SlideShare a Scribd company logo
© 2016 eMarketer Inc.
Made possible by
Navigating the Complexities of
Cross-Platform Attribution
Lauren T. Fisher
Senior Analyst
October 6, 2016
© 2016 eMarketer Inc.
Today’s Agenda
 Assess the current state of attribution
 Discuss common roadblocks to implementation
 Identify future areas of development and advancement
© 2016 eMarketer Inc.
What is attribution?
 The practice of
assigning credit to
any advertising- or
marketing-driven
interaction or other
brand-imposed
touchpoint
© 2016 eMarketer Inc.
The explosion of devices and touchpoints has
made cross-platform attribution a must
© 2016 eMarketer Inc.
This
complexity,
among other
challenges,
has made
attribution a
significant
area of
concern for
many
marketers
© 2016 eMarketer Inc.
Those using
attribution are
still in the
minority
© 2016 eMarketer Inc.
About half of US companies will shift to
multichannel attribution models by 2017
 But remember:
This is only the
number of
companies using
it for digital
media
© 2016 eMarketer Inc.
Cross-screen complexities aren’t the only things
sidelining marketers
 Cost, expertise and
organizational silos
also delay
attribution
adoption and usage
© 2016 eMarketer Inc.
Roadblocks to Overcome:
Setting Up for Attribution
Success
© 2016 eMarketer Inc.
Interest in
more
advanced
attribution
methods is
high
© 2016 eMarketer Inc.
But getting sign-off on these tools and platforms
can be difficult
© 2016 eMarketer Inc.
Winning budgets and justifying long-term value
can prove tricky, but essential
“When it comes to implementing a DMP or an
attribution platform, these are new cost items
that haven’t necessarily been slotted into the budget
mix. How do you fund that? How do you make a case
to management that this six-digit investment
that probably won’t be up and running at any sort of
meaningful scale for six to 13 months is going to
pay for itself on the other end of this?”
—Rob Griffin, Chief Innovation Officer, Almighty
© 2016 eMarketer Inc.
For most, arriving at a ‘single point of truth’
proves a powerful case for attribution
© 2016 eMarketer Inc.
But to get
there requires
companies to
make multiple
tools and tech
work together
© 2016 eMarketer Inc.
Achieving
full-scale
system
integration
is still an
industrywide
pursuit
© 2016 eMarketer Inc.
Many lack
proper
coordination
and expertise
required for
such
integration
© 2016 eMarketer Inc.
But more often, organizational and departmental
silos impede proper data share and integrations
© 2016 eMarketer Inc.
Data
integration
and
accessibility
are key to
making
attribution
work, but so
is data quality
© 2016 eMarketer Inc.
When it comes to attribution, it’s garbage in,
garbage out
© 2016 eMarketer Inc.
Common data quality issues marketers must vet
for include:
 Fidelity of owned and tracked data
 Fraud and viewability
 Accuracy of purchased data
 Compatibility of device-level data; online-to-offline
(O2O) data
© 2016 eMarketer Inc.
Cross-device and the lack of a universal
identifier poses even bigger challenges
“Cross-device targeting came ‘to the masses’ in
2015. That resulted in a lot of conversations about
deterministic and probabilistic data. But what we still
see is that cross-device attribution is still very much
in its infancy and there are no standards around
it yet. Between what ID you have on one platform vs.
what kind of ID you have on another, it ends up
becoming very complicated … and the resulting
attribution can look entirely different.”
—Alan Beiagi, Senior Director, Products, DataXu
© 2016 eMarketer Inc.
To combat this and other data quality issues,
many utilize first-party data for known users
© 2016 eMarketer Inc.
The Attribution To-Do List:
 Audit current digital
measurement strategy
 Identify where sales data
is being captured
 Take stock of tech
 Decide: Is this enough?
 If not, choose a vendor
© 2016 eMarketer Inc.
Looking Ahead: The Future
of Attribution
© 2016 eMarketer Inc.
A more cohesive blending of top-down
and bottom-up models
1
© 2016 eMarketer Inc.
In the past, most marketers have used one of
two attribution methods:
Top Down
 Branding, traditional media
(print, TV, etc.)
 Econometrics, marketing-
mix modeling
 Uses GRPs/impressions,
spend, sales data, external
data to model attribution
 Budget allocation and media
planning
Bottom Up
 Native to digital
 Path analysis
 Uses cookies and other
web data to analyze
actual activity
 Optimization,
performance
© 2016 eMarketer Inc.
But of late, many seek greater convergence
between the two
Image credit: Shutterstock
© 2016 eMarketer Inc.
Still, such solutions are works in progress
“We see a lot of people doing this today, but they’re kind
of duct-taping it together and trying to fit a
square peg in a round hole. And so marketers are left
scratching their heads and wanting to try and figure
out how to improve the data discrepancies and how
to make user-level and aggregate-level data
work together.”
—Amy Mitchell, General Manager, Convertro, an AOL company
© 2016 eMarketer Inc.
A move to more real-time reporting
2
© 2016 eMarketer Inc.
As capabilities improve, rear-view reporting will
become a thing of the past
© 2016 eMarketer Inc.
The ultimate aim of attribution is to move from
this:
To this:Company
Level
Is my
advertising/marketing
successful?
Department Level
Was this campaign successful?
Team Level
Was this tactic successful?
Company Level
Is my advertising/marketing
successful?
Department Level
Was this campaign
successful?
Team Level
Was this tactic successful?
© 2016 eMarketer Inc.
Addressable media is the most malleable at the
moment
 But long-term,
traditional
media will
become more
and more
addressable
© 2016 eMarketer Inc.
But remember: Automation and algorithms are
marketing assets, not marketing absolutes
“Oftentimes marketers immediately want
to jump to optimizing. Shut that off. Turn that
on. Move money from here to here. And they
may be reacting to an anomaly and not a
true pattern. So there’s definitely a balance in
terms of timely data availability and a deep
enough and big enough sample to make sound
decisions.”
—Jon Schulz, CMO, Viant
© 2016 eMarketer Inc.
Greater online-to-offline sales
attribution
3
© 2016 eMarketer Inc.
Attributing all
the way down to
the sale has
become ‘mission
critical’
© 2016 eMarketer Inc.
Tying online to offline (O2O) may be an
imperative, but it’s still a work in progress
“Even with just online, being able to connect the
dots across different devices and publishers is
already hard enough. And to be able to tie
online transactions to offline sales is
even harder, because a lot of the time
advertisers just don’t have the right
technologies and tools in order to do so.”
—Anh Bui, Product Marketing Manager, Facebook
© 2016 eMarketer Inc.
Data matching capabilities are improving, but
there are still pitfalls
© 2016 eMarketer Inc.
Location data
can prove a
rich source of
O2O info, but
its fidelity is
increasingly
being called
into question
© 2016 eMarketer Inc.
Example in action: an entertainment client
looking to activate monthly paid subscriptions
 The goal: Develop a cross-channel, integrated model that drives
business results using paid, owned and earned media.
 Data in: Primary acquisition KPIs (subscriptions, registrations, monthly
active users), paid media (TV, out-of-home, social, online video, display,
search), owned and earned media (site traffic, organic social and
search, email, PR), competitor media spend, holidays and events,
macroeconomic data, company-/industry-specific business metrics.
 Model used: Marketing mix modeling (MMM) with ‘nested’ levels for
optimization (digital).
 Benefits: Determine optimal channel spend; understand how to
optimize and execute marketing dollars to maximize subscriptions and
lifetime value.
© 2016 eMarketer Inc.
Key Takeaways:
 Marketers can no longer afford to ignore cross-platform
attribution.
 Organizational, technological and data quality issues all
pose significant challenges to making attribution work.
 Blended, more real-time models are now a must.
 Tying O2O data will be critical to truly gauge company
success.
Ads that workThe retargeting and prospecting platform of choice
for more than 25,000 advertisers worldwide.
Do more with first-party data
Maximize the value of your customer data with AdRoll.
Website and app data
Use the AdRoll pixel to leverage your customer
data and make it work for you
Partner data
Easily integrate the data from your CRM lists
or favorite marketing apps to power more
effective campaigns
Opt-in data
Take advantage of the AdRoll IntentMap™,
the world’s largest opt-in co-op data set, to
gain access to valuable new audiences
A solution for every stage of the customer journey
Every marketer has three core objectives.
ATTRACT
Reach new prospects
CONVERT
Turn prospects into customers
GROW
Maximize customer lifetime value
PROSPECTINGRETARGETING
Prospecting: Reach new, high-quality audiences
Attract the prospects that are most likely to be in the market for your products.
Your site’s customer behavior is
analyzed and matched with
similar prospects in the IntentMap
Prospects come to your website
via ads. With every new
customer, Prospecting gets
smarter.
“AdRoll Prospecting has been a game changer for our marketing team, helping us find new customers at a rate that eclipses
those of other platforms. We’ve seen excellent backend quality metrics from these conversions.”
—Rob DiNuzzo, demand generation manager at Datadog
Your
website
Power your campaigns with advanced
machine learning
BidIQ™, our proprietary algorithm, learns from every bid it makes to buy high impact impressions
and meet your performance targets.
SCALE
1.2 billion user profiles
and growing
DIVERSITY
An increasingly stronger
data set that constantly
gathers data across
verticals, from SMBs to
Fortune 500 companies
QUALITY
Over 100m variables instantly
evaluated per impression (pages
visited, time of day, device type, etc.)
© 2016 eMarketer Inc.
Learn more about digital marketing with an
eMarketer corporate subscription
Around 200 eMarketer reports are published
each year. Here are some recent reports you
may be interested in:
Q&A Session
Made possible by
You will receive an email tomorrow
with a link to view the deck and
webinar recording.
To learn more: www.emarketer.com/products
800-405-0844 or webinars@emarketer.com
Lauren T. Fisher
Navigating the Complexities of
Cross-Platform Attribution
 Cracking Cross-Device Attribution in 2016: Data Quality,
Blended Models and Merging Online-Offline Data
 Search and Social Platforms: How Facebook and Others Can
Steal Dollars from Search
 Making Multichannel Marketing Work: Four Tactics Required
for Omnichannel Success
 Ad Fraud in the US: How More Sophisticated Methods Are
Hurting Mobile, Video and Performance Measurement

More Related Content

eMarketer Webinar: Navigating the Complexities of Cross-Platform Attribution

  • 1. © 2016 eMarketer Inc. Made possible by Navigating the Complexities of Cross-Platform Attribution Lauren T. Fisher Senior Analyst October 6, 2016
  • 2. © 2016 eMarketer Inc. Today’s Agenda  Assess the current state of attribution  Discuss common roadblocks to implementation  Identify future areas of development and advancement
  • 3. © 2016 eMarketer Inc. What is attribution?  The practice of assigning credit to any advertising- or marketing-driven interaction or other brand-imposed touchpoint
  • 4. © 2016 eMarketer Inc. The explosion of devices and touchpoints has made cross-platform attribution a must
  • 5. © 2016 eMarketer Inc. This complexity, among other challenges, has made attribution a significant area of concern for many marketers
  • 6. © 2016 eMarketer Inc. Those using attribution are still in the minority
  • 7. © 2016 eMarketer Inc. About half of US companies will shift to multichannel attribution models by 2017  But remember: This is only the number of companies using it for digital media
  • 8. © 2016 eMarketer Inc. Cross-screen complexities aren’t the only things sidelining marketers  Cost, expertise and organizational silos also delay attribution adoption and usage
  • 9. © 2016 eMarketer Inc. Roadblocks to Overcome: Setting Up for Attribution Success
  • 10. © 2016 eMarketer Inc. Interest in more advanced attribution methods is high
  • 11. © 2016 eMarketer Inc. But getting sign-off on these tools and platforms can be difficult
  • 12. © 2016 eMarketer Inc. Winning budgets and justifying long-term value can prove tricky, but essential “When it comes to implementing a DMP or an attribution platform, these are new cost items that haven’t necessarily been slotted into the budget mix. How do you fund that? How do you make a case to management that this six-digit investment that probably won’t be up and running at any sort of meaningful scale for six to 13 months is going to pay for itself on the other end of this?” —Rob Griffin, Chief Innovation Officer, Almighty
  • 13. © 2016 eMarketer Inc. For most, arriving at a ‘single point of truth’ proves a powerful case for attribution
  • 14. © 2016 eMarketer Inc. But to get there requires companies to make multiple tools and tech work together
  • 15. © 2016 eMarketer Inc. Achieving full-scale system integration is still an industrywide pursuit
  • 16. © 2016 eMarketer Inc. Many lack proper coordination and expertise required for such integration
  • 17. © 2016 eMarketer Inc. But more often, organizational and departmental silos impede proper data share and integrations
  • 18. © 2016 eMarketer Inc. Data integration and accessibility are key to making attribution work, but so is data quality
  • 19. © 2016 eMarketer Inc. When it comes to attribution, it’s garbage in, garbage out
  • 20. © 2016 eMarketer Inc. Common data quality issues marketers must vet for include:  Fidelity of owned and tracked data  Fraud and viewability  Accuracy of purchased data  Compatibility of device-level data; online-to-offline (O2O) data
  • 21. © 2016 eMarketer Inc. Cross-device and the lack of a universal identifier poses even bigger challenges “Cross-device targeting came ‘to the masses’ in 2015. That resulted in a lot of conversations about deterministic and probabilistic data. But what we still see is that cross-device attribution is still very much in its infancy and there are no standards around it yet. Between what ID you have on one platform vs. what kind of ID you have on another, it ends up becoming very complicated … and the resulting attribution can look entirely different.” —Alan Beiagi, Senior Director, Products, DataXu
  • 22. © 2016 eMarketer Inc. To combat this and other data quality issues, many utilize first-party data for known users
  • 23. © 2016 eMarketer Inc. The Attribution To-Do List:  Audit current digital measurement strategy  Identify where sales data is being captured  Take stock of tech  Decide: Is this enough?  If not, choose a vendor
  • 24. © 2016 eMarketer Inc. Looking Ahead: The Future of Attribution
  • 25. © 2016 eMarketer Inc. A more cohesive blending of top-down and bottom-up models 1
  • 26. © 2016 eMarketer Inc. In the past, most marketers have used one of two attribution methods: Top Down  Branding, traditional media (print, TV, etc.)  Econometrics, marketing- mix modeling  Uses GRPs/impressions, spend, sales data, external data to model attribution  Budget allocation and media planning Bottom Up  Native to digital  Path analysis  Uses cookies and other web data to analyze actual activity  Optimization, performance
  • 27. © 2016 eMarketer Inc. But of late, many seek greater convergence between the two Image credit: Shutterstock
  • 28. © 2016 eMarketer Inc. Still, such solutions are works in progress “We see a lot of people doing this today, but they’re kind of duct-taping it together and trying to fit a square peg in a round hole. And so marketers are left scratching their heads and wanting to try and figure out how to improve the data discrepancies and how to make user-level and aggregate-level data work together.” —Amy Mitchell, General Manager, Convertro, an AOL company
  • 29. © 2016 eMarketer Inc. A move to more real-time reporting 2
  • 30. © 2016 eMarketer Inc. As capabilities improve, rear-view reporting will become a thing of the past
  • 31. © 2016 eMarketer Inc. The ultimate aim of attribution is to move from this: To this:Company Level Is my advertising/marketing successful? Department Level Was this campaign successful? Team Level Was this tactic successful? Company Level Is my advertising/marketing successful? Department Level Was this campaign successful? Team Level Was this tactic successful?
  • 32. © 2016 eMarketer Inc. Addressable media is the most malleable at the moment  But long-term, traditional media will become more and more addressable
  • 33. © 2016 eMarketer Inc. But remember: Automation and algorithms are marketing assets, not marketing absolutes “Oftentimes marketers immediately want to jump to optimizing. Shut that off. Turn that on. Move money from here to here. And they may be reacting to an anomaly and not a true pattern. So there’s definitely a balance in terms of timely data availability and a deep enough and big enough sample to make sound decisions.” —Jon Schulz, CMO, Viant
  • 34. © 2016 eMarketer Inc. Greater online-to-offline sales attribution 3
  • 35. © 2016 eMarketer Inc. Attributing all the way down to the sale has become ‘mission critical’
  • 36. © 2016 eMarketer Inc. Tying online to offline (O2O) may be an imperative, but it’s still a work in progress “Even with just online, being able to connect the dots across different devices and publishers is already hard enough. And to be able to tie online transactions to offline sales is even harder, because a lot of the time advertisers just don’t have the right technologies and tools in order to do so.” —Anh Bui, Product Marketing Manager, Facebook
  • 37. © 2016 eMarketer Inc. Data matching capabilities are improving, but there are still pitfalls
  • 38. © 2016 eMarketer Inc. Location data can prove a rich source of O2O info, but its fidelity is increasingly being called into question
  • 39. © 2016 eMarketer Inc. Example in action: an entertainment client looking to activate monthly paid subscriptions  The goal: Develop a cross-channel, integrated model that drives business results using paid, owned and earned media.  Data in: Primary acquisition KPIs (subscriptions, registrations, monthly active users), paid media (TV, out-of-home, social, online video, display, search), owned and earned media (site traffic, organic social and search, email, PR), competitor media spend, holidays and events, macroeconomic data, company-/industry-specific business metrics.  Model used: Marketing mix modeling (MMM) with ‘nested’ levels for optimization (digital).  Benefits: Determine optimal channel spend; understand how to optimize and execute marketing dollars to maximize subscriptions and lifetime value.
  • 40. © 2016 eMarketer Inc. Key Takeaways:  Marketers can no longer afford to ignore cross-platform attribution.  Organizational, technological and data quality issues all pose significant challenges to making attribution work.  Blended, more real-time models are now a must.  Tying O2O data will be critical to truly gauge company success.
  • 41. Ads that workThe retargeting and prospecting platform of choice for more than 25,000 advertisers worldwide.
  • 42. Do more with first-party data Maximize the value of your customer data with AdRoll. Website and app data Use the AdRoll pixel to leverage your customer data and make it work for you Partner data Easily integrate the data from your CRM lists or favorite marketing apps to power more effective campaigns Opt-in data Take advantage of the AdRoll IntentMap™, the world’s largest opt-in co-op data set, to gain access to valuable new audiences
  • 43. A solution for every stage of the customer journey Every marketer has three core objectives. ATTRACT Reach new prospects CONVERT Turn prospects into customers GROW Maximize customer lifetime value PROSPECTINGRETARGETING
  • 44. Prospecting: Reach new, high-quality audiences Attract the prospects that are most likely to be in the market for your products. Your site’s customer behavior is analyzed and matched with similar prospects in the IntentMap Prospects come to your website via ads. With every new customer, Prospecting gets smarter. “AdRoll Prospecting has been a game changer for our marketing team, helping us find new customers at a rate that eclipses those of other platforms. We’ve seen excellent backend quality metrics from these conversions.” —Rob DiNuzzo, demand generation manager at Datadog Your website
  • 45. Power your campaigns with advanced machine learning BidIQ™, our proprietary algorithm, learns from every bid it makes to buy high impact impressions and meet your performance targets. SCALE 1.2 billion user profiles and growing DIVERSITY An increasingly stronger data set that constantly gathers data across verticals, from SMBs to Fortune 500 companies QUALITY Over 100m variables instantly evaluated per impression (pages visited, time of day, device type, etc.)
  • 46. © 2016 eMarketer Inc. Learn more about digital marketing with an eMarketer corporate subscription Around 200 eMarketer reports are published each year. Here are some recent reports you may be interested in: Q&A Session Made possible by You will receive an email tomorrow with a link to view the deck and webinar recording. To learn more: www.emarketer.com/products 800-405-0844 or webinars@emarketer.com Lauren T. Fisher Navigating the Complexities of Cross-Platform Attribution  Cracking Cross-Device Attribution in 2016: Data Quality, Blended Models and Merging Online-Offline Data  Search and Social Platforms: How Facebook and Others Can Steal Dollars from Search  Making Multichannel Marketing Work: Four Tactics Required for Omnichannel Success  Ad Fraud in the US: How More Sophisticated Methods Are Hurting Mobile, Video and Performance Measurement