How to Optimize Your
Marketing Strategy Using
Facebook Topic Data
February 25th • 2016
Live Webinar with Q&A
Tim Shea
Agency, Brand and Data Partnerships (US)
DATASIFT
James Ritchie
Agency, Brand and Data Partnerships (EMEA)
DATASIFT
Dary Hsu
Product Marketing Manager
DATASIFT
SPEAKERS
What is Facebook Topic Data?
Agenda
1
Use Cases2
Q&A3
What is Facebook Topic
Data?
52
%21%
13%
4%
10%
Facebook is the #1 Social Platform for Marketers
5
Source: 2015 Social Media Marketing Industry Report, May 2015
OTHERS
“THE MOST IMPORTANT
SOCIAL PLATFORM FOR MARKETERS
IS…”
For years, companies struggled to get a complete view of their
audience on Facebook and turn that information into useful insights
until….
DATASIFT + FACEBOOK Partnership
ENGAGEMENT ACROSS FACEBOOK
FACEBOOK TOPIC DATA
Topic Data Unlocks Unique Insights for Marketers
What is Facebook Topic Data?
8
What’s on your mind?
CONTENT DEMOGRAPHICS LIKES and SHARES
Anonymized and aggregate topic data
• Posts
• Pages Posts
Plus engagement data
• Likes on Posts
• Shares on Posts
• Comments (no text) on Posts
Data enriched with
• Demographics
• Topics
• Sentiment
• Real-time access to the entire newsfeed with over 4.75 billion pieces of content shared a day.
• Gain anonymous & aggregated insights about specific activities, events, brand names, and other
subjects that people are sharing on Facebook.
Not a Data Feed. Topic Data is Aggregated and
Anonymized
9
New approach to provide privacy-first insights:
Facebook is not a public social network.
User identity is removed from posts and engagement data
processing.
Text and meta data from anonymized posts are indexed
within Facebook’s infrastructure for analysis.
Developers query data collected in real-time to perform
analysis. Data is aggregated at query time to provide
aggregate results.
Privacy controls ensure results only provided if audience size
thresholds are met.
Insights From a Network of 1.59 Billion People
WITHOUT FACEBOOK TOPIC DATA + FACEBOOK TOPIC DATA
Analysis across public social data sources
Example: Analysis of automotive brand
6x
Analysis includes Twitter, Tumblr, blogs, forums.10
11
Brand Analytics
How Companies Are Using Topic Data
BRAND • PRODUCT CONTENT • LINKS INDUSTRY • TOPIC AUDIENCE
Content & Media Analysis Industry & Topic Research
Market research to inform creative &
campaigns
Brand Reputation Management
Campaign Analysis
Competitive Analysis
Influential Media Analysis
Topic Analysis
Content Discovery
Industry Benchmarking
Topic-Specific Analysis
Vertical Applications (e.g. TV)
Creative & Campaign Design
Audience Affinity Analysis
Audience Discovery/Expansion
Ad Agency
Cross-shopping For Car
Brands
13
The Problem
The agency struggled to quantify how & when
consumers were “cross-shopping” between
competitors' vehicles, what motivated them to do
so, and where they were in the purchase funnel.
They also lacked the ability to measure cultural
differences and care-abouts between Spanish &
English speaking audiences.
Our Approach
๏ DataSift developed 1 filter focused on the brand, its
models & competitors, as well as car features. The
second filter focused on understanding cross-
shopping behavior.
๏ DataSift used VEDO to classify 10 automobile
features across 6 different languages, as well as
purchase intent into consideration, rejection, or
ownership.
๏ The brand index captured 180k interactions in 3
weeks.
Ad Agency Car Brands
Using topic data the agency was able to
uncover multiple audience insights for their
client:
๏ Significant differences between Spanish &
English-speaking audiences for various
automobile features. For example, style & price
barely showed up within Spanish conversation.
๏ How often the client brand was mentioned together
with a competitor brand as an indication of cross-
shopping. We also enabled them to quantify what
features were being compared as well as what was
motivating the comparison.
๏ Client brand had the lowest level of brand
consideration, whereas another competitor
enjoyed zero brand rejection.
Recommended Actions
๏ Emphasize different features and vernacular when advertising in Spanish versus English.
๏ Understand how to position against competition and quantify changes in risk to market share.
Ad Agency Car Brands
CROSS
SHOPPING
Competitor
C
5%
3%
10%
TOP 3 FEATURES ENGLISH VS. SPANISH
ESPECIFICACIÓN MECÁNICA
AMBIENTE
COSTO DE FUNCIONAMENTO
STYLE
PRACTICALITY
PERFORMANCE
Global Brand
Crisis Management
16
The Problem
The global brand was subject to a protest
campaign driven by a small number of protest
groups, whose campaign video was “going viral”.
The brand could measure engagement on other
social networks but they lacked demographic
insight and the ability to measure the scale of the
backlash on Facebook.
Our Approach
๏ DataSift developed 1 filter focused on the crisis -
tracking engagement with the protest video,
references to the crisis and links to the global
brands’ websites.
๏ DataSift used VEDO to classify interactions which
linked to stories already identified as influential.
Another classifier tagged the key consumer brands.
๏ The index captured 40k interactions in 8 days.
Global Brand - Crisis Management
Through analysis of topic data the brand was
able to uncover the following insights:
๏ The brand could see how different demographic
groups engaged in the issue, with women over
65 over-indexing massively.
๏ The brand had been focused on stories with
some traction on other social networks but was
unaware of stories in France & Germany which
were driving much larger engagement on
Facebook.
๏ They could identify which product brands were
being most targeted for boycott.
Recommended Actions
๏ By understanding the demographics of people actually engaging in the crisis the brand could develop better targeted messaging.
๏ Identify specific product brands needing additional defence / support.
CPG Brand - Crisis
Management
MISSING THE BIGGER PICTURE
AVERAGE ENGAGEMENT
FEMALE
65 +
Ad Agency
Audience Research
Ad Agency For Drinks Brand
The Problem
Ad agency wanted to understand how women
engaged with their client’s brand and with hot
drinks. They also wanted to get a deeper
understand of the media consumption / magazine
stories which most engaged their target consumers
Our Approach
๏ DataSift developed 2 filters focused around the
campaign. The first filter focused on the brand, its
key competitors and on hot drinks. The 2nd index
focused on engagement with links to 10 top female
magazines
๏ DataSift used VEDO to categorize hot drinks
(across multiple languages) and to tag magazines
based on URLs
๏ The 1st index captured 2.5m interactions in 12
days. The 2nd index captured 950k interactions in
16 days
Under 35s, as a % of brand’s total
engagement
47%
The brand was able to derive multiple
actionable creative insights using topic data:
๏ There are clear variations in preferences for
hot drinks across nations and demographic
groups
๏ The brand found that they suffered with the
lowest relative engagement amongst
millennials
๏ We could see the publications, stories &
celebrities which drove most engagement
amongst the target demographic
Recommended Actions
๏ Look at featuring different hot drinks when advertising in different markets and to different demographic groups
๏ Use story insight for media placement as well as for identifying potential influencers
CPG Brand
1st
2nd
3rd
LATTE
HOT CHOCOLATE
ESPRESSO
HOT CHOCOLATE
LATTE
CAPPUCINO
CAPPUCINO
LATTE
ESPRESSO
USA UK GERMANY
Client Brand Competitor A Competitor B
31%
58%
18-24
25-34
35-44
45-54
55-
64
65+
JENNIFER LAWRENCE
SHARON STONE
TAYLOR SWIFT
KYLIE JENNER
JUSTIN TIMBERLAKE
Celebrity stories driving most engagement
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 1000 10500 11000 11500 12000 12500 13000 13500 14000 14500
CPG Brand
Audience Research
CPG Brand
The Problem
A CPG brand was running a 2nd iteration of
their flagship campaign. They wanted new
ways to understand how people engaged with
the campaign and its underlying theme on
Facebook. They also wanted to develop
creative insight to support future campaigns.
Our Approach
๏ DataSift developed 2 filters focused around
the campaign. The first filter focused on the
campaign itself: the underlying theme as well
as hashtags and urls. The second filter
focused on how the campaign tag line was
used in normal discussions
๏ DataSift used VEDO to categorize themes and
phrases
๏ The 1st index captured 6.9m interactions in 7
days. The 2nd index captured 461k
interactions in 14 days.
The brand was able to derive audience and
creative insights using topic data:
๏ The campaign had targeted a theme which
had much less engagement than the
themes of pride or determination.
๏ When people engage with these more
popular themes, the majority of links they
share are for charity runs or causes.
๏ We could track the volume (and
demographics) of engagement in phrases
the brand had previously featured and
identify new phrases which should
resonate with their target audience.
Recommended Actions
๏ Look at targeting different themes and phrases, which are proven to resonate with target audience
๏ Consider sponsoring charity runs / events as part of future marketing mix
CPG Brand
2,481.900
299,000
Campaign theme
Determination
Pride
Top websites linked when sharing emotional
state
CHARITY RUNS
CHARITY RAISING FOR INDIVIDUALS
OTHER
11,300
CPG Brand
Audience Research
CPG Brand
The Problem
A CPG company believed that Millennials are
fundamentally different to older consumers in
their attitudes to food. They wanted to use
Facebook to measure and understand these
differences so they could develop new
creative campaigns to better engage
Millennial consumers.
Our Approach
๏ DataSift developed 2 filters focused around
the campaign. The first filter focused on
general engagement with food as a topic. The
2nd filter focused on the specific food trends
the company had already identified
๏ DataSift used VEDO to categorize themes,
trends, consumption locations and meals
๏ The 1st index captured 13.2m interactions in
6 days. The 2nd index captured 11m
interactions in 32 days.
The brand was able to derive multiple
actionable creative insights using topic data:
๏ We could see very significant different levels
of engagement in most food trends between
young millennials (18-24) and older millennials
(25-34) but little distinction between
millennials versus older consumers
๏ Small differences in the media types
millennials post (more photos) but a negligible
difference in the content they engage with
when discussing food
๏ Identified a range of national difference in
engagement with food themes and meals
Recommended Actions
๏ Don’t treat Millennials as a homogenous group - in many ways they are like the rest of us!
๏ Optimize campaign messaging to reflect national engagement in different food trends
CPG Brand
Media type when posting about
food
Engagement on media type
about food
American Males engagement
in Reduced Meat
British Females engagement
in Natural Food
Q&A
THANK YOU!

How to Optimize Your Marketing Strategy using Facebook Topic Data

  • 1.
    How to OptimizeYour Marketing Strategy Using Facebook Topic Data February 25th • 2016 Live Webinar with Q&A
  • 2.
    Tim Shea Agency, Brandand Data Partnerships (US) DATASIFT James Ritchie Agency, Brand and Data Partnerships (EMEA) DATASIFT Dary Hsu Product Marketing Manager DATASIFT SPEAKERS
  • 3.
    What is FacebookTopic Data? Agenda 1 Use Cases2 Q&A3
  • 4.
    What is FacebookTopic Data?
  • 5.
    52 %21% 13% 4% 10% Facebook is the#1 Social Platform for Marketers 5 Source: 2015 Social Media Marketing Industry Report, May 2015 OTHERS “THE MOST IMPORTANT SOCIAL PLATFORM FOR MARKETERS IS…”
  • 6.
    For years, companiesstruggled to get a complete view of their audience on Facebook and turn that information into useful insights until….
  • 7.
    DATASIFT + FACEBOOKPartnership ENGAGEMENT ACROSS FACEBOOK FACEBOOK TOPIC DATA Topic Data Unlocks Unique Insights for Marketers
  • 8.
    What is FacebookTopic Data? 8 What’s on your mind? CONTENT DEMOGRAPHICS LIKES and SHARES Anonymized and aggregate topic data • Posts • Pages Posts Plus engagement data • Likes on Posts • Shares on Posts • Comments (no text) on Posts Data enriched with • Demographics • Topics • Sentiment • Real-time access to the entire newsfeed with over 4.75 billion pieces of content shared a day. • Gain anonymous & aggregated insights about specific activities, events, brand names, and other subjects that people are sharing on Facebook.
  • 9.
    Not a DataFeed. Topic Data is Aggregated and Anonymized 9 New approach to provide privacy-first insights: Facebook is not a public social network. User identity is removed from posts and engagement data processing. Text and meta data from anonymized posts are indexed within Facebook’s infrastructure for analysis. Developers query data collected in real-time to perform analysis. Data is aggregated at query time to provide aggregate results. Privacy controls ensure results only provided if audience size thresholds are met.
  • 10.
    Insights From aNetwork of 1.59 Billion People WITHOUT FACEBOOK TOPIC DATA + FACEBOOK TOPIC DATA Analysis across public social data sources Example: Analysis of automotive brand 6x Analysis includes Twitter, Tumblr, blogs, forums.10
  • 11.
    11 Brand Analytics How CompaniesAre Using Topic Data BRAND • PRODUCT CONTENT • LINKS INDUSTRY • TOPIC AUDIENCE Content & Media Analysis Industry & Topic Research Market research to inform creative & campaigns Brand Reputation Management Campaign Analysis Competitive Analysis Influential Media Analysis Topic Analysis Content Discovery Industry Benchmarking Topic-Specific Analysis Vertical Applications (e.g. TV) Creative & Campaign Design Audience Affinity Analysis Audience Discovery/Expansion
  • 12.
  • 13.
    13 The Problem The agencystruggled to quantify how & when consumers were “cross-shopping” between competitors' vehicles, what motivated them to do so, and where they were in the purchase funnel. They also lacked the ability to measure cultural differences and care-abouts between Spanish & English speaking audiences. Our Approach ๏ DataSift developed 1 filter focused on the brand, its models & competitors, as well as car features. The second filter focused on understanding cross- shopping behavior. ๏ DataSift used VEDO to classify 10 automobile features across 6 different languages, as well as purchase intent into consideration, rejection, or ownership. ๏ The brand index captured 180k interactions in 3 weeks. Ad Agency Car Brands
  • 14.
    Using topic datathe agency was able to uncover multiple audience insights for their client: ๏ Significant differences between Spanish & English-speaking audiences for various automobile features. For example, style & price barely showed up within Spanish conversation. ๏ How often the client brand was mentioned together with a competitor brand as an indication of cross- shopping. We also enabled them to quantify what features were being compared as well as what was motivating the comparison. ๏ Client brand had the lowest level of brand consideration, whereas another competitor enjoyed zero brand rejection. Recommended Actions ๏ Emphasize different features and vernacular when advertising in Spanish versus English. ๏ Understand how to position against competition and quantify changes in risk to market share. Ad Agency Car Brands CROSS SHOPPING Competitor C 5% 3% 10% TOP 3 FEATURES ENGLISH VS. SPANISH ESPECIFICACIÓN MECÁNICA AMBIENTE COSTO DE FUNCIONAMENTO STYLE PRACTICALITY PERFORMANCE
  • 15.
  • 16.
    16 The Problem The globalbrand was subject to a protest campaign driven by a small number of protest groups, whose campaign video was “going viral”. The brand could measure engagement on other social networks but they lacked demographic insight and the ability to measure the scale of the backlash on Facebook. Our Approach ๏ DataSift developed 1 filter focused on the crisis - tracking engagement with the protest video, references to the crisis and links to the global brands’ websites. ๏ DataSift used VEDO to classify interactions which linked to stories already identified as influential. Another classifier tagged the key consumer brands. ๏ The index captured 40k interactions in 8 days. Global Brand - Crisis Management
  • 17.
    Through analysis oftopic data the brand was able to uncover the following insights: ๏ The brand could see how different demographic groups engaged in the issue, with women over 65 over-indexing massively. ๏ The brand had been focused on stories with some traction on other social networks but was unaware of stories in France & Germany which were driving much larger engagement on Facebook. ๏ They could identify which product brands were being most targeted for boycott. Recommended Actions ๏ By understanding the demographics of people actually engaging in the crisis the brand could develop better targeted messaging. ๏ Identify specific product brands needing additional defence / support. CPG Brand - Crisis Management MISSING THE BIGGER PICTURE AVERAGE ENGAGEMENT FEMALE 65 +
  • 18.
  • 19.
    Ad Agency ForDrinks Brand The Problem Ad agency wanted to understand how women engaged with their client’s brand and with hot drinks. They also wanted to get a deeper understand of the media consumption / magazine stories which most engaged their target consumers Our Approach ๏ DataSift developed 2 filters focused around the campaign. The first filter focused on the brand, its key competitors and on hot drinks. The 2nd index focused on engagement with links to 10 top female magazines ๏ DataSift used VEDO to categorize hot drinks (across multiple languages) and to tag magazines based on URLs ๏ The 1st index captured 2.5m interactions in 12 days. The 2nd index captured 950k interactions in 16 days
  • 20.
    Under 35s, asa % of brand’s total engagement 47% The brand was able to derive multiple actionable creative insights using topic data: ๏ There are clear variations in preferences for hot drinks across nations and demographic groups ๏ The brand found that they suffered with the lowest relative engagement amongst millennials ๏ We could see the publications, stories & celebrities which drove most engagement amongst the target demographic Recommended Actions ๏ Look at featuring different hot drinks when advertising in different markets and to different demographic groups ๏ Use story insight for media placement as well as for identifying potential influencers CPG Brand 1st 2nd 3rd LATTE HOT CHOCOLATE ESPRESSO HOT CHOCOLATE LATTE CAPPUCINO CAPPUCINO LATTE ESPRESSO USA UK GERMANY Client Brand Competitor A Competitor B 31% 58% 18-24 25-34 35-44 45-54 55- 64 65+ JENNIFER LAWRENCE SHARON STONE TAYLOR SWIFT KYLIE JENNER JUSTIN TIMBERLAKE Celebrity stories driving most engagement 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 1000 10500 11000 11500 12000 12500 13000 13500 14000 14500
  • 21.
  • 22.
    CPG Brand The Problem ACPG brand was running a 2nd iteration of their flagship campaign. They wanted new ways to understand how people engaged with the campaign and its underlying theme on Facebook. They also wanted to develop creative insight to support future campaigns. Our Approach ๏ DataSift developed 2 filters focused around the campaign. The first filter focused on the campaign itself: the underlying theme as well as hashtags and urls. The second filter focused on how the campaign tag line was used in normal discussions ๏ DataSift used VEDO to categorize themes and phrases ๏ The 1st index captured 6.9m interactions in 7 days. The 2nd index captured 461k interactions in 14 days.
  • 23.
    The brand wasable to derive audience and creative insights using topic data: ๏ The campaign had targeted a theme which had much less engagement than the themes of pride or determination. ๏ When people engage with these more popular themes, the majority of links they share are for charity runs or causes. ๏ We could track the volume (and demographics) of engagement in phrases the brand had previously featured and identify new phrases which should resonate with their target audience. Recommended Actions ๏ Look at targeting different themes and phrases, which are proven to resonate with target audience ๏ Consider sponsoring charity runs / events as part of future marketing mix CPG Brand 2,481.900 299,000 Campaign theme Determination Pride Top websites linked when sharing emotional state CHARITY RUNS CHARITY RAISING FOR INDIVIDUALS OTHER 11,300
  • 24.
  • 25.
    CPG Brand The Problem ACPG company believed that Millennials are fundamentally different to older consumers in their attitudes to food. They wanted to use Facebook to measure and understand these differences so they could develop new creative campaigns to better engage Millennial consumers. Our Approach ๏ DataSift developed 2 filters focused around the campaign. The first filter focused on general engagement with food as a topic. The 2nd filter focused on the specific food trends the company had already identified ๏ DataSift used VEDO to categorize themes, trends, consumption locations and meals ๏ The 1st index captured 13.2m interactions in 6 days. The 2nd index captured 11m interactions in 32 days.
  • 26.
    The brand wasable to derive multiple actionable creative insights using topic data: ๏ We could see very significant different levels of engagement in most food trends between young millennials (18-24) and older millennials (25-34) but little distinction between millennials versus older consumers ๏ Small differences in the media types millennials post (more photos) but a negligible difference in the content they engage with when discussing food ๏ Identified a range of national difference in engagement with food themes and meals Recommended Actions ๏ Don’t treat Millennials as a homogenous group - in many ways they are like the rest of us! ๏ Optimize campaign messaging to reflect national engagement in different food trends CPG Brand Media type when posting about food Engagement on media type about food American Males engagement in Reduced Meat British Females engagement in Natural Food
  • 27.
  • 28.