SOCIAL INFLUENCE ANALYSIS:
A NEW INSIGHT TERRITORY
Anthony Fradet - Linkfluence
UK Chief Operating Officer
@anthonyfradet
Luis Fernandes - Expedia
EMEA Research and Consumer
Insights Director
The mission of the researcher
QUANT QUAL
The traditional toolkit
QUAL
The new toolkit
SOCIALQUANT
Qualitative data
on a quantitative scale
In vivo (vs in vitro)
Social data: qualitative data on a quantitative scale
Vast collection
of individual opinions
Identification of major trends
and weak signal
Wide possibilities
from social KPI measurement
to in-depth content analysis
Social data: in vivo (vs in vitro)
Untamed, spontaneous material
Opinions expressed reach
an audience, triggering tangible
influence mechanisms
No interrogation bias
What makes research methodologies fit
for insight generation?
Framed Analysis
Business Question
Traditional research
Quotas
Samples
Representativeness
Discussion guides
Valid Insight
Structured / bias-controlled data
Quotas
Samples
Representativeness
Discussion guides
Unstructured / biased data
Framed Analysis
Social media research (for him)
“Lack of control on
key parameters”
Business Question
Valid Insight
Social media research
Structured
data
Unstructured
data
Business Question
Valid InsightFramed Analysis
1) Definition of the data universe
(semantic profiles)
2) Data segmentation
3) Data contextualisation
through relevant social analysis frames
1) Definition of the data universe
(semantic profiles)
2) Data segmentation
3) Data contextualisation
through relevant social analysis frames
Social influence is
an element of context unique to social data
Social influence?
An individual’s ability to affect other people's
thinking in an online community
Numbers?
We are in the age of the tyranny of numbers
Numbers are good liars
Social influence goes beyond numbers,
it is 100% contextual and based on
relationships and influence dynamics
Understand
the context
TECHNOLOGY
(RADARLY)
CLIENT
SERVICES
MARKETING
INTELLIGENCE
01
UK POLITICAL WEB OBSERVATORY
HOW NETWORK ECOSYSTEM ANALYSIS
PREDICTED THE EU REFERENDUM RESULT
Published early June 2016
www.politicalweb.co.uk
www.politicalweb.co.uk
Colours refer to
communities
Nodes are websites
Links are hypertext
links between two
websites
(inbound/outbound)
Node size indicates
the level of
influence of a
website
02
DANONE CITIES
HOW SOCIAL AUDIENCE MAPPING HELPS
UNDERSTAND CONSUMER MINDSET
2 objectives
Identify the key communities of interest
structuring the social landscape in 4 major cities
Recommend smart actions to engage
these communities in a meaningful way
How to capture the social conversation
happening in a city and put it into context?
Can we trust individuals to tell the truth
about what they are influential on?
Good proportion of what I post
Good proportion of what I post
Almost never post
Post rarely
Interest is not influence
To make sense of audiences,
content analysis at scale is necessary
The key
Organising individuals in communities based
on the topics they really talk about
(not what they say they are)
Topic Drops
Key outcomes for Danone
Sharp understanding of cultural background
Identification of relevant content area
Food for innovation in brand initiative
03
EXPEDIA - DIGITAL FOOTPRINT
HOW HIGHLY CONTEXTUALISED INFLUENCE
ANALYSIS INFORMED EXPEDIA’S
INFLUENCER STRATEGY
Luis Fernandes - Expedia
EMEA Research and Consumer
Insights Director
Modes
&
context
121
Business objective
Increase site visits
and number of transactions
How to achieve this?
Increase brand presence throughout
the decision making process prior to booking
Identify the right influencers
for the right mode at the right time
How is the travel ecosystem structured
in different markets?
The UK Travel Ecosystem
Colours refer to
communities
Nodes are websites
Links are hypertext
links between two
websites
(inbound/outbound)
Node size indicates
of the level of
influence of a
website
Travel Online Ecosystem UK vs Germany
1,685	websites
25	sub-communities
2,051	websites
Travel Online Ecosystem UK vs Germany
Travel	News
1%
Sports
2%
Family
4%
Innovation	&	Tech
4%
Business	&	Finance
5%
Food	&	Drinks
10%
Lifestyle
11%
Travel
12%
Beauty	&	Fashion
15%
Online	Media
22%
Travel
31%
Online	Media
15%
Lifestyle 9%
Food	&	Drinks
7%
Family
5%
Beauty	&	Fashion
4%
Business	&	Finance
4%
Culture	&	
Entertainment 3%
Travel Online Ecosystem UK vs Germany
Travel	News
1%
Sports
2%
Family
4%
Innovation	&	Tech
4%
Business	&	Finance
5%
Food	&	Drinks
10%
Lifestyle
11%
Travel
12%
Beauty	&	Fashion
15%
Online	Media
22%
Travel
31%
Online	Media
15%
Lifestyle 9%
Food	&	Drinks
7%
Family
5%
Beauty	&	Fashion
4%
Business	&	Finance
4%
Culture	&	
Entertainment 3%
Travel Online Ecosystem UK vs Germany
Travel	News
1%
Sports
2%
Family
4%
Innovation	&	Tech
4%
Business	&	Finance
5%
Food	&	Drinks
10%
Lifestyle
11%
Travel
12%
Beauty	&	Fashion
15%
Online	Media
22%
Travel
31%
Online	Media
15%
Lifestyle 9%
Food	&	Drinks
7%
Family
5%
Beauty	&	Fashion
4%
Business	&	Finance
4%
Culture	&	
Entertainment 3%
How to integrate context into affluence
and make it operational?
Contextualised influencer identification
Social audience
Monthly visits and
views
Potential
for working
with
brands
Relevance for
Expedia
Influence
score*
*Influence score: Score (from 0 to 100) based on the inbound links received by the websites within the ecosystem.
Blogs with a
high influence
score*
Sites with
strong social
following
Identification of
matching
Expedia Travel
Mode
Travel modes
in the UK travel ecosystem
• TRAVEL MODE A, B & D are the three main travel
modes in the Travel Ecosystem, with 83% of the
top influencers within the UK Travel ecosystem.
• These three travel modes provided most traffic.
Travel mode B gathers the stronger follower base,
with an average 880k followers.
• The most influential opinion leaders:
AVERAGE	MONTHLY	VISITS AVERAGE	SOCIAL	FOLLOWERS
TRAVEL MODE A 368K 197K
TRAVEL MODE B 35,5k 880K
TRAVEL MODE C 71,4k 84,9K
TRAVEL MODE D 103K 295K
TRAVEL MODE E 13,8K 18,1K
TRAVEL MODE F 95,3K 46,2K
Key influencers
Cumulated
social audience
Dominant Travel
Mode
25.5m à TRAVEL MODE A
4.2m à TRAVEL MODE B
1.8m à TRAVEL MODE C
1.5m à TRAVEL MODE B
1.4m à TRAVEL MODE D
31%
26%
26%
10%
10%
7%
Share	of	influencers	
by	travel	mode
Travel mode C
What kind of content do they produce?
• Visual posts, documenting breaks that are largely relaxation,
culture or luxury focused (e.g. warm destinations, spa days,
boutique hotels)
Place of travel within their life
Travels to relax mentally and physically
Booking Behaviour
• Destination is less important than an ideal resort. Likely to book all
inclusive packages, sometimes offline
• Want hassle-free booking
Other interests
Food | Fashion | Style | Weekend breaks |
Health
103k MONTHLY SITE VISITS ON AVERAGE
295k SOCIAL FOLLOWERS PER INFLUENCER (AGGREGATED AVERAGE)
Outcome for Expedia:
Involvement in contextual conversations
- Opportunities to develop more relevant content for specific
context or simply to share it in the right place…
- Key influencers to work with beyond number of followers
Two things to remember…
QUAL
The new toolkit
SOCIALQUANT
One thing to remember
Context
is
everything
Anthony Fradet - Linkfluence
UK Chief Operating Officer
@anthonyfradet
Luis Fernandes - Expedia
EMEA Research and Consumer
Insights Director

"Social influence analysis: a new insight territory"

  • 1.
    SOCIAL INFLUENCE ANALYSIS: ANEW INSIGHT TERRITORY Anthony Fradet - Linkfluence UK Chief Operating Officer @anthonyfradet Luis Fernandes - Expedia EMEA Research and Consumer Insights Director
  • 2.
    The mission ofthe researcher
  • 3.
  • 4.
  • 5.
    Qualitative data on aquantitative scale In vivo (vs in vitro)
  • 6.
    Social data: qualitativedata on a quantitative scale Vast collection of individual opinions Identification of major trends and weak signal Wide possibilities from social KPI measurement to in-depth content analysis
  • 7.
    Social data: invivo (vs in vitro) Untamed, spontaneous material Opinions expressed reach an audience, triggering tangible influence mechanisms No interrogation bias
  • 9.
    What makes researchmethodologies fit for insight generation?
  • 10.
    Framed Analysis Business Question Traditionalresearch Quotas Samples Representativeness Discussion guides Valid Insight Structured / bias-controlled data
  • 11.
    Quotas Samples Representativeness Discussion guides Unstructured /biased data Framed Analysis Social media research (for him) “Lack of control on key parameters” Business Question Valid Insight
  • 12.
  • 14.
    1) Definition ofthe data universe (semantic profiles) 2) Data segmentation 3) Data contextualisation through relevant social analysis frames
  • 15.
    1) Definition ofthe data universe (semantic profiles) 2) Data segmentation 3) Data contextualisation through relevant social analysis frames
  • 16.
    Social influence is anelement of context unique to social data
  • 17.
    Social influence? An individual’sability to affect other people's thinking in an online community
  • 18.
  • 19.
    We are inthe age of the tyranny of numbers
  • 20.
  • 21.
    Social influence goesbeyond numbers, it is 100% contextual and based on relationships and influence dynamics
  • 22.
  • 24.
  • 25.
    01 UK POLITICAL WEBOBSERVATORY HOW NETWORK ECOSYSTEM ANALYSIS PREDICTED THE EU REFERENDUM RESULT
  • 26.
    Published early June2016 www.politicalweb.co.uk
  • 27.
    www.politicalweb.co.uk Colours refer to communities Nodesare websites Links are hypertext links between two websites (inbound/outbound) Node size indicates the level of influence of a website
  • 31.
    02 DANONE CITIES HOW SOCIALAUDIENCE MAPPING HELPS UNDERSTAND CONSUMER MINDSET
  • 32.
    2 objectives Identify thekey communities of interest structuring the social landscape in 4 major cities Recommend smart actions to engage these communities in a meaningful way
  • 33.
    How to capturethe social conversation happening in a city and put it into context?
  • 34.
    Can we trustindividuals to tell the truth about what they are influential on?
  • 35.
    Good proportion ofwhat I post Good proportion of what I post Almost never post Post rarely Interest is not influence
  • 36.
    To make senseof audiences, content analysis at scale is necessary
  • 37.
    The key Organising individualsin communities based on the topics they really talk about (not what they say they are)
  • 38.
  • 39.
    Key outcomes forDanone Sharp understanding of cultural background Identification of relevant content area Food for innovation in brand initiative
  • 40.
    03 EXPEDIA - DIGITALFOOTPRINT HOW HIGHLY CONTEXTUALISED INFLUENCE ANALYSIS INFORMED EXPEDIA’S INFLUENCER STRATEGY Luis Fernandes - Expedia EMEA Research and Consumer Insights Director
  • 41.
  • 42.
  • 43.
    Business objective Increase sitevisits and number of transactions
  • 44.
    How to achievethis? Increase brand presence throughout the decision making process prior to booking Identify the right influencers for the right mode at the right time
  • 45.
    How is thetravel ecosystem structured in different markets?
  • 46.
    The UK TravelEcosystem Colours refer to communities Nodes are websites Links are hypertext links between two websites (inbound/outbound) Node size indicates of the level of influence of a website
  • 47.
    Travel Online EcosystemUK vs Germany 1,685 websites 25 sub-communities 2,051 websites
  • 48.
    Travel Online EcosystemUK vs Germany Travel News 1% Sports 2% Family 4% Innovation & Tech 4% Business & Finance 5% Food & Drinks 10% Lifestyle 11% Travel 12% Beauty & Fashion 15% Online Media 22% Travel 31% Online Media 15% Lifestyle 9% Food & Drinks 7% Family 5% Beauty & Fashion 4% Business & Finance 4% Culture & Entertainment 3%
  • 49.
    Travel Online EcosystemUK vs Germany Travel News 1% Sports 2% Family 4% Innovation & Tech 4% Business & Finance 5% Food & Drinks 10% Lifestyle 11% Travel 12% Beauty & Fashion 15% Online Media 22% Travel 31% Online Media 15% Lifestyle 9% Food & Drinks 7% Family 5% Beauty & Fashion 4% Business & Finance 4% Culture & Entertainment 3%
  • 50.
    Travel Online EcosystemUK vs Germany Travel News 1% Sports 2% Family 4% Innovation & Tech 4% Business & Finance 5% Food & Drinks 10% Lifestyle 11% Travel 12% Beauty & Fashion 15% Online Media 22% Travel 31% Online Media 15% Lifestyle 9% Food & Drinks 7% Family 5% Beauty & Fashion 4% Business & Finance 4% Culture & Entertainment 3%
  • 51.
    How to integratecontext into affluence and make it operational?
  • 52.
    Contextualised influencer identification Socialaudience Monthly visits and views Potential for working with brands Relevance for Expedia Influence score* *Influence score: Score (from 0 to 100) based on the inbound links received by the websites within the ecosystem. Blogs with a high influence score* Sites with strong social following Identification of matching Expedia Travel Mode
  • 53.
    Travel modes in theUK travel ecosystem • TRAVEL MODE A, B & D are the three main travel modes in the Travel Ecosystem, with 83% of the top influencers within the UK Travel ecosystem. • These three travel modes provided most traffic. Travel mode B gathers the stronger follower base, with an average 880k followers. • The most influential opinion leaders: AVERAGE MONTHLY VISITS AVERAGE SOCIAL FOLLOWERS TRAVEL MODE A 368K 197K TRAVEL MODE B 35,5k 880K TRAVEL MODE C 71,4k 84,9K TRAVEL MODE D 103K 295K TRAVEL MODE E 13,8K 18,1K TRAVEL MODE F 95,3K 46,2K Key influencers Cumulated social audience Dominant Travel Mode 25.5m à TRAVEL MODE A 4.2m à TRAVEL MODE B 1.8m à TRAVEL MODE C 1.5m à TRAVEL MODE B 1.4m à TRAVEL MODE D 31% 26% 26% 10% 10% 7% Share of influencers by travel mode
  • 54.
    Travel mode C Whatkind of content do they produce? • Visual posts, documenting breaks that are largely relaxation, culture or luxury focused (e.g. warm destinations, spa days, boutique hotels) Place of travel within their life Travels to relax mentally and physically Booking Behaviour • Destination is less important than an ideal resort. Likely to book all inclusive packages, sometimes offline • Want hassle-free booking Other interests Food | Fashion | Style | Weekend breaks | Health 103k MONTHLY SITE VISITS ON AVERAGE 295k SOCIAL FOLLOWERS PER INFLUENCER (AGGREGATED AVERAGE)
  • 55.
    Outcome for Expedia: Involvementin contextual conversations - Opportunities to develop more relevant content for specific context or simply to share it in the right place… - Key influencers to work with beyond number of followers
  • 56.
    Two things toremember…
  • 57.
  • 58.
    One thing toremember Context is everything
  • 59.
    Anthony Fradet -Linkfluence UK Chief Operating Officer @anthonyfradet Luis Fernandes - Expedia EMEA Research and Consumer Insights Director