Do you have a Data Strategy?
Andreas Grasel
Thorsten Feldhege
ProgrammatiCon 11.10.2018
Introduction to Adform
!2
Adform is the Leading European AdTech Provider
!3
Gustav Mellentin
CEO & Co-Founder
Jakob Bak
CTO & Co-Founder
Stefan Juricic
CPO & Co-Founder
Founded in Denmark in 2002 by 3 code-writing entrepreneurs, Adform is the industry’s only
enterprise full-stack platform built from the ground up as one integrated advertising platform.
Founded and Focused in Europe with a Global Footprint
!4
880+
Employees
19
Countries
360
Developers
250
Client Services
8
Data Points of
Presence
APAC Activation Hub
Global Competency
Center & EMEA Hub
Americas Activation
Hub
Adform is Uniquely Positioned
!5
Adform provides Full-Stack capabilities
CROSS- DEVICE /

CROSS-CHANNEL
MobileDesktop Video DOOHPrint TV
Demand Side Platform (DSP)Data Management Platform (DMP) Ad Server (TPAS) Creative Site Platform (CSP)
Audio Native
Rich Media – Site tracking - HTML5 Studio -
Dynamics / Rotators - Dynamic Creative
Optimization (DCO) - Creative Library
Direct (AG/ IO) - RTB - Marketplace –
PMP - Real-Time Forecasting - CPC, CPM,
CPA, ROAS, vCPM
Audience Builder - Audience Extension -
High-Reach 3rd Party Data - Taxonomy
Builder
Integrations - API
Intuitive UI →
Simple workflows
across platforms
DATA CREATIVITY TRADING
Cross-Device - Client-Side Encryption of Identifiers (PII) - Probabilistic + Deterministic Data
Social
CROSS-DEVICE DATA ATTRIBUTIONTRACKINGSUPPLY
+ many more + many more
→
Adform takes care of our Customers Data
• Fully compliant with international privacy laws
• Data is never collected without user consent
• Only non-PII information for probabilistic cookie matching
• All profiles based on Cookies & Device IDs
• All user-relevant data is one-way encrypted before stored on Adform system
• Annual Deloitte audits and security penetration tests
Retain protection, control and management of all 1st, 2nd and 3rd party data as well as ROI and ROAS associated with
your campaigns from one single UI. Our data centers have the most stringent security processes.
Protection Ownership Control Integrations
!7
… with deep Vertical Expertise
!8
Who we are
!9
Thorsten Feldhege
Platform Solutions Consultant DACH
Andreas Grasel
Commercial Director AT & CH
What to expect from this Session
!10
Demystifying Data
Strategy
The Four Cornerstones
of a Data Strategy
Man versus MachineResponsibilities and
Operations
Demystifying Data Strategy
!12
DMPs segment data from multiple
different sources, configure it, and make
it available to marketers to activate in
digital marketing
But: DMPs are no Data Warehouses
Data Strategy and Goal Definition
Process of DMP Implementation
!13
Preparation Setup
Data Privacy Alignment
• Data Security
• Data Privacy
• Data Storage
• Data Usage
Data Strategy and
Goal Definition
• Goal Definition
• Business Case
• Use Cases
• Stakeholder
Technology
Selection
• Requirements
• RFI/RFP
• Scoring
• Contract
Data Ingestion
• 1st Party Data
• 2nd Party Data
• 3rd Party Data
• APIs
DMP Setup
• Naming Conventions
• Taxonomy
• Segmentation
• KPI Definition
DMP Workflows
• Accounts
• Access Authorization
• Business Processes
• Campaign Setup
Change Management
• Organization
• Responsibilities
• Know-How
• Workflows
Run
DMP Operations
• Analysis
• Optimization
• Campaign Management
• Quality Management
1 2 3 4 5 6 7 8
!14
The Four Cornerstones of a
Data Strategy
What is a Data Strategy?
The four cornerstones:
• Collection
• Hierarchy
• Management
• Activation
If one part of the data strategy fails, the
whole project will fail!
!16
Collection
• All channels have to be integrated to
ensure a holistic overview.
• In a perfect world the customer journey is
taken into consideration to show
interactions between channels.
• Data privacy is not be taken lightly and
should be considered from the very
beginning.
• Data collection is the foundation of a
functioning data hierarchy and the
beginning of every data optimization
circle!
!17
Involved Systems
Collection
!18
Collection - Do's and Don'ts
• Plan ahead - every bit of acquired
information can be worth a lot if combined
with existing data
• Track every channel - a complete picture
will yield better results
• Start small, think big
!19
• Rush tracking implementation
• Hesitate to invest development resources
into tracking
• Leave out channels because of “collection
investment”
Hierarchy
• DMPs do not offer a truly persistent long
term data history. This task is handled by
a specialized system within the enterprise
environment. Namely this can be a
business intelligence system and / or a
CRM.
• Ad-Hoc data should be handled by the
DMP.
• A working and connected data
hierarchy is mandatory for efficient
and correct data management!
!20
Hierarchy – Involved Systems
!21
Customer Data
POS
Offline
BI / CRM PLATFORM DATA MANAGEMENT PLATFORM
Web &

App Data
Campaign
Data
Search & 

Social Data
Hierarchy - Do's and Don'ts
• Link any in-house or on premise system to
online sources via an ID matching
• Integrate the campaign data stream with all
On-site and CRM streams within the data
hierarchy
• Integrate offline systems (store visits, etc.)
!22
• Underestimate the complexity of combining
data sources
• Hesitate to build up in-house infrastructure
• De-prioritize data process documentation
Management
• The data provided by the data hierarchy
must be analyzed and segmented.
• A liaison between business intelligence
1st party data and external DMP
information (2nd and 3rd party) will yield
the best results.
• Data has to be available in a pre-
segmented, exportable way within the
enterprise owned environment and
ready to use in the DMP to be able to
activate it!
!23
Management – Involved Systems
!24
POS
Offline
BI / CRM PLATFORM DATA MANAGEMENT PLATFORM
Web &

App Data
Campaign
Data
Search & 

Social Data
3rd Party 

Branded Data
2nd Party
Partner Data
Customer Data
Management - Do's and Don'ts
• Keep the complexity low and understandable
• Compare and adapt segments on a regular
basis
• Structure the data in an easy to understand
taxonomy
• Think bottom up: Single attributes can be
combined in the DMP
• Be creative and innovative - even if an idea fails
you have learned something
!25
• Hoard data that is not necessary for any
channel or use case
• Be to granular - combining data is easy -
subtracting audiences causes confusion
Activation
• The segmented data has to be used in
every system that is capable of data
usage to ensure a seamless
communication.
• Every data usage outcome has to be
measurable and reportable to allow an
optimization of the segmented data.
• A loopback mechanism to
retrospectively optimize the used data
is mandatory to achieve a closed data
loop.
!26
Activation – Involved Systems
!27
POS
Offline
BI / CRM PLATFORM DATA MANAGEMENT PLATFORM
Web &

App Data
Campaign
Data
Search & 

Social Data
Ad Server Demand Side
Platform
Google Adwords &
Facebook
Other Platforms
3rd Party 

Branded Data
2nd Party
Partner Data Data
Marketplace
Data Deals with
Demand Side
Data Monetization
Data Activation
Customer Data
Activation - Do's and Don'ts
• Test datasets against each other
• Activate data in a structured way -
otherwise learnings are hard to see
• Generate data about your data -
documentation is king
!28
• Let the knowledge be unused
• Wait for the “right” moment to use the data
• Overreact when the planned results do not
show up immediately
• Overengineer data and campaign setup
Responsibilities and
Operations
Which Department owns the DMP?
“Which department is the main sponsor for the data management platform
strategy, rollout, and management at your organization?”
!30
Source: Forrester Study, “Savvy Publishers Deploy DMP Insights Across Their Organizations” (2016)
Which Department owns the DMP?
Persistence Prediction Activation
CRM / BI Platform Data Management Platform
Continuous Feedback Loop
!31
Business Intelligence
Marketing
IT
• Segmented Pseudonymous Data
• Non-PII Data
• 2nd & 3rd Party Data
• Audience Extension / Cross Device Mgt.
• Own Custom Algorithms
• Based on own IT Security
and Privacy
• CRM / Data Warehouse / CDP
• PII-Data
- Flexible Structures
- Virtually no Restrictions
Human versus Machine
!33
There are different Types of Data
Derived DataDeclared Data
● Name: Thorsten Feldhege
● Gender: Male
● Age: 40-45
● Email: thorsten.feldhege@adform.com
● Address: Louise-Dumont-Str. 5, 40211
Düsseldorf
Hard Data is often User Declared
!34
Declared Data
● Name: Thorsten Feldhege
● Gender: Male
● Age: 40-45
● Email: thorsten.feldhege@adform.com
● Address: Louise-Dumont-Str. 5, 40211
Düsseldorf
Do Not underestimate Yourself
!35
Every individual lives in statistical patterns...
!36
...which can be used to build Personas
!37
Every Individual lives in Statistical Patterns
!38
Conclusion: Human - Machine Collaboration
!39
!40
Appendix
!41
The Platform Matters
The Platform Matters
Each Adform product powers and informs our full ecosystem to give you the broadest, best
informed advertising platform on the market.
!43
SHARED BENEFITS
Cross-Device | Reporting | Real-Time |
Privacy & Security | Fraud Detection |
Machine Learning Platform |
Measurement & Tracking | Audience & ID
Management | User Experience
Data Management
Platform
Demand Side
Platform
Ad Server
Creative Platform
CREATIVE PLATFORM: DYNAMIC ADS
& 

ADFORM SIGNALS
!44
Dynamic Ads
!45
Active
Lifestyle
Family
Segment
Book now Book now Book now
Book now Book now Book now
Dynamic Ads 2.0
!46
Efficiency & automation
• Runs across multiple formats
• Integrated directly into Ad
Server workflow
• Efficient spreadsheet workflow
• Powerful live preview
• Flexible reporting
Data signal activation
• Date & time
• Geo
• Site-tracking
• DMP data (1st, 3rd party, Ad
Server, combined)
• Adform Signals
Optimization
• Auto-optimization based on
powerful Bayesian math and ML
• Lower operations cost 

(adops and creative production)

• Faster campaign launch
(from initiation to campaign launch)

• Saving from 60% to 85% 

on production and adops budgets
• More relevance 

(that results in engagement and
conversion)

• Higher ROI

(more relevant and engaging)
• Higher ROI and lower
operations cost

(finding most relevant message)
User Journey
!47
Adform Signals. Supported Macros.
Creative
• Creative (banner) dimensions in a form WxH
(__ADFDIM__)
• Creative shell name
!48
Trading and account structure
• Placement (Line Item) Name or ID
• RTB deal or package ID
• RTB domain
• Mobile application name
• Domain / URL (where impression was shown)
(__ADFBNRDOM__)
• External Publisher ID (__ADFPUBID__)
• Inventory Source Name or ID
• Mobile app bundle id (__ADFAPPID__)
• Media Name or ID
• Order Name or ID
• Campaign Name or ID
Use Case Example: Versions Determined on Shell Dimensions
!49
thorsten.feldhege@adform.com
andreas.grasel@adform.com

Andreas Grasel & Thorsten Feldhege (Adform) Wir haben eine DMP - Was nun? - ProgrammatiCon 2018

  • 1.
    Do you havea Data Strategy? Andreas Grasel Thorsten Feldhege ProgrammatiCon 11.10.2018
  • 2.
  • 3.
    Adform is theLeading European AdTech Provider !3 Gustav Mellentin CEO & Co-Founder Jakob Bak CTO & Co-Founder Stefan Juricic CPO & Co-Founder Founded in Denmark in 2002 by 3 code-writing entrepreneurs, Adform is the industry’s only enterprise full-stack platform built from the ground up as one integrated advertising platform.
  • 4.
    Founded and Focusedin Europe with a Global Footprint !4 880+ Employees 19 Countries 360 Developers 250 Client Services 8 Data Points of Presence APAC Activation Hub Global Competency Center & EMEA Hub Americas Activation Hub
  • 5.
    Adform is UniquelyPositioned !5
  • 6.
    Adform provides Full-Stackcapabilities CROSS- DEVICE /
 CROSS-CHANNEL MobileDesktop Video DOOHPrint TV Demand Side Platform (DSP)Data Management Platform (DMP) Ad Server (TPAS) Creative Site Platform (CSP) Audio Native Rich Media – Site tracking - HTML5 Studio - Dynamics / Rotators - Dynamic Creative Optimization (DCO) - Creative Library Direct (AG/ IO) - RTB - Marketplace – PMP - Real-Time Forecasting - CPC, CPM, CPA, ROAS, vCPM Audience Builder - Audience Extension - High-Reach 3rd Party Data - Taxonomy Builder Integrations - API Intuitive UI → Simple workflows across platforms DATA CREATIVITY TRADING Cross-Device - Client-Side Encryption of Identifiers (PII) - Probabilistic + Deterministic Data Social CROSS-DEVICE DATA ATTRIBUTIONTRACKINGSUPPLY + many more + many more →
  • 7.
    Adform takes careof our Customers Data • Fully compliant with international privacy laws • Data is never collected without user consent • Only non-PII information for probabilistic cookie matching • All profiles based on Cookies & Device IDs • All user-relevant data is one-way encrypted before stored on Adform system • Annual Deloitte audits and security penetration tests Retain protection, control and management of all 1st, 2nd and 3rd party data as well as ROI and ROAS associated with your campaigns from one single UI. Our data centers have the most stringent security processes. Protection Ownership Control Integrations !7
  • 8.
    … with deepVertical Expertise !8
  • 9.
    Who we are !9 ThorstenFeldhege Platform Solutions Consultant DACH Andreas Grasel Commercial Director AT & CH
  • 10.
    What to expectfrom this Session !10 Demystifying Data Strategy The Four Cornerstones of a Data Strategy Man versus MachineResponsibilities and Operations
  • 11.
  • 12.
    !12 DMPs segment datafrom multiple different sources, configure it, and make it available to marketers to activate in digital marketing But: DMPs are no Data Warehouses
  • 13.
    Data Strategy andGoal Definition Process of DMP Implementation !13 Preparation Setup Data Privacy Alignment • Data Security • Data Privacy • Data Storage • Data Usage Data Strategy and Goal Definition • Goal Definition • Business Case • Use Cases • Stakeholder Technology Selection • Requirements • RFI/RFP • Scoring • Contract Data Ingestion • 1st Party Data • 2nd Party Data • 3rd Party Data • APIs DMP Setup • Naming Conventions • Taxonomy • Segmentation • KPI Definition DMP Workflows • Accounts • Access Authorization • Business Processes • Campaign Setup Change Management • Organization • Responsibilities • Know-How • Workflows Run DMP Operations • Analysis • Optimization • Campaign Management • Quality Management 1 2 3 4 5 6 7 8
  • 14.
  • 15.
    The Four Cornerstonesof a Data Strategy
  • 16.
    What is aData Strategy? The four cornerstones: • Collection • Hierarchy • Management • Activation If one part of the data strategy fails, the whole project will fail! !16
  • 17.
    Collection • All channelshave to be integrated to ensure a holistic overview. • In a perfect world the customer journey is taken into consideration to show interactions between channels. • Data privacy is not be taken lightly and should be considered from the very beginning. • Data collection is the foundation of a functioning data hierarchy and the beginning of every data optimization circle! !17
  • 18.
  • 19.
    Collection - Do'sand Don'ts • Plan ahead - every bit of acquired information can be worth a lot if combined with existing data • Track every channel - a complete picture will yield better results • Start small, think big !19 • Rush tracking implementation • Hesitate to invest development resources into tracking • Leave out channels because of “collection investment”
  • 20.
    Hierarchy • DMPs donot offer a truly persistent long term data history. This task is handled by a specialized system within the enterprise environment. Namely this can be a business intelligence system and / or a CRM. • Ad-Hoc data should be handled by the DMP. • A working and connected data hierarchy is mandatory for efficient and correct data management! !20
  • 21.
    Hierarchy – InvolvedSystems !21 Customer Data POS Offline BI / CRM PLATFORM DATA MANAGEMENT PLATFORM Web &
 App Data Campaign Data Search & 
 Social Data
  • 22.
    Hierarchy - Do'sand Don'ts • Link any in-house or on premise system to online sources via an ID matching • Integrate the campaign data stream with all On-site and CRM streams within the data hierarchy • Integrate offline systems (store visits, etc.) !22 • Underestimate the complexity of combining data sources • Hesitate to build up in-house infrastructure • De-prioritize data process documentation
  • 23.
    Management • The dataprovided by the data hierarchy must be analyzed and segmented. • A liaison between business intelligence 1st party data and external DMP information (2nd and 3rd party) will yield the best results. • Data has to be available in a pre- segmented, exportable way within the enterprise owned environment and ready to use in the DMP to be able to activate it! !23
  • 24.
    Management – InvolvedSystems !24 POS Offline BI / CRM PLATFORM DATA MANAGEMENT PLATFORM Web &
 App Data Campaign Data Search & 
 Social Data 3rd Party 
 Branded Data 2nd Party Partner Data Customer Data
  • 25.
    Management - Do'sand Don'ts • Keep the complexity low and understandable • Compare and adapt segments on a regular basis • Structure the data in an easy to understand taxonomy • Think bottom up: Single attributes can be combined in the DMP • Be creative and innovative - even if an idea fails you have learned something !25 • Hoard data that is not necessary for any channel or use case • Be to granular - combining data is easy - subtracting audiences causes confusion
  • 26.
    Activation • The segmenteddata has to be used in every system that is capable of data usage to ensure a seamless communication. • Every data usage outcome has to be measurable and reportable to allow an optimization of the segmented data. • A loopback mechanism to retrospectively optimize the used data is mandatory to achieve a closed data loop. !26
  • 27.
    Activation – InvolvedSystems !27 POS Offline BI / CRM PLATFORM DATA MANAGEMENT PLATFORM Web &
 App Data Campaign Data Search & 
 Social Data Ad Server Demand Side Platform Google Adwords & Facebook Other Platforms 3rd Party 
 Branded Data 2nd Party Partner Data Data Marketplace Data Deals with Demand Side Data Monetization Data Activation Customer Data
  • 28.
    Activation - Do'sand Don'ts • Test datasets against each other • Activate data in a structured way - otherwise learnings are hard to see • Generate data about your data - documentation is king !28 • Let the knowledge be unused • Wait for the “right” moment to use the data • Overreact when the planned results do not show up immediately • Overengineer data and campaign setup
  • 29.
  • 30.
    Which Department ownsthe DMP? “Which department is the main sponsor for the data management platform strategy, rollout, and management at your organization?” !30 Source: Forrester Study, “Savvy Publishers Deploy DMP Insights Across Their Organizations” (2016)
  • 31.
    Which Department ownsthe DMP? Persistence Prediction Activation CRM / BI Platform Data Management Platform Continuous Feedback Loop !31 Business Intelligence Marketing IT • Segmented Pseudonymous Data • Non-PII Data • 2nd & 3rd Party Data • Audience Extension / Cross Device Mgt. • Own Custom Algorithms • Based on own IT Security and Privacy • CRM / Data Warehouse / CDP • PII-Data - Flexible Structures - Virtually no Restrictions
  • 32.
  • 33.
    !33 There are differentTypes of Data Derived DataDeclared Data ● Name: Thorsten Feldhege ● Gender: Male ● Age: 40-45 ● Email: thorsten.feldhege@adform.com ● Address: Louise-Dumont-Str. 5, 40211 Düsseldorf
  • 34.
    Hard Data isoften User Declared !34 Declared Data ● Name: Thorsten Feldhege ● Gender: Male ● Age: 40-45 ● Email: thorsten.feldhege@adform.com ● Address: Louise-Dumont-Str. 5, 40211 Düsseldorf
  • 35.
  • 36.
    Every individual livesin statistical patterns... !36
  • 37.
    ...which can beused to build Personas !37
  • 38.
    Every Individual livesin Statistical Patterns !38
  • 39.
    Conclusion: Human -Machine Collaboration !39
  • 40.
  • 41.
  • 42.
  • 43.
    The Platform Matters EachAdform product powers and informs our full ecosystem to give you the broadest, best informed advertising platform on the market. !43 SHARED BENEFITS Cross-Device | Reporting | Real-Time | Privacy & Security | Fraud Detection | Machine Learning Platform | Measurement & Tracking | Audience & ID Management | User Experience Data Management Platform Demand Side Platform Ad Server Creative Platform
  • 44.
    CREATIVE PLATFORM: DYNAMICADS & 
 ADFORM SIGNALS !44
  • 45.
    Dynamic Ads !45 Active Lifestyle Family Segment Book nowBook now Book now Book now Book now Book now
  • 46.
    Dynamic Ads 2.0 !46 Efficiency& automation • Runs across multiple formats • Integrated directly into Ad Server workflow • Efficient spreadsheet workflow • Powerful live preview • Flexible reporting Data signal activation • Date & time • Geo • Site-tracking • DMP data (1st, 3rd party, Ad Server, combined) • Adform Signals Optimization • Auto-optimization based on powerful Bayesian math and ML • Lower operations cost 
 (adops and creative production)
 • Faster campaign launch (from initiation to campaign launch)
 • Saving from 60% to 85% 
 on production and adops budgets • More relevance 
 (that results in engagement and conversion)
 • Higher ROI
 (more relevant and engaging) • Higher ROI and lower operations cost
 (finding most relevant message)
  • 47.
  • 48.
    Adform Signals. SupportedMacros. Creative • Creative (banner) dimensions in a form WxH (__ADFDIM__) • Creative shell name !48 Trading and account structure • Placement (Line Item) Name or ID • RTB deal or package ID • RTB domain • Mobile application name • Domain / URL (where impression was shown) (__ADFBNRDOM__) • External Publisher ID (__ADFPUBID__) • Inventory Source Name or ID • Mobile app bundle id (__ADFAPPID__) • Media Name or ID • Order Name or ID • Campaign Name or ID
  • 49.
    Use Case Example:Versions Determined on Shell Dimensions !49
  • 50.