube and TV advertising on brand
perceptions and consideration

Methodology

Biometrics Eye tracking Depth interviews Online survey

(48 people) (24 people) (48 people) (1,800 people)

All samples were representative of both advertisers’ target audience

(18-29 years)

37
 Test Stimuli
TV Ad YouTube Ad Impression YouTube Video Ad Click YouTube Viral Video

so
un
Hompage
d)
silent
autoroll
Corsa

InVideo
banner
Motorola

38
 YouTube ads stand out and are noticed without
interrupting the user experience

Eye-tracking shows that autoroll InVideo ads appear on uncluttered watch

homepage ads become the hottest pages and are closest to the viewing
part of the homepage hotspot

Users spend
50% longer
viewing the
companion
display
banner than
standard
display ads*

Ads receive attention without annoying users: the majority of users say the ads on
YouTube are non-obtrusive and do not interrupt the user experience.

"I really like it. You notice it but it doesn't get in the way, and if it was relevant to
what I was watching then that would be even better” Male, 22, UK 39

* Compared to average ad dwell time measured in an Eyetools study, 2004.

 Users are 1.5x more attentive when browsing YouTube than when
watching TV

Biometrics data* collected whilst

respondents watched TV and
browsed YouTube demonstrates:
• Attention levels are consistently
higher - 1.5x overall - whilst
browsing YouTube than when

Intensity (attention)
watching TV
• This confirms the value of YouTube
as a lean forward platform, where
consumers actively chose the
content they watch
• In addition, across all tested ads,
engagement on YouTube is
amplified ~140% over TV

* Biometrics data measured: Time (seconds)

Heart rate
Physical movement
Respiratory rate
Skin conductance

40

Source: Biometric data collected Sept 2008 in the UK. TV respondents were watching In Plain Sight a US FBI drama. Web respondents were surfing on YouTube
 Combining YouTube with TV lifts perception of Motorola’s
core brand attributes above TV alone
• For Motorola, the addition of YouTube ad impressions drives brand metrics up to 4x higher
than TV alone. For GM Corsa, metrics are driven up to 3x higher
• Advertisers can also take advantage of the YouTube brand halo effect to lift attributes not
communicated in the campaign

Motorola: indexed uplift in brand association with attribute (from control)

Indexed uplift in brand association (from control)

% associating attribute with brand

Q. Looking at this list of phones, which would you say is… You can click on as many or as few brands as you wish. (n=300) * Significant uplift from control 41
GM YouTube Video

Google Confidential and Proprietary

2009 Research Themes
Search influences decisions to buy Search & online turbo-charges your marketing

Search and online help build brands effectively YouTube has a unique supporting role in the media mix

Google Confidential and Proprietary 43