From data collection
to action:
Practical use-cases
MAXIME VANDENBUSSCHE
eBusiness Consulting Manager
WHO ARE WE?
FREDERIC PALELLA
eBusiness Consultant
FOOD
&
DRINK
SEMETIS
BUFFAL
OOOOO
AUTHOR
(to be)
REAL
MADRID
TECH
Data-Driven Digital Media Agency
Connecting digital advertising & business intelligence with data
Our services
We offer unique capabilities in the
management of digital advertising &
digital business intelligence, data being
their connector
DIGITAL ADVERTISING DIGITAL BUSINESS INTELLIGENCE
data
Facing digital marketing evolutions
with our DNA
#open #share #knowledge
Business minded
Transparent & technology agnostic
Data & technology-driven
Our 4-step model
collection unification
activation insights
The 4 core capabilities of your future architecture
1 2
34
The 4 core capabilities of your future architecture
1
2
4
3
Case 1: Eurotunnel
Audience segmentation
A short introduction to Eurotunnel
The Challenge of Eurotunnel
Eurotunnel presence in the Benelux is still in an early-stage business compared to the UK
and France. The steps toward business growth:
• Market development: expansion of the potential market to new customers or new
users
• Market penetration: Competing to capture a greater share of an established
customer base
Our own data
How are we going to achieve
market growth on the continent?
Leverage our own data and
audiences
Audiences or segments are
clusters of users that have
common denominators
Leveraging the full potential of audiences
What most implementations
look like
What implementations should
look like
audiences
generic
messages
messa
ges
audien
ces
audiences relevant
messages
reasons
Data pools for acquisition & retention
BRAND
FARE FINDER
TRAVEL OPTIONS
PERSONAL DETAILS
PAYMENT
SALE
POST SALE FLOW
ACQUISITION
WEB SCORING
SYSTEM
RETENTION
IN HOUSE
SCORING SYSTEM
Predicting based on collected data
Profile A
Profile B
Profile C
Message A
Message B
Message C
The objective of the audience-based approach is to determine a profile based on what we can measure in
the funnel. This allows us to tailor a message to each audience, increasing our impact on user that entered
the funnel but did not buy.
Impacting ROI & retention through
improved acquisition
As a consequence, our delivered messages are more relevant to the customers:
(1) So we can convert / acquire more users (within the same volume)
Immediate impact - With a more granular segmentation, we’ll also be able to have a better prioritization of
remarketing budgets
Profile A
Profile B
Profile C
High ROI
Medium ROI
Low ROI
High investment
Medium investment
No investment
(2) So we can be more relevant after the sale so we can retain the customer
How do we get there?
Leveraging the backend - Checkout
Building a future-proof infrastructure
Tag Management
System
Analytics Tool
Ad Server DSP
dataLayer
Tag Management Implementation
Audience Data Integration
Implementing tag management structure cross-countries
Ensuring consistency and good naming convention
Mapping all the steps of the different funnels of Eurotunnel
Creating of multi-funnel step dataLayer tracking
Enhanced ecommerce implementation
Funnel tracking improvement
Integrating multiple data sources (CRM, Sitecore,..)
Pushing data sources to relevant advertising platform
Integrating custom Javascript to identify last source before
conversion
From data to audience clusters
Sales
data
Profile building
Remarketing with
adequate message
Retrieving
consumer insights
for future
informed
decision making
digital
marketing
We created dimensions and
metrics for all variables available
in the backend in the Web
Analytics tool. We conducted an
analysis to discover the key
revenue denominators
Location is a key variable
Provinces closed to the coast are the biggest sales generators. Brussels is the cheapest in
terms of acquisition. The further away the location to the terminal, the higher the AOV.
Province #Sales Revenue CPA AOV Distance
South Holland
1000 € 200.000 € 5,37 € 200
321 km
Flemish Brabant
978 € 185.771 € 3,95 € 190
216 km
Antwerp
974 € 183.082 € 5,70 € 188
204 km
East Flanders
889 € 176.429 € 5,75 € 199
151 km
West Flanders
864 € 173.680 € 4,87 € 201
123 km
Brussels
767 € 152.431 € 3,73 € 199
198 km
North Brabant
756 € 177.894 € 4,87 € 235
296 km
North Holland
725 € 144.379 € 5,67 € 199
390 km
Booking horizon & vehicle type
CAR VAN CAMPER
URGENT
SHORT
TERM
LONG
TERM
URGENT
SHORT
TERM
LONG
TERM
URGENT
SHORT
TERM
LONG
TERM
VEHICLE TYPE BOOKING HORIZON % REVENUE ROI
CAR URGENT  22% 3.2
CAR SHORT TERM 15% 1.8
CAR LONG TERM   10% 2.4
VAN URGENT   15%  2.3
VAN SHORT TERM 10% 1.9
VAN LONG TERM   8% 3.1
CAMPERVAN & MINIBUS URGENT   3%  3.1
CAMPERVAN & MINIBUS SHORT TERM   5% 2.4
CAMPERVAN & MINIBUS LONG TERM 12%  2.2
Pushing those dimensions into your
platforms to create audiences
Bottom line impact
CONVERSION RATE
+ 13 %
over last 2 years
RETURN ON INVESTMENT
+ 18 %
over last 2 years
Case 2: Carglass
Increase funnel conversion rate
through DCO
Top Belgian brand
Willingness to explore &
improve
Changing market conditions
Increasing competition
The Problem
Main Challenges
Main Challenges
High volume of combinations Possibility to test and adapt
Option to test at least 2 call to
action variables.
2 call to action variations
1680 banner variations
15 top car brands
2 product possibilities
2 languages
7 formats
2 weather types
840 banner variations
*
*
*
* *
Huge production costs
No need to say that the
production of (at least) 840
banners is a time consuming
and thus expensive task.
Moreover, this approach isn’t
easy to scale as more brands
become available.
Increased risk of error
As all tasks are manual, the
risk of errors being introduced
in high. Moreover, as multiple
stakeholders are involved the
process can be heavy / slow.
Main Challenges
Dynamic Creative Automation
Use DoubleClick Studio in order to automate
the asset creation, assembly and delivery
process using a transparent build-up process.
The Solution
Variable 1 = repair / replace
Variable 2 = car brand
We decided on 2 variables that make a
difference visually and are also at the start of
the funnel to be able to personalize early on
Variables
The Variables
Different animations based on product
The Variables
WHEN IT ALL COMES TOGETHER
Variables and content come together in
a feed
Based on the variables, the banners
dynamically load in the content
matching the scenario
The feed
ZOOM IN ON EXAMPLE
The feed
Video
What’s next?
DCO for Video
Take-aways
Take-aways
Your own data (GA) can be leveraged in many different ways
#1 Segment your data to discover new business opportunities
#2 Use your data to personalize your messaging
#3 Explore new digital KPI’s to make business decisions
Thank you

From data collection to action

  • 1.
    From data collection toaction: Practical use-cases
  • 2.
    MAXIME VANDENBUSSCHE eBusiness ConsultingManager WHO ARE WE? FREDERIC PALELLA eBusiness Consultant FOOD & DRINK SEMETIS BUFFAL OOOOO AUTHOR (to be) REAL MADRID TECH
  • 3.
    Data-Driven Digital MediaAgency Connecting digital advertising & business intelligence with data
  • 4.
    Our services We offerunique capabilities in the management of digital advertising & digital business intelligence, data being their connector DIGITAL ADVERTISING DIGITAL BUSINESS INTELLIGENCE data
  • 5.
    Facing digital marketingevolutions with our DNA #open #share #knowledge Business minded Transparent & technology agnostic Data & technology-driven
  • 6.
  • 7.
    collection unification activation insights The4 core capabilities of your future architecture 1 2 34
  • 8.
    The 4 corecapabilities of your future architecture 1 2 4 3
  • 9.
  • 10.
    A short introductionto Eurotunnel
  • 11.
    The Challenge ofEurotunnel Eurotunnel presence in the Benelux is still in an early-stage business compared to the UK and France. The steps toward business growth: • Market development: expansion of the potential market to new customers or new users • Market penetration: Competing to capture a greater share of an established customer base
  • 12.
    Our own data Howare we going to achieve market growth on the continent? Leverage our own data and audiences
  • 13.
    Audiences or segmentsare clusters of users that have common denominators
  • 14.
    Leveraging the fullpotential of audiences What most implementations look like What implementations should look like audiences generic messages messa ges audien ces audiences relevant messages reasons
  • 15.
    Data pools foracquisition & retention BRAND FARE FINDER TRAVEL OPTIONS PERSONAL DETAILS PAYMENT SALE POST SALE FLOW ACQUISITION WEB SCORING SYSTEM RETENTION IN HOUSE SCORING SYSTEM
  • 16.
    Predicting based oncollected data Profile A Profile B Profile C Message A Message B Message C The objective of the audience-based approach is to determine a profile based on what we can measure in the funnel. This allows us to tailor a message to each audience, increasing our impact on user that entered the funnel but did not buy.
  • 17.
    Impacting ROI &retention through improved acquisition As a consequence, our delivered messages are more relevant to the customers: (1) So we can convert / acquire more users (within the same volume) Immediate impact - With a more granular segmentation, we’ll also be able to have a better prioritization of remarketing budgets Profile A Profile B Profile C High ROI Medium ROI Low ROI High investment Medium investment No investment (2) So we can be more relevant after the sale so we can retain the customer
  • 18.
    How do weget there?
  • 19.
  • 20.
    Building a future-proofinfrastructure Tag Management System Analytics Tool Ad Server DSP dataLayer Tag Management Implementation Audience Data Integration Implementing tag management structure cross-countries Ensuring consistency and good naming convention Mapping all the steps of the different funnels of Eurotunnel Creating of multi-funnel step dataLayer tracking Enhanced ecommerce implementation Funnel tracking improvement Integrating multiple data sources (CRM, Sitecore,..) Pushing data sources to relevant advertising platform Integrating custom Javascript to identify last source before conversion
  • 21.
    From data toaudience clusters Sales data Profile building Remarketing with adequate message
  • 22.
    Retrieving consumer insights for future informed decisionmaking digital marketing We created dimensions and metrics for all variables available in the backend in the Web Analytics tool. We conducted an analysis to discover the key revenue denominators
  • 23.
    Location is akey variable Provinces closed to the coast are the biggest sales generators. Brussels is the cheapest in terms of acquisition. The further away the location to the terminal, the higher the AOV. Province #Sales Revenue CPA AOV Distance South Holland 1000 € 200.000 € 5,37 € 200 321 km Flemish Brabant 978 € 185.771 € 3,95 € 190 216 km Antwerp 974 € 183.082 € 5,70 € 188 204 km East Flanders 889 € 176.429 € 5,75 € 199 151 km West Flanders 864 € 173.680 € 4,87 € 201 123 km Brussels 767 € 152.431 € 3,73 € 199 198 km North Brabant 756 € 177.894 € 4,87 € 235 296 km North Holland 725 € 144.379 € 5,67 € 199 390 km
  • 24.
    Booking horizon &vehicle type CAR VAN CAMPER URGENT SHORT TERM LONG TERM URGENT SHORT TERM LONG TERM URGENT SHORT TERM LONG TERM VEHICLE TYPE BOOKING HORIZON % REVENUE ROI CAR URGENT  22% 3.2 CAR SHORT TERM 15% 1.8 CAR LONG TERM   10% 2.4 VAN URGENT   15%  2.3 VAN SHORT TERM 10% 1.9 VAN LONG TERM   8% 3.1 CAMPERVAN & MINIBUS URGENT   3%  3.1 CAMPERVAN & MINIBUS SHORT TERM   5% 2.4 CAMPERVAN & MINIBUS LONG TERM 12%  2.2
  • 25.
    Pushing those dimensionsinto your platforms to create audiences
  • 26.
    Bottom line impact CONVERSIONRATE + 13 % over last 2 years RETURN ON INVESTMENT + 18 % over last 2 years
  • 27.
    Case 2: Carglass Increasefunnel conversion rate through DCO
  • 28.
    Top Belgian brand Willingnessto explore & improve Changing market conditions Increasing competition The Problem
  • 29.
  • 30.
    Main Challenges High volumeof combinations Possibility to test and adapt Option to test at least 2 call to action variables. 2 call to action variations 1680 banner variations 15 top car brands 2 product possibilities 2 languages 7 formats 2 weather types 840 banner variations * * * * *
  • 31.
    Huge production costs Noneed to say that the production of (at least) 840 banners is a time consuming and thus expensive task. Moreover, this approach isn’t easy to scale as more brands become available. Increased risk of error As all tasks are manual, the risk of errors being introduced in high. Moreover, as multiple stakeholders are involved the process can be heavy / slow. Main Challenges
  • 32.
    Dynamic Creative Automation UseDoubleClick Studio in order to automate the asset creation, assembly and delivery process using a transparent build-up process. The Solution
  • 33.
    Variable 1 =repair / replace Variable 2 = car brand We decided on 2 variables that make a difference visually and are also at the start of the funnel to be able to personalize early on Variables The Variables
  • 34.
    Different animations basedon product The Variables
  • 35.
    WHEN IT ALLCOMES TOGETHER
  • 36.
    Variables and contentcome together in a feed Based on the variables, the banners dynamically load in the content matching the scenario The feed
  • 37.
    ZOOM IN ONEXAMPLE The feed
  • 38.
  • 39.
  • 40.
  • 41.
  • 42.
    Take-aways Your own data(GA) can be leveraged in many different ways #1 Segment your data to discover new business opportunities #2 Use your data to personalize your messaging #3 Explore new digital KPI’s to make business decisions
  • 43.