This document discusses transforming marketing approaches to focus more on customers. It advocates taking a customer-centric view of marketing performance rather than focusing on individual touchpoints. This involves defining customers and looking at their overall behavior over multiple interactions. It also discusses using data and modeling to better understand customer value, segmentation, experimentation and integrating marketing metrics with other business goals and indicators. The overall message is that marketing needs to move beyond isolated touchpoint views to consider the full customer experience and journey.
2. KEY TAKE-AWAYS
> Attribution problem points out flaw in structuring GTM around
traditional funnel – better is to look at cycle of buying and align
metrics along a common thread
> Customer-centric approach and data suggests making customer-level
ROI the common thread for marketing performance, not per-touch
ROI
> A critical step to providing marketing performance metrics that
matter is to define the customer and look at customer behavior for a
collection of touches
2
3. ANSWERS YOU CAN TRUST
MORE IMPORTANT TO MAKE IT REAL EMPIRICAL AND
IMPORTANT THAN REAL TIME AND TRIVIAL
> Strategic means big numbers, enterprise level
> Challenges happen when numbers don’t correspond to other known
realities or expectations – avoid internal inconsistencies and limit
assumptions
> Evidence need not be highly granular or zero latency
“The most exciting phrase
to hear in science, the
one that heralds new
discoveries, is not
‘Eureka!’ (I found it!) but
‘That's funny.’”
Isaac Asimov
3
4. CRAWL – WALK – RUN
Crawl
> Mobilize and normalize data across media,
touchpoints
Where do I place my
> Model individual channel elasticity
bets?
> Model cross-channel interaction
> Benchmark top line metrics Best ROI for next $
> Adjust budgets periodically by channel /
product / region
Walk
> Incorporate new media in multichannel
models What does it mean for
> Allow for rapid fine-tuning and optimization what customers will do?
with increased granularity
> Build Designs of Experiment for test / learn Embrace complexity,
> Provide segment view enhance engagement
> Consider non-MARCOMS aspects of GTM
Run
> Include all touches in 360o framework – How to have real
inbound and outbound conversations that
> Enhance customer-centricity through next- increase CLV
gen models 4
5. ATTRIBUTION DECISIONS
Identify goals and measures of success
Align and charter what will make actionable work acted upon
Behavioral segmentation and CLV modeling
Connect touches to people
Connect touches to outcomes
Connect people to outcomes
Evaluate interaction effects / synergies / halo
Connect other aspects of GTM to people and outcomes
DOE, market tests and feedback
5
6. THE PROBLEM OF ATTRIBUTION
> What is at stake here?
> What is wrong with the question?
> What is a pragmatic way to deal with this?
7. THE PROBLEM OF ATTRIBUTION
" I know what some of my tactics contribute to sales
because I can follow the contact stream all the way
through.
In other cases I don't know because I don't have a direct
link in the data from a message we sent, to a customer's
receipt of the message, to the customer connecting and
completing a transaction.
How do I resolve this to learn what my real ROI is by
channel, campaign and tactic?"
8. THE PROBLEM OF ATTRIBUTION
Trying to get high
performance out of too
small an engine:
The relationship is
what must pay, not the
touch
9. HOW PURCHASES HAPPEN
More like this
Less like this
But metrics still assume
it happens this way
9
10. CRAWL
WHERE DO I PLACE MY BETS?
> Compare elasticity by channel against a common KPI, e.g. sales, to
allocate media spend more efficiently
Search
Response Curves Limitations
16.00
> Single metric may not apply to
Millions
Display
15.80
each channel in straight
Radio
forward ways
Response curves 15.60
> Tends to mask interaction
Total Product Units
15.40 Print between touches and other
E-mail TV
factors
15.20
OOH
> Good for limited range of
15.00 channels or input variables
14.80 Direct Mail > Not so flexible for new data
streams or what if scenarios
14.60
> Limited predictive applicability
$10
$15
$20
$25
$30
$34
$40
$45
$50
$55
$60
$65
$70
$75
$80
$85
$90
$95
$100
$105
$110
$115
$120
$125
$130
$135
$140
$145
$150
$155
$160
$165
$170
$175
$180
$185
$190
$195
$200
$205
$0
$5
Media Investment ($M)
10
11. WALK
MAKING MARKETING MORE CUSTOMER-
CENTRIC THROUGH SEGMENTATION AND
EXPERIMENTATION
> Identify individuals and
households across
touches and segment
> Bucket style and
+ 500%
frequency of
communication by
segment
> Natural clusters Contribution to
> CLV Overall Profit
> Identify key inbound Customer Value
variables and match to Top 30% Next 50% Bottom 20%
Segment
segments
> Taylor outbound Contribution to
messaging to segment Overall Profit
variables for increased
relevance - 400%
> Test, test, test
11
12. RUN
DRIVING CUSTOMER VALUE AS KPI – LINKS
OTHER INDICATORS IN A SCORECARD
> Develop triggers based on life
stage or key predictors of Messaging
preference or need
Responding
> Customer journey
> Translate operational metrics
into scorecard metric, ideally
CLV Customer
> Integrate “back office” and Work
“front office” customer
touches in one conversation
moving in both directions Family
> Layer individual contact via
social strategy and personal
contact / offer strategy
12