Post on 12-Jan-2017
© 2016 eMarketer Inc.
Made possible by
Predictive Marketing—Using Data
Decisively at Every Stage of the Funnel
Jillian Ryan
Analyst
June 23, 2016
© 2016 eMarketer Inc.
Today’s Agenda
Predictive analytics and marketing defined
The importance of data in building predictive models
Exploring predictive applications across the customer
funnel
Measuring predictive marketing’s value
The current state of adoption and what’s on the horizon
© 2016 eMarketer Inc.
Defining Predictive
© 2016 eMarketer Inc.
Many B2B
marketers in
the US don’t
understand
the benefits
of predictive
analytics
© 2016 eMarketer Inc.
Predictive analytics delivers more accurate
insights
Predictive analytics uses
predictive models to estimate
which marketing actions are most
likely to pay off in the future
With more accurate insights,
marketers can determine how to
best encourage sales from
existing and new customers
© 2016 eMarketer Inc.
It’s all about the data—and it’s getting more
sophisticated
The way B2B marketers are
using data is evolving
Predictive is rooted in
forward-looking analysis
Rather than relying on
guesswork, predictive is based
on a trusted, analytical fact
© 2016 eMarketer Inc.
Data:
The Engine Powering
Predictive Technology
© 2016 eMarketer Inc.
Marketers are using data, but sophisticated
practitioners are winning
90%+maintained databases on
customers/prospects,
segmented data for
targeting, and measured
campaign results across
channelsSource: Global Alliance of Data-Driven Marketing Associations and
Winterberry Group, Sept 2015
50%+considered themselves
sophisticated
practitioners, while the
remaining portion were
just performing these
data-driven approaches
to some extent
© 2016 eMarketer Inc.
Historical data reveals patterns in behavior
Data tells a story based
on the past experiences
it has recorded
Predictive analyzes data
touchpoints to determine
patterns
It learns from data to
render predictions
© 2016 eMarketer Inc.
The most common
types of data used for
predictive marketing
included:
Website data (47%)
Demographics (44%)
Digital transactions
(41%)
Social (39%)
© 2016 eMarketer Inc.
Data touchpoints
must be managed
properly
Top DQM services
included:
Big data
Master data
management
Data cleansing and
validation
© 2016 eMarketer Inc.
Bigger companies
are more likely to
leverage historical
campaign analytics
for predictive
modeling
© 2016 eMarketer Inc.
Marketers face
several
challenges
leveraging
historical data
for predictive:
Too expensive
Lack of expertise
Vendor tech isn’t
mature enough
Hard to prove value
© 2016 eMarketer Inc.
Machine learning delivers fact-based predictive
insight
The machine
ingests data from
past successes and
failures, then runs a
model to deliver
predictive scores for
various desired
outcomes
68% of respondents
used it for
predictive analytics
© 2016 eMarketer Inc.
Ensuring accuracy in the predictive model
requires two types of data:
Training Data
Used to construct the model
The majority of data points available
Testing Data
Used to verify the predictive model’s precision
A small subset kept separate from the training data
© 2016 eMarketer Inc.
As the machine learns, the model gets smarter
The predictive engine
continues to learn from
data touchpoints
With more experience it
renders more accurate
predictions
© 2016 eMarketer Inc.
Predictive Marketing:
Putting Insights into
Action
© 2016 eMarketer Inc.
Moving from predictive analytics to marketing
Every application of predictive analytics follows the
same two-part structure:
1. Understanding
what is being
predicted
2. Figuring out what
to do with that
prediction
© 2016 eMarketer Inc.
A standard predictive team requires three players*
1. A demand marketing expert
2. A data scientist
3. A technologist
*They all need to communicate well
© 2016 eMarketer Inc.
Putting the pieces of the predictive puzzle in place
Marketers need to know what
type of business problem they
are trying to solve to take
action
Figuring out the business case
will kick off the process
© 2016 eMarketer Inc.
Predictive marketing can be used to improve
the funnel at every stage
43%
of US B2B marketers
used predictive analytics
to get insights about
where prospects are in
the sales funnel
Source: 6sense and OnTarget Consulting & Research, April 2015
26%
of marketing executives
in North America said
that a benefit of
predictive marketing was
better funnel conversions
Source: Forbes Insights, April 2015
© 2016 eMarketer Inc.
Primary objectives for predictive spanned across
the customer funnel
The major objective, at
33%, was customer
acquisition
At 17% each:
– measuring behavior and insights
– ad/campaign effectiveness
– calculating and improving
lifetime value
– retention
© 2016 eMarketer Inc.
Applying predictive marketing at every stage
of the funnel
Finding new prospects
Lead scoring
Lifetime value of current
customers
Personalization
Sales/channel support
Recommendations
Awareness
Consideration
Purchase
Loyalty
Advocacy
© 2016 eMarketer Inc.
Finding new prospects
51%
of US B2B marketers
used predictive analytics
to find new prospects
Source: 6sense and OnTarget Consulting & Research, April 2015
37%
of marketing executives
in North America used
predictive marketing
technologies to find new
prospects
Source: Forbes Insights, April 2015
© 2016 eMarketer Inc.
Kasasa uses predictive marketing for pre-funnel
scoring
“They use predictive modeling to score potential
buyers before they enter the funnel. If a bank
doesn’t have the right combination of attributes like, for
example, specific turnover rates, profit margins or number
of customers, it’s eliminated. But prospects with a
high predictive score are then proactively
marketed to, because the model found they are most
beneficial to talk to and more likely to convert.”
—Phil Winters, Author, “Customer Impact Agenda: Doing Business
from the Customer’s Perspective”
© 2016 eMarketer Inc.
Lead scoring
B2B marketers can use
predictive models to score
leads to determine which
individuals have a higher
propensity to convert
Experts note this is the
most common use case
Still, only 7% of B2Bs used
predictive lead scoring
© 2016 eMarketer Inc.
Getty Images scores leads with predictive to
prioritize sales efforts
“Using custom-built predictive indicators, we can take a set
of 10,000 new leads that we’ve identified and then sort
them in priority order based on the data that is
available to us, and then use a predictive
model to understand which leads are more
likely to convert. … Sales can understand which leads
have the highest propensity to buy. That information
allows them to prioritize their efforts.”
—Jason Widup, Senior Director, Demand Generation and Marketing
Operations, Getty Images
© 2016 eMarketer Inc.
Increasing customer lifetime value
32%
of data management
professionals worldwide
used predictive analytics
to predict lifetime value
of each customerSource: Experian Data Quality and Dynamic Markets Ltd.,
Dec 2014
14%
of marketing executives
used predictive
marketing to forecast
and minimize churn
within current accountsSource: Forbes Insights, April 2015
© 2016 eMarketer Inc.
Major US wireless carrier uses predictive to
understand the value of its B2B customers
Predictive models help the
carrier acquire companies
In the funnel, it determines
which customers will have
the highest lifetime value
– add more lines
– upgrade to higher data buckets
– add more premium devices
© 2016 eMarketer Inc.
Getty Images predicts churn within its current
accounts to minimize cancellations
When the predictive engine
detects a decline in usage
or another attribute
associated with churn, the
machine flags that account
This triggers a sales or
marketing action to prevent
the loss
© 2016 eMarketer Inc.
Predictive techniques for personalization still
have a way to go
While 41% used
predictive to accelerate
personalization, only 12%
said predictive analysis
was an essential criterion
for personalization
success
Personalization might be
the end goal, but it is a
strategy of predictive that
needs refinement
© 2016 eMarketer Inc.
A shortcoming to predictive personalization is
the lack of a human touch
“Most marketers want to hit the ‘easy’ button, but
marketers need to tailor the message and be
more empathetic. The problem is that empathy is
not something that predictive analytics—the way
it’s being sold today—is doing at all. It doesn’t provide you
with empathy to understand the situation and the
disposition of who you are talking to.”
—Tim Hayden, Vice President, Marketing, Zignal Labs
© 2016 eMarketer Inc.
HP employs predictive personalization through
recommendations in partner channels
Partner channel Sales Central uses predictive to tailor
solutions to each third-party channel partner or reseller
The hub delivers personalized suggestions for more
successful campaigns and sales tactics for each partner
Then, personalized recommendations provide a tailored
campaign or piece of content
This isn’t a direct-to-customer approach, but overlaps with
two other essential uses of predictive marketing
technology: sales support and recommendations
© 2016 eMarketer Inc.
Measuring Predictive
Marketing’s Value
© 2016 eMarketer Inc.
Predictive marketing is seen as moderately
valuable
Extent to which investments in intelligent software, like predictive, are
delivering expect business value for worldwide executives:
Exceeding: 16%
Expected value: 33%
Some value: 39%
Underperforming: 8%
Source: Economist Intelligence Unit (EIU), sponsored by Accenture and Pegasystems, April 2015
© 2016 eMarketer Inc.
How impactful is predictive marketing?
Just under half of B2B
marketing decision-makers
said predictive analytics
had a very high impact on
their business
Over one-third (36%) said
there was considerable
impact
13% cited some impact
Only 5% claimed to see
little or no impact
© 2016 eMarketer Inc.
Benchmarking predictive vs. nonpredictive
determines effectiveness
55% of US B2B
marketers who used
predictive said their
strategy was effective,
compared with 18%
who didn’t use
predictive
Dell uses A/B tests and
compares the results
to see predictive’s
effect
© 2016 eMarketer Inc.
The most common
metrics used to
measure the success
of predictive
marketing included:
customer retention rates
customer value
cost per lead
projected ROI
total conversions
© 2016 eMarketer Inc.
Predictive is not perfect … but it can still work
Data-derived predictions
will never be 100% accurate
all of the time
This isn’t science fiction or
a crystal ball
Predictive equips marketers
with information that they
need to be smarter and
more efficient
© 2016 eMarketer Inc.
Current State of B2B
Predictive Marketing Adoption
© 2016 eMarketer Inc.
Predictive tech firms say the tipping point is now
While only 31% currently used a predictive analytics
tool, 54% were likely to invest in predictive analytics in
the next 12 to 36 months. Source: 6sense, April 2015
38% planned to significantly increase the role of
predictive analytics in the next year. Source: Lattice, Oct 2015
68% believed predictive marketing would be key to the
marketing stack. 47% were investigating how to use
predictive, and 25% were currently using it. Source: Everstring, Sept 2015
© 2016 eMarketer Inc.
Sources other than vendors tell a different story
Tactics that occupied vs. will occupy time and resources in 2015 and 2016
2015 2016
Predictive modeling 44.4% 44.4%
Source: IAB Data Center of Excellence and Winterberry Group, January 2016
© 2016 eMarketer Inc.
Why such a discrepancy?
Yes, predictive vendors might have a self-serving
reason to be biased …
… but the more likely scenario is that predictive
techniques can be bundled into larger marketing
automation tools
Thus, predictive marketing’s adoption might be much
larger than we think, because the capability might be
folded into other marketing tech products
© 2016 eMarketer Inc.
Predictive undercover in Google Analytics
Google Analytics offers predictive capabilities without
actually labeling them as such
The Smart Goals feature uses machine learning to
identify which website visits are most high-value and
have a higher propensity to convert
Google doesn’t call Smart Goals predictive because
many marketers using it might be turned off or
intimidated by words like “predictive” or “machinery”
© 2016 eMarketer Inc.
49%
expected growth of
predictive intelligence,
according to marketing
leaders in the US
Source: Salesforce, Jan 2016
38%of marketing executives
worldwide said predictive
analytics will have the
biggest impact on
marketing companies by
2020
Source: Economist Intelligence Unit (EIU), April 2016
Predictive is booming today, and likely tomorrow
© 2016 eMarketer Inc.
Predictive will be here for a while and has a long
way to go
“As investors continue to pump funds into
predictive vendors and technology, the
capabilities will mature very rapidly. We
are in the early stages right now, but
predictive marketing isn’t going
anywhere just yet.”
—David Rabb, Principal, Marketing Technology and
Analytics, Raab Associates
© 2016 eMarketer Inc.
Key Takeaways
Predictive models, built from past customer touchpoint data,
guide marketers by estimating which actions are most likely to
pay off in the future
Using this and other data allows for more efficiency at all
points of the marketing funnel
Measuring ROI can be a challenge, but benchmarking past
performance will indicate if the method is adding value
B2B companies are slowly adopting predictive marketing and
using it within their marketing technology stack
It remains uncertain whether this is the year predictive
marketing goes mainstream
©2016 EverString
Predictive Marketing Making Marketing More Relevant
June 2016
©2016 EverString
Predictive marketing starts with Audience Selection
Audience Selectionis the process of using data
to identify the most relevant
prospects…
So you can market and
sell to the right prospects,
even if they are not in
your funnel
©2016 EverString
Focus of predictive marketing to date
Build Pipelinewith Predictive Demand
Generation
Increase Conversion with Predictive Scoring
Total Potential Market (all companies)
Predictive Scoring prioritizes your
existing funnel
customer
model
A
B
C
DTotal Captured Marketwithin your CRM
& marketing automation
Predictive Demand Generation delivers net new, high value accounts
©2016 EverString
A breakthrough in predictive marketing
EverString Audience Platform
powered by the
EverString Company Graph
Uses semantic similarity to map
the connections & similarities
between companies
Like the Google knowledge graph
and Facebook people graph,
but for B2B company data
The EverString Company Graph maps the
known universe of 11M B2B Accounts
Total Potential Market (TPM)
©2016 EverString
An evolution in predictive marketing
Data science helps identify and expand
segments beyond the boundaries of your
funnel, creating…
Predictive Segments
…which use semantic similarity to find
look-alike accounts that are similar on
specified dimensions
These segments can be mapped to your
products, territories, teams, events,
campaigns and more
EverString B2B Graph: known universe of accounts
Total Potential Market
Total Captured Market (TCM)
Segment
Segmentsegment
Segment
©2016 EverString
An evolution in predictive marketing
EverString B2B Graph: known universe of accounts
Total Potential Market
Total Captured Market (TCM)
Segment
Segmentsegment
Segment
Use cases to consider1. Funnel Prioritization – use scoring to prioritize
the prospects in your funnel
2. Segmented nurture – uses segments to optimize
the nurture paths for your existing database
3. Multi-channel Programs – create and use a
segment to market a campaign, program or
event
4. New market expansion – identify a wholly new
market to expand into and get accounts today
5. Territory Planning – create a segment for an
individual territory for a rep or team
6. Account Based Marketing – identify target
accounts and preform more directed campaigns
© 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
Jillian Ryan
Predictive Marketing—Using
Data Decisively at Every Stage
of the Funnel
Predictive Analytics in B2B Marketing: Using Data Decisively,
at Every Stage of the Funnel
B2B Sales Enablement: Driving Strategic Efficiencies Along the
Path to Purchase
Marketing Technology: The Six Developments that Matter the
Most in 2016
B2B Content Marketing in the US: Maximizing ROI and
Cost-Effectiveness over Time