INTENT & AI IN B2B ADVERTISING
Matt Aaronson, Director of Product Marketing
Seth Myers, Director of Data Science
MARCH 2019
AGENDA
§ THE IDENTITY TRAP & THE PROMISE OF INTENT
§ THE SCIENCE OF INTENT
§ PUTTING INTENT TO WORK
THE IDENTITY TRAP
& THE PROMISE OF INTENT
SLIDE 4
PRECISE TARGETING IS EVERYTHING IN B2B
Consumer
campaign
B2B
campaign
Qualified accounts and buying committees
Core target demographic / behavioral profile
General population – Less likely to purchase
General population – Can not purchase
SLIDE 5
FUN WITH THIRD-PARTY SEGMENTS
MATT’S FIRMOGRAPHICS:
Χ Company Size > 10-50 Employees
Χ Company Size > 50-100 Employees
Χ Company Size > 5,000+ Employees
Χ Sales Volume > $1MM - $5MM
Χ Small Business
SLIDE 6
FUN WITH THIRD-PARTY B2B SEGMENTS
Don’t blindly trust
job function
audiences
SLIDE 7
HOW IDENTITY-BASED SEGMENTS ARE BUILT
Contact on-boarding
Form data
On-boarding:
70-85%
drop-off
Initial contact list Campaign audience
Form fill coverage Campaign audience
SLIDE 8
HOW IDENTITY-BASED SEGMENTS ARE BUILT
Keyword behaviors:
§ Payroll management
§ Payroll software
§ Compensation management
§ Accounting software
§ Joint pain
§ Arthritis treatment
§ John Bolton
§ State Department
§ G8 Summit
Job functions assigned:
§ Accounting
§ Finance
§ Healthcare practitioner
§ Government > Foreign Policy
Should B2B advertisers be starting
with job titles in the first place?
SLIDE 10
JOB
TITLE
EXPLOSION
SLIDE 11
CONSUMER ADVERTISERS HAVE ADOPTED BEHAVIORAL
SLIDE 13
CONSUMER ADVERTISERS HAVE ADOPTED BEHAVIORAL
Source: Millward Brown Digital
SLIDE 14
WHY INTENT? BECAUSE IDENTITY OBSCURES REAL DEMAND.
Job functions assigned:
§ Accounting
§ Finance
Keyword behaviors:
§ Payroll management
§ Payroll software
§ Compensation management
§ Accounting software
Served ads for:
§ Billing and invoicing technology
§ Forecasting software
SLIDE 15
WHY INTENT? BECAUSE IDENTITY OBSCURES REAL DEMAND.
Keyword behaviors:
§ Payroll management
§ Payroll software
§ Compensation management
§ Accounting software
Served ads for:
§ Payroll software
§ Billing and invoicing technology
§ Forecasting software
INTENT BRINGS THE ANONYMOUS BUYERS’ JOURNEY INTO THE LIGHT
Initial
Research
Social
Network
Social
Media
First
Website
Visit
Competitive
Site Visits
Buying
committee
research
Form fill /
hand-raise
Pipeline
Customer
INTEREST
Evaluation
ANONYMOUS
BUYERS
KNOWN
BUYERS
If increasing numbers of
buyers are starting their
journeys to choose a life
partner anonymously, B2B
marketers have to learn
how to engage with them
on the same terms — for
as long as it takes for B2B
consumers to reveal who
they are and take the next
step with their company.
Steven Casey – “The
Birth of the B2B
Consumer”
October 2017
SLIDE 17
KEYWORD PROFILE INTENT MONITORING
Boehringer Ingelheim employees
Read 12 articles about:
account based marketing
Read 7 articles about:
funnel management
Recently read 2 articles about:
b2b marketing
Read 3 articles about:
b2b marketing
Read 1 articles about:
marketing automation
Read 2 articles about:
target accounts
AI BID OPTIMIZATION
Buyer influence score 94
Optimal bid $14.31 CPM
Buyer influence score 68
Optimal bid $7.89 CPM
Buyer influence score 11
Optimal bid $1.31 CPM
IDENTIFYING BUYING COMMITTEES WITH INTENT
SLIDE 18
SCALE More visibility, higher coverage
RELEVANCE In-market buyers
PRIORITIZATION Understand relative intent levels within accounts
FULL FUNNEL Understand stages of buyers’ journey
THE CASE
FOR INTENT
THE SCIENCE OF INTENT
SLIDE 20
SO WHERE DOES INTENT COME FROM?
www.forbes.com/three-things-to-
know-about-cybersecurity
Cookie 1234
at IP Address
2.1.4.5
www.techrepublic.com/your-
enterprise-cybersecurity-plan/
Cookie 4657
at IP Address
8.7.1.2
www.fortune.com/google-nsa-rsa-
conference-highlights
Cookie 9876
at IP Address
7.7.6.4
500 MillionGlobal cookies
8.5 BillionDaily content consumption events
SO WHERE DOES INTENT COME FROM?
Cookie works at
Cookie works at
Cookie works at
Cookie works at
Cookie works at
Cookie works at
Demandbase
company identification
Content
extraction
SLIDE 22
IDENTIFYING RELEVANT ARTICLES
Which is relevant to digital security?
Article 1 Article 2
Local World Sports Politics Business Health Technology
digital security
security
security
security security
security
security
digital
National Security Advisor to Visit BorderRecent Innovations in Cybersecurity
SLIDE 23
AN ARTIFICIAL INTELLIGENCE THAT UNDERSTANDS CONTEXT
Step 1
Our customer identifies
specific keywords:
Digital Security
Data Encryption
Malware Prevention
Step 2
AI identifies 100s-1,000s
of other related keywords
from a vocab of 105,000+
SLIDE 24
Step 3
AI looks for all related
keywords in the article to
determine relevance
AN ARTIFICIAL INTELLIGENCE THAT UNDERSTANDS CONTEXT
Local World Sports Politics
digital security
Recent Innovations in Cybersecurity
digital security
endpoint security
threat detection
password safety
access management
Relevance
SLIDE 25
FROM RELEVANT ARTICLES TO RELEVANT BUYERS
A single relevant article does not indicate a potential buyer…
§ How many relevant articles did they read?
§ More articles imply stronger interest
§ Over how many different days?
§ A potential buyer would have sustained interest
§ How old were the articles?
§ New articles are likely “news of the day”
§ Older articles imply vendor research
§ How recently did they read the articles?
§ Timing is everything
§ How rare are the relevant keywords in the articles?
§ Rarer/more specific keywords imply cookie is farther into the buyer journey
§ Do they have intent in other topics that disqualifies them?
SLIDE 26
FROM RELEVANT ARTICLES TO RELEVANT COOKIES
Probability of Site
Engagement
AI identifies intent patterns that precede on-site engagement
SLIDE 27
AND IT WORKS!
0.00%
0.05%
0.10%
0.15%
0.20%
0.25%
0.30%
0.35%
0.40%
Highest Intent
Cookies
Medium Intent
Cookies
Low Intent
Cookies
Average Cookie Negative Intent
Cookies
CookieEngagementRate
PUTTING INTENT TO WORK
© 2018 DEMANDBASE SLIDE 29
START WITH YOUR CAMPAIGN AUDIENCE & OBJECTIVE
AWARENESS
PIPELINE
CUSTOMER
§ Top-of-funnel prospecting
§ New market entry
§ Pipeline generation
§ Pipeline acceleration
§ Close rate lift
§ ACV optimization
§ Retention
§ Up-sell
§ Cross-sell
Other ways to segment:
§ Account priority
§ Product / Solution
§ Retargeting
§ Location
§ Competitive conquesting
© 2018 DEMANDBASE SLIDE 30
§ Choose keywords that your buyers use
to research your product or category
§ Use at least 50 keywords in your profile
§ Start with SEM keywords
BEST PRACTICES: INTENT KEYWORDS
© 2018 DEMANDBASE SLIDE 31
INTENT BEST PRACTICES BY FUNNEL STAGE
Broad keywords – Category level
Ex: B2B Marketing, ABM, Demand Generation, Leads, B2B sales
Specific keywords – Product and challenge level
Ex: Demandbase, Targeting Solution, Account list, B2B intent
© 2018 DEMANDBASE SLIDE 32
INTENT BEST PRACTICES BY PRODUCT AREA
General keywords – every profile:
Ex. B2B marketing, ABM, Demand generation
Product-profile-specific keywords
Targeting
§ Programmatic advertising
§ Campaign optimization
§ B2B campaigns
§ Adwords
§ DMP
Engagement
§ Site personalization
§ Form fills
§ Content recommendations
§ Site experience
Conversion
§ Sales intelligence
§ Account alerts
§ Intent alerts
§ Prospecting
§ Sales tools
© 2018 DEMANDBASE SLIDE 33
KEYWORDS FOR COMPETITIVE CONQUESTING
Competitor names
§ 6Sense
§ Terminus
§ Engagio
§ Triblio
§ The Big Willow
Competitor products
§ Account-based
Orchestration
Platform
§ BrightFunnel
§ Playmaker
Competitor artifacts
§ Orchestration
§ Flip my funnel
§ Account plays
§ Predictive
SLIDE 34
DEMANDBASE RESOURCES FOR KEYWORD LIST BUILDING
CONTEXTUAL RELEVANCY
KEYWORD
DISCOVERY
COMING SOON:
AUTO-DETECT KEYWORDS:
KEY TAKE AWAYS
§ Intent finds the “who?” but also the
“who, specifically?” and the “when?”
§ Scale and sophistication pay off
§ Intent can serve a broad range of
campaign use cases
SLIDE 36
COMING SOON…
§ Cross-Campaign
Analytics
§ Self-service campaign
management
§ Dynamic audiences with
intent
THANK YOU

Intent and AI in B2B Advertising

  • 1.
    INTENT & AIIN B2B ADVERTISING Matt Aaronson, Director of Product Marketing Seth Myers, Director of Data Science MARCH 2019
  • 2.
    AGENDA § THE IDENTITYTRAP & THE PROMISE OF INTENT § THE SCIENCE OF INTENT § PUTTING INTENT TO WORK
  • 3.
    THE IDENTITY TRAP &THE PROMISE OF INTENT
  • 4.
    SLIDE 4 PRECISE TARGETINGIS EVERYTHING IN B2B Consumer campaign B2B campaign Qualified accounts and buying committees Core target demographic / behavioral profile General population – Less likely to purchase General population – Can not purchase
  • 5.
    SLIDE 5 FUN WITHTHIRD-PARTY SEGMENTS MATT’S FIRMOGRAPHICS: Χ Company Size > 10-50 Employees Χ Company Size > 50-100 Employees Χ Company Size > 5,000+ Employees Χ Sales Volume > $1MM - $5MM Χ Small Business
  • 6.
    SLIDE 6 FUN WITHTHIRD-PARTY B2B SEGMENTS Don’t blindly trust job function audiences
  • 7.
    SLIDE 7 HOW IDENTITY-BASEDSEGMENTS ARE BUILT Contact on-boarding Form data On-boarding: 70-85% drop-off Initial contact list Campaign audience Form fill coverage Campaign audience
  • 8.
    SLIDE 8 HOW IDENTITY-BASEDSEGMENTS ARE BUILT Keyword behaviors: § Payroll management § Payroll software § Compensation management § Accounting software § Joint pain § Arthritis treatment § John Bolton § State Department § G8 Summit Job functions assigned: § Accounting § Finance § Healthcare practitioner § Government > Foreign Policy
  • 9.
    Should B2B advertisersbe starting with job titles in the first place?
  • 10.
  • 11.
  • 12.
    CONSUMER ADVERTISERS HAVEADOPTED BEHAVIORAL
  • 13.
    SLIDE 13 CONSUMER ADVERTISERSHAVE ADOPTED BEHAVIORAL Source: Millward Brown Digital
  • 14.
    SLIDE 14 WHY INTENT?BECAUSE IDENTITY OBSCURES REAL DEMAND. Job functions assigned: § Accounting § Finance Keyword behaviors: § Payroll management § Payroll software § Compensation management § Accounting software Served ads for: § Billing and invoicing technology § Forecasting software
  • 15.
    SLIDE 15 WHY INTENT?BECAUSE IDENTITY OBSCURES REAL DEMAND. Keyword behaviors: § Payroll management § Payroll software § Compensation management § Accounting software Served ads for: § Payroll software § Billing and invoicing technology § Forecasting software
  • 16.
    INTENT BRINGS THEANONYMOUS BUYERS’ JOURNEY INTO THE LIGHT Initial Research Social Network Social Media First Website Visit Competitive Site Visits Buying committee research Form fill / hand-raise Pipeline Customer INTEREST Evaluation ANONYMOUS BUYERS KNOWN BUYERS If increasing numbers of buyers are starting their journeys to choose a life partner anonymously, B2B marketers have to learn how to engage with them on the same terms — for as long as it takes for B2B consumers to reveal who they are and take the next step with their company. Steven Casey – “The Birth of the B2B Consumer” October 2017
  • 17.
    SLIDE 17 KEYWORD PROFILEINTENT MONITORING Boehringer Ingelheim employees Read 12 articles about: account based marketing Read 7 articles about: funnel management Recently read 2 articles about: b2b marketing Read 3 articles about: b2b marketing Read 1 articles about: marketing automation Read 2 articles about: target accounts AI BID OPTIMIZATION Buyer influence score 94 Optimal bid $14.31 CPM Buyer influence score 68 Optimal bid $7.89 CPM Buyer influence score 11 Optimal bid $1.31 CPM IDENTIFYING BUYING COMMITTEES WITH INTENT
  • 18.
    SLIDE 18 SCALE Morevisibility, higher coverage RELEVANCE In-market buyers PRIORITIZATION Understand relative intent levels within accounts FULL FUNNEL Understand stages of buyers’ journey THE CASE FOR INTENT
  • 19.
  • 20.
    SLIDE 20 SO WHEREDOES INTENT COME FROM? www.forbes.com/three-things-to- know-about-cybersecurity Cookie 1234 at IP Address 2.1.4.5 www.techrepublic.com/your- enterprise-cybersecurity-plan/ Cookie 4657 at IP Address 8.7.1.2 www.fortune.com/google-nsa-rsa- conference-highlights Cookie 9876 at IP Address 7.7.6.4 500 MillionGlobal cookies 8.5 BillionDaily content consumption events
  • 21.
    SO WHERE DOESINTENT COME FROM? Cookie works at Cookie works at Cookie works at Cookie works at Cookie works at Cookie works at Demandbase company identification Content extraction
  • 22.
    SLIDE 22 IDENTIFYING RELEVANTARTICLES Which is relevant to digital security? Article 1 Article 2 Local World Sports Politics Business Health Technology digital security security security security security security security digital National Security Advisor to Visit BorderRecent Innovations in Cybersecurity
  • 23.
    SLIDE 23 AN ARTIFICIALINTELLIGENCE THAT UNDERSTANDS CONTEXT Step 1 Our customer identifies specific keywords: Digital Security Data Encryption Malware Prevention Step 2 AI identifies 100s-1,000s of other related keywords from a vocab of 105,000+
  • 24.
    SLIDE 24 Step 3 AIlooks for all related keywords in the article to determine relevance AN ARTIFICIAL INTELLIGENCE THAT UNDERSTANDS CONTEXT Local World Sports Politics digital security Recent Innovations in Cybersecurity digital security endpoint security threat detection password safety access management Relevance
  • 25.
    SLIDE 25 FROM RELEVANTARTICLES TO RELEVANT BUYERS A single relevant article does not indicate a potential buyer… § How many relevant articles did they read? § More articles imply stronger interest § Over how many different days? § A potential buyer would have sustained interest § How old were the articles? § New articles are likely “news of the day” § Older articles imply vendor research § How recently did they read the articles? § Timing is everything § How rare are the relevant keywords in the articles? § Rarer/more specific keywords imply cookie is farther into the buyer journey § Do they have intent in other topics that disqualifies them?
  • 26.
    SLIDE 26 FROM RELEVANTARTICLES TO RELEVANT COOKIES Probability of Site Engagement AI identifies intent patterns that precede on-site engagement
  • 27.
    SLIDE 27 AND ITWORKS! 0.00% 0.05% 0.10% 0.15% 0.20% 0.25% 0.30% 0.35% 0.40% Highest Intent Cookies Medium Intent Cookies Low Intent Cookies Average Cookie Negative Intent Cookies CookieEngagementRate
  • 28.
  • 29.
    © 2018 DEMANDBASESLIDE 29 START WITH YOUR CAMPAIGN AUDIENCE & OBJECTIVE AWARENESS PIPELINE CUSTOMER § Top-of-funnel prospecting § New market entry § Pipeline generation § Pipeline acceleration § Close rate lift § ACV optimization § Retention § Up-sell § Cross-sell Other ways to segment: § Account priority § Product / Solution § Retargeting § Location § Competitive conquesting
  • 30.
    © 2018 DEMANDBASESLIDE 30 § Choose keywords that your buyers use to research your product or category § Use at least 50 keywords in your profile § Start with SEM keywords BEST PRACTICES: INTENT KEYWORDS
  • 31.
    © 2018 DEMANDBASESLIDE 31 INTENT BEST PRACTICES BY FUNNEL STAGE Broad keywords – Category level Ex: B2B Marketing, ABM, Demand Generation, Leads, B2B sales Specific keywords – Product and challenge level Ex: Demandbase, Targeting Solution, Account list, B2B intent
  • 32.
    © 2018 DEMANDBASESLIDE 32 INTENT BEST PRACTICES BY PRODUCT AREA General keywords – every profile: Ex. B2B marketing, ABM, Demand generation Product-profile-specific keywords Targeting § Programmatic advertising § Campaign optimization § B2B campaigns § Adwords § DMP Engagement § Site personalization § Form fills § Content recommendations § Site experience Conversion § Sales intelligence § Account alerts § Intent alerts § Prospecting § Sales tools
  • 33.
    © 2018 DEMANDBASESLIDE 33 KEYWORDS FOR COMPETITIVE CONQUESTING Competitor names § 6Sense § Terminus § Engagio § Triblio § The Big Willow Competitor products § Account-based Orchestration Platform § BrightFunnel § Playmaker Competitor artifacts § Orchestration § Flip my funnel § Account plays § Predictive
  • 34.
    SLIDE 34 DEMANDBASE RESOURCESFOR KEYWORD LIST BUILDING CONTEXTUAL RELEVANCY KEYWORD DISCOVERY COMING SOON: AUTO-DETECT KEYWORDS:
  • 35.
    KEY TAKE AWAYS §Intent finds the “who?” but also the “who, specifically?” and the “when?” § Scale and sophistication pay off § Intent can serve a broad range of campaign use cases
  • 36.
    SLIDE 36 COMING SOON… §Cross-Campaign Analytics § Self-service campaign management § Dynamic audiences with intent
  • 37.