| October 13, 2016 |

The power of television

Capital Markets Day – October 13, 2016

DataChristof Wahl

| October 13, 2016 || October 13, 2016 |

In advertising, user data becomes a key success factor

2

From context buying … … to audience buying

Advertisers used to target their audiences based

on context of the ad …

… now, data-driven advertising enables

advertisers to target specific users (one-to-one),

independent from specific context

We leverage data from our broad Digital Entertainment, Commerce and

TV portfolio to fully exploit opportunities arising from targeted advertising

| October 13, 2016 || October 13, 2016 |

Addressable TV transfers digital targeting approach to TV …

31) Only applied to users who pressed the red button

Special

interest

national1)

Mass markets

with relevance

for larger

geographics

Geographic segments, detected via IP address

Shown after

channel switch

For national

campaigns

ALREADY AVAILABLE

Addressable TV core targeting approaches Addressable TV targeting formats (examples)

Household related interest segments based on TV usagee.g. single household with high automotive affinity

IN PREPARATION

Spot

Fullscreen

Overlay

Switch-Ins

ALREADY AVAILABLE

IN PREPARATION

Switch-In (XXL)

| October 13, 2016 || October 13, 2016 |

... and will create additional advertising opportunities in TV

4

Red Button user User tracking Segmentation Targeting

Tracking pixel

Data pool +

user

segmentation

Segment

“Family with

kids, age 6-13”

Jetzt RED BUTTON drücken

Ad overlay (Switch-In or

Spot-Overlay) with

family vacation theme

New offerings

- Incremental reach of TV ad campaigns on digital P7S1 properties

- Incremental reach of digital ad campaigns on Addressable TV layer

- Frequency capping across devices (incl. HbbTV)

Upside

- Additional revenues from existing TV customers

- Access to new advertiser segments thanks to more granular targeting

Benefits

| October 13, 2016 || October 13, 2016 |

Also in e-commerce, data significantly drives value

5

User ID 4711

Female with young kids &

preference for premium products

Pearl necklace

premium

Sport watch Family vacation

IbizaIbiza Café del Mar

Perfume top brand AftershavePearl necklace

premium

Family vacation

IbizaPerfume top brand

From top selling products for everybody … … to segment specific offers

We continue to expand our data pool to

drive personalization and thereby increase conversion rates and visits

| October 13, 2016 || October 13, 2016 |

Data type:

Behavioral data

from HbbTV usage

and in future also

registration data

Impact1):

20-40% conversion

uplift

Data will be a key value driver across our portfolio

61) Market benchmarks – indexed 2) Page Impressions/Video ViewsSource: Internal benchmarks and expert assessments

Upside potential cannot be applied to all revenue streams

as data availability and granularity is not given for all assets/units on same level

Digital ad sales:

eCPM effect

Data type:

Highly valid

sociodemographic

user data from

registrations

Commerce:

Conversion uplift

Data type:

Intent data from

P7S1 e-commerce

touch-points

Addressable TV:

Incremental revenues

Online video:

PI/VV2) uplift

Data type:

Content interest &

usage data

Impact1):

10-20% usage

uplift, e.g. for Video

Views and thus pre-

roll ad impressions

Impact1):

30-50% eCPM

increase possible

for remnant

inventory

Impact:

Incremental

revenues from new

profile based

targeting proposition

(still in test phase)

| October 13, 2016 || October 13, 2016 |

TV websites

Our vision: Leverage data potential to further address customers

7

Addressable TV

Commerce websites

Entertainment websites

INTEREST

INTENTSOCIO

DEM.

ENGAGEMENT

CONSIDER

USE

BE

CO

ME

AW

AR

EX

DEVICE

BU

Y

SELECTED EXAMPLES

APP

| October 13, 2016 || October 13, 2016 |

We have acquired a proprietary AdTech stack ...

8

Data Management Platform

Demand Side Platform Supply Side Platform Ad server

(Minority stake only)

P7S1 stakein Virtual Minds:

51%

Exchange/Marketplace

70%of German Top 20

AGOF sales houses use

Yieldlab

VM with minority stake

in ADEX

We cover the full value chain

| October 13, 2016 || October 13, 2016 |

… to enable independent & publisher friendly data-driven advertising

9

Our AdTech stack gives advertisers access to combination of

inventory and data – and prevents data leakage

Agency

Demand Side Platform Supply Side Platform Ad server

Budget AdTech Media & data

DMP

| October 13, 2016 || October 13, 2016 |

In addition, we offer with smartstream a powerful video targeting product

10

2 brand-strategy

Volume

Pri

ce

Premium

reach

Audience reach

AdTech

Advanced audience

targeting on “brand-safe”

high reach inventory

Top contents & premium

websites on top placements

Fully transparent,

AGOF-measured

| October 13, 2016 || October 13, 2016 |

Selection

Targeted

advertising

Site

personalization

Onsite recom-

mendations

Targeted

media buying

We have built a unified data pool across our key digital assets

111) To be linked with digital data pool2) Only pseudonymous data3) Estimated cross-device unique user reach (desktop web/mobile web)

Core use casesAnonymous tracking Data segmentation Financial potential

Cost per

order

Ad revenue

per user

Sales per

customer

ADDRESSABLE TV1)

Thanks to our broad asset portfolio we can leverage our e-commerce data for advertising

Data pool2) today:

~40 P7S1

websites

~30m user

profiles3)

P7S1 internal data

| October 13, 2016 || October 13, 2016 |

Case example: Targeted advertising with own DMP data works

12

Impact of data pool based targeting

Click-through rate (CTR)

Untargeted(Control group)

Targeted(Based on data pool)

User affinity to destinations (e.g. Spain,

Italy, Croatia) captured at wetter.com via

data pool in data protection compliant way

Performance display campaign for

überflieger.de based on targeting

information from P7S1 data pool

Targeted ads delivered on wetter.com

(high priority)

Test setup & insights

160%

uplift

CASE EXAMPLE

Results from pilot test case – upside potential cannot be applied to all revenue streams

as data availability and granularity are not given for all assets/units on same level

| October 13, 2016 || October 13, 2016 |

Case example: First tests show power of targeted media buying

131) Test conducted with Semasio/Adlicious 2) Post view results (30 days after end of campaign)

Look-alike display ad targeting Test setup & insights

Cost per order

Approach: Analyze conversion profiles and

buy similar profiles programmatically in

marketplace1)

Insight: Targeting decreases cost per

order by more than 20%2)

Next step: Combine look-alike media

buying with retargeting to drive CPO

further down

From different campaigns

>20% Cost

reduction

CASE EXAMPLE

Untargeted(Control group)

Targeted

Results from pilot test case – upside potential cannot be applied to all revenue streams

as data availability and granularity are not given for all assets/units on same level

| October 13, 2016 || October 13, 2016 |

We push registration via 7Pass to enable cross-device targeting

14

1) Already integrated into 7Pass product – used in 7TV as log-in mechanism

Note: Marketing usage of real identity attributes requires opt-in of user or anonymization of all Personally Identifiable Information (PII)

Applications for all devices

QR scan technology1) to link

mobile to smart TV

Cross-device logins via smartphone, tablet, web & smart TV

Key enabler:

7Pass as cross-

device Single-Sign-

On (SSO)-tool

| October 13, 2016 || October 13, 2016 |

Data pool Digital Advertising

User profiling covers now

>30M UU1) at >40 P7S1

majority assets

DMP based ad targeting

works – with own and 3rd

party data

7Pass SSO roll-out will

improve cross-device

targeting &

sociodemographic

profiling

SSO

Data pool is

used for onsite

personalization and

targeted media buying

Digital Commerce

HbbTV targeting generates

incremental TV ad

revenue streams

HbbTV

Usage of own AdTech

maximizes data control

along value chain

AdTech

Summary: We create value for digital and TV with data plus tech

151) Estimated cross-device Unique User (UU) reach (desktop web/mobile web)

| October 13, 2016 || October 13, 2016 |

Key strategic next steps to further develop impact of data

Track and use further

interest & intent data

on our Digital

Entertainment and

Commerce assets

We have a unique value proposition

in data

We can combine Commerce,

Entertainment and Addressable TV

data for our core use cases

Ad Targeting

Content and Commerce

personalization

In addition we consider to further

develop our proprietary tech stack

and build data partnerships with 3rd

parties

Illustrative strategic impact of data Summary

Further create

registrations for

cross-device in

Digital Entertainment

and Commerce;

further push and use

socio-demographic

data

Capture further

Addressable TV data

and link addressable

with digital data

(cross-device

addressable & digital)

16