SEATTLE INTERACTIVE CONFERENCE 
Paid Social Advertising: 
Now, Next and What Works Best
Why are we talking about search in a discussion about paid social? 
The early days of search 
• Marketers’ dream comes true 
• We built it, they came 
• Few tools, big results 
• Highly targeted, reactive – 
Customers flocked to us 
• At low cost 
Search was brilliant.
Why are we talking about search in a discussion about paid social? 
The middle years 
• Battle for top rankings intensifies 
• Fierce competition for keywords 
• The age of SEO begins 
• Fuzzy metrics 
• ROI uncertain 
• Enter paid search
Why are we talking about search in a discussion about paid social? 
Today 
• SEOs have tried all sorts of things 
• Diminished search experience 
• Algorithms (Vince, Panda) 
Most brands have to pay for reach and scale 
• Brands can’t control search 
results 
• Paid Search is a multi-billion dollar 
a year industry.
Current Landscape 
Social Marketing Today 
• Social reach in decline 
• Battle for the newsfeed 
• Algorithms control eyeballs 
Brands have lost control of who 
sees their social content. Sound Familiar? 
• Confusing metrics 
• If brands build content, consumption is 
not guaranteed. This is a problem. 
• Enter Paid Social
Current Landscape 
Search Marketing Today/ Social 
Reach in decline Climate Visibility in decline 
Front Page Battle Zone Newsfeed 
Algorithms Control Algorithms 
Keywords Targeting Gender, age, geo 
Fuzzy Metrics Confusing
Current Landscape 
Social Marketing Today 
• Followers, retweets 
• Fans, likes, comments, shares 
• Likes, comments, shares 
This model leaves lots of unanswered questions
Current Landscape 
Who is on my social channels? 
Why is it important to know? 
How you can find out? 
What to do with the information?
Current Landscape 
How can I reach them? 
• Location 
• Age, Gender 
• Interests 
• Behavior 
• Education Levels 
• Connections 
• Keywords 
• Interests 
• Location 
• Gender 
• Language 
• Devices 
• By job title and 
function 
• By industry and 
company size 
• By seniority 
FACEBOOK TWITTER LINKEDIN
Current Landscape 
What content is working? 
Why is it important to know? 
How you can find out? 
What to do with the information?
Current Landscape 
Who is sharing it? 
Why is it important to know? 
How you can find out? 
What to do with the information?
Current Landscape 
Limitations 
The customer journey is not limited to just social channels 
Analytics should give deep insight across channels 
Multichannel/Full Lifecycle approach to managing the customer experience
The future of social advertising 
Programmatic 
• What does programmatic mean to you? 
• Then & Now
The future of social advertising 
Audience Profiles 
• The best audience profile is data-driven 
• Utilize a combination of native targeting on 
social networks in addition to your 
first-party data 
• First-party data 
• Social network native targeting
The future of social advertising 
Audience Profiles continued… 
3 KEY WAYS OF TARGETING AUDIENCES ON FACEBOOK AND TWITTER: 
1. Core Audiences 
2. Custom or Tailored Audiences 
3. Look-a-like Audiences
The future of social advertising 
Social Retargeting 
Why retarget? 
How do you retarget? 
What are the use causes?
The future of social advertising 
Forecasted campaigns and predictive analytics 
• A single source of truth bringing data together 
• Adopt consistent ad tech and marketing tech 
solution 
• No single user journey is the same 
• Leverage other unique data points such as rich 
analytics data as early signals that correlate to a 
conversion rate to address data sparsity issues.
The future of social advertising 
Predictive analytics to inform ad buying 
Some company have access to hundreds of 
analytics metrics 
Identify ones that have a high correlation with 
conversion rate and leverage as revenue signals. 
This may differ from industry vertical to vertical. 
Common revenue signals are: 
1. Page views in the first visit 
2. Total time spent on multiple visits 
3. Bounce rate 
4. Total page views across multiple visits 
5. Time spent onsite in first visit
The future of social advertising 
Best practices for targeting in a 
multi-channel ecosystem 
• Common business objectives 
• Leverage customer profiles & 
audience targeting 
• Segment audiences at scale and test 
different levers that influence performance 
• Leverage early predictors of performance 
using web analytics 
• Optimize channels holistically to maximize 
digital ROI 
• Experiment and iterate
Case Studies 
Lead Generation for Education client 
OBJECTIVE: TEST AND MAXIMIZE FULL POTENTIAL OF 1ST PARTY CRM DATA 
Solution: Test overlay of core native targeting 
with Look-a-like targeting 
Results: Benchmarked against Core native 
targeting on Facebook 
Click through rate increased +447% 
Conversion rate increased +51% 
Cost per lead improved 46% 
Solution: Look-a-like audiences generated from 
Custom Audiences 
Results: Benchmarked against Core native 
targeting on Facebook 
Conversion rate increased +199% 
Cost per lead improved by 59%
-66% 
-45% 
-5% 
3% 
4% 
166% 
100% 
-4% 
1% 
-14% 
-100% -50% 0% 50% 100% 150% 200% 
ROAS CPO 
Case Studies 
Online sales for eCommerce client 
LOOKALIKE AUDIENCES 
LIKES & INTERESTS 
LOOKALIKE AUDIENCES 
+ LIKES & INTERESTS 
CUSTOM AUDIENCES 
WEBSITE CUSTOM AUDIENCES 
Custom audiences & other Audience Targets relative to avg. CPO and ROAS
0% 
2% 
4% 
6% 
Search Behavior Lifetime Website Custom Audiences Last 60 
Days 
Search Behavior Last 7 Days 
CTR CVR 
Case Studies 
Online sales for Travel client 
ROAS: 5.69 6.64 8.11 
OBJECTIVE: DRIVE INCREMENTAL SCALE BEYOND SEARCH, COST-EFFICIENTLY 
Solution: Social retargeting through Website and CRM Custom Audiences 
Results:
Case Studies 
Online sales for Travel client 
0.41% 0.61% 0.50% 0.56% 
1.02% 
3.06% 
8.25% 
9.70% 
0% 
2% 
4% 
6% 
8% 
10% 
12% 
Engaged Newsletter 
Subscribers 
High Value Customers Lower Tier Loyalty Program Higher Tier Loyalty Program 
CTR CVR 
ROAS: 5.69 12.12 15.55 24.85
Thank You

Paid Social Advertising: Now, Next and What Works Best

  • 1.
    SEATTLE INTERACTIVE CONFERENCE Paid Social Advertising: Now, Next and What Works Best
  • 4.
    Why are wetalking about search in a discussion about paid social? The early days of search • Marketers’ dream comes true • We built it, they came • Few tools, big results • Highly targeted, reactive – Customers flocked to us • At low cost Search was brilliant.
  • 6.
    Why are wetalking about search in a discussion about paid social? The middle years • Battle for top rankings intensifies • Fierce competition for keywords • The age of SEO begins • Fuzzy metrics • ROI uncertain • Enter paid search
  • 8.
    Why are wetalking about search in a discussion about paid social? Today • SEOs have tried all sorts of things • Diminished search experience • Algorithms (Vince, Panda) Most brands have to pay for reach and scale • Brands can’t control search results • Paid Search is a multi-billion dollar a year industry.
  • 10.
    Current Landscape SocialMarketing Today • Social reach in decline • Battle for the newsfeed • Algorithms control eyeballs Brands have lost control of who sees their social content. Sound Familiar? • Confusing metrics • If brands build content, consumption is not guaranteed. This is a problem. • Enter Paid Social
  • 11.
    Current Landscape SearchMarketing Today/ Social Reach in decline Climate Visibility in decline Front Page Battle Zone Newsfeed Algorithms Control Algorithms Keywords Targeting Gender, age, geo Fuzzy Metrics Confusing
  • 12.
    Current Landscape SocialMarketing Today • Followers, retweets • Fans, likes, comments, shares • Likes, comments, shares This model leaves lots of unanswered questions
  • 13.
    Current Landscape Whois on my social channels? Why is it important to know? How you can find out? What to do with the information?
  • 14.
    Current Landscape Howcan I reach them? • Location • Age, Gender • Interests • Behavior • Education Levels • Connections • Keywords • Interests • Location • Gender • Language • Devices • By job title and function • By industry and company size • By seniority FACEBOOK TWITTER LINKEDIN
  • 15.
    Current Landscape Whatcontent is working? Why is it important to know? How you can find out? What to do with the information?
  • 16.
    Current Landscape Whois sharing it? Why is it important to know? How you can find out? What to do with the information?
  • 17.
    Current Landscape Limitations The customer journey is not limited to just social channels Analytics should give deep insight across channels Multichannel/Full Lifecycle approach to managing the customer experience
  • 19.
    The future ofsocial advertising Programmatic • What does programmatic mean to you? • Then & Now
  • 20.
    The future ofsocial advertising Audience Profiles • The best audience profile is data-driven • Utilize a combination of native targeting on social networks in addition to your first-party data • First-party data • Social network native targeting
  • 21.
    The future ofsocial advertising Audience Profiles continued… 3 KEY WAYS OF TARGETING AUDIENCES ON FACEBOOK AND TWITTER: 1. Core Audiences 2. Custom or Tailored Audiences 3. Look-a-like Audiences
  • 22.
    The future ofsocial advertising Social Retargeting Why retarget? How do you retarget? What are the use causes?
  • 23.
    The future ofsocial advertising Forecasted campaigns and predictive analytics • A single source of truth bringing data together • Adopt consistent ad tech and marketing tech solution • No single user journey is the same • Leverage other unique data points such as rich analytics data as early signals that correlate to a conversion rate to address data sparsity issues.
  • 24.
    The future ofsocial advertising Predictive analytics to inform ad buying Some company have access to hundreds of analytics metrics Identify ones that have a high correlation with conversion rate and leverage as revenue signals. This may differ from industry vertical to vertical. Common revenue signals are: 1. Page views in the first visit 2. Total time spent on multiple visits 3. Bounce rate 4. Total page views across multiple visits 5. Time spent onsite in first visit
  • 25.
    The future ofsocial advertising Best practices for targeting in a multi-channel ecosystem • Common business objectives • Leverage customer profiles & audience targeting • Segment audiences at scale and test different levers that influence performance • Leverage early predictors of performance using web analytics • Optimize channels holistically to maximize digital ROI • Experiment and iterate
  • 27.
    Case Studies LeadGeneration for Education client OBJECTIVE: TEST AND MAXIMIZE FULL POTENTIAL OF 1ST PARTY CRM DATA Solution: Test overlay of core native targeting with Look-a-like targeting Results: Benchmarked against Core native targeting on Facebook Click through rate increased +447% Conversion rate increased +51% Cost per lead improved 46% Solution: Look-a-like audiences generated from Custom Audiences Results: Benchmarked against Core native targeting on Facebook Conversion rate increased +199% Cost per lead improved by 59%
  • 28.
    -66% -45% -5% 3% 4% 166% 100% -4% 1% -14% -100% -50% 0% 50% 100% 150% 200% ROAS CPO Case Studies Online sales for eCommerce client LOOKALIKE AUDIENCES LIKES & INTERESTS LOOKALIKE AUDIENCES + LIKES & INTERESTS CUSTOM AUDIENCES WEBSITE CUSTOM AUDIENCES Custom audiences & other Audience Targets relative to avg. CPO and ROAS
  • 29.
    0% 2% 4% 6% Search Behavior Lifetime Website Custom Audiences Last 60 Days Search Behavior Last 7 Days CTR CVR Case Studies Online sales for Travel client ROAS: 5.69 6.64 8.11 OBJECTIVE: DRIVE INCREMENTAL SCALE BEYOND SEARCH, COST-EFFICIENTLY Solution: Social retargeting through Website and CRM Custom Audiences Results:
  • 30.
    Case Studies Onlinesales for Travel client 0.41% 0.61% 0.50% 0.56% 1.02% 3.06% 8.25% 9.70% 0% 2% 4% 6% 8% 10% 12% Engaged Newsletter Subscribers High Value Customers Lower Tier Loyalty Program Higher Tier Loyalty Program CTR CVR ROAS: 5.69 12.12 15.55 24.85
  • 31.