LEOtrace Mobile DMI Academy New York 12 06

LEOtrace® Mobile DMI Academy New York, 12. 06. 2013 Torsten Müller © Gf. K 2012 | LEOtrace® Mobile | March 2013 1

LEOtrace® Mobile stands for the highest standards in terms of quality, privacy, security, integrity and compliance. User-Centered Metering for Smartphones and Tablets LEOtrace® Mobile is an extension of Gf. K nurago‘s LEOtrace® technology. A wide range of measured mobile insights is constantly extended to cover more device types and mobile operating systems. Data is collected directly, … in a non-reactive, unobtrusive way. … independent of network operators, device manufacturers*, application and web service providers. … enriched by attitudinal insights using event- and location-triggered surveys. * Single data points can slightly differ or nor be available for particular devices. © Gf. K 2012 | LEOtrace® Mobile | March 2013 2

Use Cases © Gf. K 2012 | LEOtrace® Mobile | March 2013 3

Audience Research Why? Credible measures of reach and frequency to enable advertisers to plan brand advertising across media and devices. Mobile Usage is a significant part of the global media usage and is getting even more important. Our consumercentric hybrid audience measurement system deliver a comprehensive measure of all digital activity, across devices and across locations. What? Examples Recommendations for strategic media planning and allocation of media budgets; channel selection based on cross-media GRP analysis. Relevant Mobile Metrics: § Socio demographics § URL usage § App usage § Device types Part of Gf. K’s cross-device and crossmedia capabilities. © Gf. K 2012 | LEOtrace® Mobile | March 2013 4

Consumer Journeys Why? Customers use their smartphones for in-store product research and browsing or tablets for online shopping. With consumer journeys we deliver insights into the behavior a target market displays when investigating and progressing through the product research and purchase process as well as insights into expectation, motivations and attitudes. With LEOtrace® it’s also possible to understand cross-device interactions between desktop computers/laptops, tablets and smartphones. © Gf. K 2012 | LEOtrace® Mobile | March 2013 What? Examples Comprehensive Consumer Journeys for smaller or larger sample sizes for specific products, industries, etc. Relevant Mobile Metrics: § Socio demographics § URL usage § App usage § Search requests § Locations § Device types § Event-triggered questionnaires 5

Event-Triggered Surveys and Diaries Why? Mobility of devices establish new opportunities to get insights about consumers in relevant moments. Surveying participants when entering a specific geographical area, using a particular app or right after searching for a product on Google. Getting insights by diary approaches and linking short surveys on a frequent basis to location information or device usage. © Gf. K 2012 | LEOtrace® Mobile | March 2013 What? Examples Quick surveys which can be filled out on the mobile device online or offline and which are triggered by time, location or device usage. Relevant Mobile Metrics: § Socio demographics § URL usage § App usage § Search requests § Locations § Device types § Event-triggered questionnaires 6

Location Based Insights Why? Mobile devices are literally mobile. Insights about the movements of your consumers helps you to adjust out of home advertising, insights about shopping or lifestyle behavior. Understanding where specific target groups go and how their behavior in using their smartphones changes based on their current positions. What? Examples Combination of geo location information and consumer insights. Relevant Mobile Metrics: § Socio demographics § Locations § Event-triggered questionnaires § Device Usage Furthermore it’s a key insight to know what people are looking up in the internet while being at a POS. © Gf. K 2012 | LEOtrace® Mobile | March 2013 7

User Experience Why? Conversion rates on m. Commerce apps or the success of devices , apps or mobile websites is always dependent from a good user experience. Device-orientated measurement of usage and habits combined with insights in expectations, evaluations and needs of users helps identifying pain points and room for improvement. © Gf. K 2012 | LEOtrace® Mobile | March 2013 What? Examples Quick surveys which can be filled out on the mobile device online or offline and which are triggered by time, location or device usage. Relevant Mobile Metrics: § Socio demographics § Device usage (e. g. Calls, Apps, Data Traffic, Browser usage, etc. ) § Locations § Information about the device § Event-triggered questionnaires 8

LEOtrace ® for Mobile Capabilities © Gf. K 2012 | LEOtrace® Mobile | March 2013 9

Scope of our Mobile Measurement Smartphones and Tablets Android 2. 1+; support of Tablets, Smartphones, Mini Smartphones, Kindle Fire and Nook Tablets. i. OS 5; support of i. Phones and i. Pads. further OS Support for Symbian, RIM (Blackberry) and Windows Mobile has been discontinued due to low market shares. Upcoming platforms are under consideration (e. g. Windows Phone). © Gf. K 2012 | LEOtrace® Mobile | March 2013 10

Getting participants easily involved Easily downloading and installing the app for smartphones and tablets 1 3 2 Recruitment Email or text message to participants with a link for opt-in and download* The app runs in background, logs mobile metrics, only a slight impact on battery drain * Download via Google Play Store (Android) or direct download from Gf. K servers (Android + i. OS); Opt-In © Gf. K 2012 | LEOtrace® Mobile | March 2013 11

Insight Tables Who? What? Where and when? Structural Information Apps & Browser Time Device • Socio demographics from the respective panel • Application Usage • Browser URL Request • Browser Search Request (coming soon) • Local Time • UTC time • Durations for App and URL usage. • Distinguishing Smartphones and Tablets • Model • Manufacturer • Operating System Communication Location • Call • Longitude/Latitude • Accuracy and source Attitudinal Insights • Possibility to survey people on the mobile device triggered by time. Triggers by usage or location are coming soon. How? Device usage • Data Traffic © Gf. K 2012 | LEOtrace® Mobile | March 2013 12

Compatibility of Insights & Metrics Metering Data Application Usage Browser Search Request (coming soon) Browser URL Requests Location (Optional) Device State Change (Optional) Call Offline Surveys Data Traffic App-Centric Data Metrics In-App Measurement for Client’s Apps Durations Browser Search Requests for i. OS are under development. © Gf. K 2012 | LEOtrace® Mobile | March 2013 13

LEOtrace® Mobile Privacy & Transparency Opt-In & Transparency – every respondent is aware of data collection and research purposes. No „hiding“ of the actual functionality of the app. Data protection & Encryption – All data is encrypted on the device and transferred via HTTPS connections to our secured servers. Data economy – We only track what is necessary by configurable data points. Aggregation & Anonymizing data – We won‘t share PII with clients. Stopping data collection by deleting the app. © Gf. K 2012 | LEOtrace® Mobile | March 2013 February 14 2013

Known Limitations General challenges of Mobile Measurement § § § lack of standards technological challenges complex ecosystem ambiguous regulatory environment privacy features in mobile operating systems. This causes some limitations in term of mobile measurement you might know. © Gf. K 2012 | LEOtrace® Mobile | March 2013 Current limitations for passive metering approaches: § § § For Android it’s only possible to measure standard browsers. SSL websites are currently not allowed to be measured on i. OS. It’s not possible to manipulate contents displayed on the device. There is no authentication possibility to measure multi-user behavior on tablets. In-App behavior is only measurable for client’s apps by integrating a measurement component. We are continuously testing upcoming versions of operating systems and new devices to extend passive metering capabilities. 15

„ Competitive Environment „…it is apparent that nurago has greatly improved its metering services on almost every front. ” (Gf. K MRI after testing LEOtrace® Mobile) Competior Intelligence: Based on strong relationships with our clients we have good insights into our competitor’s offerings. Even though the solutions are technically on a similar level there are important differences: § We do i. OS and hold an agreement with Apple § Our user experience for participants is better § We offer Mobile Metering as a cross-platform (computers, tablets, smartphones) and crossmedia (TV, magazines, etc. ) solution within Gf. K. © Gf. K 2012 | LEOtrace® Mobile | March 2013 Benchmark: Within a comprehensive test Gf. K MRI benchmarked Gf. K nurago’s mobile metering solution for Android and i. OS against four competitors. Within all dimensions of the test Gf. K nurago’s LEOtrace® mobile meter was capable of competing and performed overall as the most reliable solution. 16

Data & Project Examples © Gf. K 2012 | LEOtrace® Mobile | March 2013 17

Reach of Apps and Mobile Websites in Germany Mobile Apps Mobile Websites Net Reach Duration in sec Ø Whats. App Messenger 67% Google Search 55% facebook 55% Gmail You. Tube 26% wetter. com 24% DB Navigator 12% 409 465 270 137 308 110 52% facebook 39% amazon 163 15% 70% wikipedia 107 36% e. Bay WEB. DE Mail 497 45% Duration in sec Ø Google Search 258 59% Google Maps Net Reach 31% You. Tube e. Bay 27% 25% gutefrage. net 18% bild. de 17% focus. de 14% accuweather. com 14% 36 74 55 68 40 108 35 69 43 77 Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user n=844 (app user n=840, internet via browser user=758) © Gf. K 2012 | LEOtrace® Mobile | March 2013 18

App vs. Mobile Browser n=546 n=395 n=306 58% 28% Via Browser Only 15% 27% App only Via Browser Only 24% 23% 48% App only Via Browser Only 56% 21% App only Interesting fact: 8% of the app user did not use internet via browser. Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user n=844 © Gf. K 2012 | LEOtrace® Mobile | March 2013 19

Top 10 of App Categories Duration of daily app-usage in sec Ø Written Communication e. g. Angry Birds, Sudoku Free, Solitaire. 1499 Gaming 375 Social 279 162 Self Organization Shopping 105 Location & Maps 71 Offline Music & Video 65 News 60 Tools, Pics & Videos 58 Personalization 55 e. g. Facebook, Twitter, XING e. g. Facebook Messenger, Whats. App, Gmail Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user n=844 © Gf. K 2012 | LEOtrace® Mobile | March 2013 20

Mobile vs. Desktop Usage Google Maps is more likely to be used on mobile devices. e. Commerce is still more popular than m. Commerce. Net Reach Mobile Google Search Net Reach Desktop 72% 79% Whats. App Messenger facebook Google Maps 67% 64% 68% 61% You. Tube 45% e. Bay 38% wikipedia 38% amazon 37% Google Mail 18% 58% youtube. com 58% ebay. de/com wikipedia. org 44% 65% 28% 12% facebook. com Google Maps 37% 45% wetter. com whatsapp. com 1% Gmail Google Search amazon. de/com wetter. com Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user x Desktop user (N=666) © Gf. K 2012 | LEOtrace® Mobile | March 2013 21

Mobile Usage – Usage over the day 2 1. 5 Mobile Usage 1 Desktop Usage 0. 5 59 02 : 0 0 - 03 : 59 00 : 0 0 - 01 : 59 22 : 0 0 - 23 : 59 20 : 0 0 - 21 : 59 18 : 0 0 - 19 : 59 16 : 0 0 - 17 : 59 14 : 0 0 - 15 : 59 12 : 0 0 - 13 : 59 10 : 0 0 - 11 : 59 08 : 0 0 - 09 : 59 07 : 0: 0 06 : 0 0 - 05 : 59 0 04 Index (daily average = 1) based on duration The mobile internet is constantly used during the day while the fixed internet is used during the day with special raises in the late evening and strong decrease at night. Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user n=844, Mobile Android user with usage via desktop PC n=666 © Gf. K 2012 | LEOtrace® Mobile | March 2013 22

Incremental Reach via Mobile 94% 86% 81% 76% 29% 65% 30% 55% 20% 62% 18% Mobile 26% 27% 49% 65% 55% 27% 62% 55% 29% Desktop and Mobile 29% 37% Desktop 16% Source: Media Efficiency Panel 11/2012, Mobile external app/browser Android user n=844, Mobile Android user with usage via desktop PC n=666 © Gf. K 2012 | LEOtrace® Mobile | March 2013 24

Mobile Insights on Individual Level – This is Arne! Attitudinal Insights 33 years old Frequent Flyer Interests (IT) news, music, playstation, soccer, social media Miles & More programs, “preferred guest” in hotels Griesheim, Germany middle school education Ad agency in Mainz Adventurer young, dynamic, trendsetter, seeking for success and autonomy, Consumption orientated on innovation and leisure © Gf. K 2012 | LEOtrace® Mobile | March 2013 Gastronomic wine gourmet, french restaurants in a relationship with a 30 year-old woman Windows 7 and Firefox 8 Samsung Galaxy II (Android) 25

a we eken d in Fr ankfu rt plans 9: 00 am – 10: 30 am: quality time: Google Mail, heise. de, golem. de, vielfliegertreff. de, starwoodhotels. de, banking. sparda. de, play 3. de 7: 25 am: waking up: bild. de, Google Mail, facebook. com 11: 47 am: discussion with colleagues gamescom. com, facebook. de 5: 50 pm: break bild. de, sport news on n-tv. de and ndr. de, facebook. com 21: 03 pm: extended break haushelden. de (looking for a gardener), facebook. com, Youtube, Google+, Twitter, groupon. de, Restaurant finder Due to data privacy issues we displaced residence and workplace © Gf. K 2012 | LEOtrace® Mobile | March 2013 26

Wi. Fi versus 3 G/4 G Data Volume Preliminary Findings Average Monthly Usage Per Student Median Wi-Fi 3 G/4 G Total Usage 7. 60 GB 5. 67 GB 13. 27 GB 1. 18 GB 0. 49 GB 1. 83 GB Distribution of Monthly Usage 30% Less than 1 GB 14% 1 GB to less than 1. 5 GB 8% 1. 5 GB to less than 2 GB 16% 2 GB to less than 5 GB 21% 10% 5 GB to less than 25 GB or more BASE: Total (n = 107) Confidential and proprietary material for authorized Verizon Wireless personnel only. Use, disclosure or distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 27

Wi. Fi versus 3 G/4 G Data Volume By Day of Week Preliminary Findings based on data volume 100% Wi. Fi Usage 53% 57% 60% 56% 3 G/4 G 58% 64% Average/week Mix of Wi. Fi, 3 G/4 G usage 50% 47% 43% 40% 44% 42% 36% tu y da Su n ay rd ay id Fr Sa Th ur s da y ay sd ed ne es d ay W School Week Tu M on da y 0% Weekend BASE: Total (n = 107) Confidential and proprietary material for authorized Verizon Wireless personnel only. Use, disclosure or distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 28

Wi. Fi versus 3 G/4 G Data Volume By Time of Day Preliminary Findings based on data volume Index (daily average = 1) 2. 5 2 1. 5 Wi. Fi Usage 1 3 G/4 G Usage 0. 5 9 3 : 5 0 to 2: 0 to 00 AM 12 : 00 to 1 1 1: : 5 9 59 9 10 : 0 to 8: 0 7 : 5 0 to 9 : 5 9 9 6: 0 0 to 5 : 5 9 4: 0 3 : 5 0 to 2: 0 to 00 PM 12 : 00 to 1 1 1: : 5 9 59 9 10 : 8: 0 0 to 9 : 5 9 7 : 5 0 to 6: 0 AM 4: 0 0 to 5 : 5 9 0 BASE: Total (n = 107) Confidential and proprietary material for authorized Verizon Wireless personnel only. Use, disclosure or distribution of this material is not permitted to any unauthorized persons or third parties except by written agreement. 29

Case Story Mobile Project Switzerland Case Story Mobileproject Switzerland 30

Campaign Background 31

Product push in Switzerland Our Client is a publishing house with a high number of various media brands. Our task was to measure the ad efficacy of a campaign for a coffee drink competing with several long -established companies on the Swiss beverage market. 60% 13 websites 11/02/2013 11/03/2013 40% 2 22 Banner apps 2 Pre-Interstitial languages 32

Measurement & Recruitment 33

Measurement – In App The publisher integrated a tracking component (SDK) into his apps. SDK APP With this SDK it was possible to measure inapp ad contacts for all app users. 34

Measurement – MEW (Mobile Enabled Web) via Cookie & Tag Panelists (recruited via Gf. K’s Online Access Panel) were equipped with a cookie by visiting an opt-in page on their mobile device. Tagged Ad The publisher integrated „tracking tags“ into his ads and websites to measure ad contacts. 35

Recruitment For matching the data we got from in-app measurement with the data of the cookie panel we asked our participants to provide their device ID. Panellists were able to report the ID in a questionnaire or to download an app which automatically transfers this information to us. Download Gf. K Mobile. Monitor App @ § Basic Survey/ Pretest WITH Gf. K MOBILE MONITOR Recruitment Email Registration Mobile Cookie Panel Transfer Device ID WITHOUT Gf. K MOBILE MONITOR Asking for Device ID Transfer Device ID 36

Conclusions 37

Effective in driving Brand Awareness +31% Client Mobile is an effective platform to launch new brands and build brand awareness +2% Competitor 1 0% Competitor 2 +2% Competitor 3 38

Frequency matters +20% 1 -2 Contacts More exposures drove higher uplifts in brand awareness +31% 3 -5 Contacts +44% 6+ Contacts 39

Conclusions Ad efficacy research for mobile campaigns is as successful as common research approaches and delivers powerful insights. It’s perfect to be enriched by desktop, TV or print research data in crossmedia projects. 40

Experience with projects and panels § Media Efficiency Panel Germany – Measurement of fixed and mobile usage, TV ratings, purchase behavior, etc. Since 2011, currently approx. 800 users § Media Efficiency Panel Netherlands – Measurement of fixed and mobile usage, purchase behavior, etc. Starting Q 1/2013 , approx 1000 users in 2013. § Connected. Life. dx – Cross-platform measurement of Windows computers and Android Phones in US, UK, China and Germany. 11/201112/2012 – More than 600 Users. § Ad Efficiency Research for In-App and In-Browser ads for a large Swiss publishing house. Since 01/2013 § Out-Of-Home Campaign tracking using a location based diary approach in Germany. Starting Q 1/2013 § Qualitative tracking of political opinion leaders for Voice of America in Indonesia. 2012 - 42 Users. § Verizon US: Measurement of Android tablet usage, focus on data traffic. 20123 - 100 Users. 2 Follow-Up projects won, 2 more are likely to come © Gf. K 2012 | LEOtrace® Mobile | March 2013 41 February 2013

Questions? © Gf. K 2012 | LEOtrace® Mobile | March 2013 42

Contacts Torsten Mueller Global Head of Gf. K DMI Gf. K nurago, Hamburg Tel. +49 40 696669 150 Mobile +49 171 7562449 Email: Torsten. Mueller@gfk. com © Gf. K 2012 | LEOtrace® Mobile | March 2013 43

Appendix © Gf. K 2012 | LEOtrace® Mobile | March 2013 44

Insights in Detail © Gf. K 2012 | LEOtrace® Mobile | March 2013 45

Log Entries in Detail Net Reach in Germany Application Usage 67% Whats. App Messenger • App usage - An app changed its focus - An app is terminated incl. app name and vendor of the app 59% Google Maps Google Search 55% facebook 55% 45% Gmail 36% You. Tube e. Bay 26% wetter. com 24% WEB. DE Mail DB Navigator 15% 12% Source: Media Efficiency Panel 11/2012, Android user n=840 © Gf. K 2012 | LEOtrace® Mobile | March 2013 46

Log Entries in Detail Net Reach in Germany Browser URL request • Track all URLs* 70% Google Search 52% facebook 39% wikipedia 31% amazon Browser Search request 27% You. Tube 25% e. Bay • • Includes all search terms and Com ing s oon the search engine Uses the pre-installed browser and the linked searches gutefrage. net 18% bild. de 17% focus. de 14% accuweather. com 14% Source: Media Efficiency Panel 11/2012, Android user n=758) *On Android it’s only possible to track URLs in the default browser © Gf. K 2012 | LEOtrace® Mobile | March 2013 47

Log Entries in Detail Information on location (Optional) • Location information Ø Latitude Ø Longitude • Accuracy Location data is also added to every event Source: Test of Gf. K MRI in New York City © Gf. K 2012 | LEOtrace® Mobile | March 2013 48

Log Entries in Detail Information on calls • • • State - Initiated - Received - Terminated Action for received calls - Missed - Rejected - Taken Country code of the phone number Source: Example of a single user’s log © Gf. K 2012 | LEOtrace® Mobile | March 2013 49

Log Entries in Detail (Offline) Surveys • • Triggered from our app • By date Eve nt & • By events L are ocation com • By location ing s trigger oon The look can be adapted Surveys with one or more pages Different types of questions © Gf. K 2012 | LEOtrace® Mobile | March 2013 50

Log Entries in Detail Information on data traffic on the device • • 1. 5 M o bi le. . . 1 0. 5 0 04 : 0 0 06 -05 : 0 : 5 0 9 08 -07 : 0 : 5 0 9 10 -09 : 0 : 5 0 9 12 -11 : 0 : 5 0 9 14 -13 : 0 : 5 0 9 16 -15 : 0 : 5 0 9 18 -17 : 0 : 5 0 9 20 -19 : 0 : 5 0 9 22 -21 : 0 : 5 0 9 00 -23 : 0 : 5 0 9 02 -01 : 0 : 5 0 - 9 03 : 5 9 • Number of bytes that is sent or received Timestamps for start and end of the transfer Connection types (Wi. Fi or Cellular Connection) 2 Source: Media Efficiency Panel 11/2012, Androis users n=844. Index (daily average = 1) based on duration. © Gf. K 2012 | LEOtrace® Mobile | March 2013 51

Insight Tables in Detail Device State Change (Optional) • Device Switched On / Off • Device Screen Locked/Unlocked © Gf. K 2012 | LEOtrace® Mobile | March 2013 52

Insight Tables in Detail In-App Measurement for Client’s Apps. • • Interaction, Views of screens, advertising contacts, etc. within apps. Only available for Client’s Apps by integration of our library (SDK) © Gf. K 2012 | LEOtrace® Mobile | March 2013 53

Further Collected Data Points § Text/SMS sent § Sound is recorded and saved § Text SMS received § New ringtone installed § Song starts/stops playing § Bookmark added § Video starts/stops playing § … § Picture is taken and saved We will provide more data points in 2013. Ask us. § Video is taken and saved © Gf. K 2012 | LEOtrace® Mobile | March 2013 54

Outlook © Gf. K 2012 | LEOtrace® Mobile | March 2013 55

LEOtrace® Mobile - Roadmap for 2013 July/Aug Sep-Dec Android + i. OS and Backend Systems • • • Important Security Improvements Event and Location Based Triggers (was planned for 02/2013) Diary Approach/Manual Survey Start Validation • • Finishing Data Points Validation 1 st Round (was planned for 04/2013) Location Tracking Validation Round 2 (Walk-around with Apps + GPS Tracker Data Processing • • Mobile Data Feed – Launch of the Standard Format Flagging Relevant URLs from URL Requests Event for i. OS Concepts and Proof of Concepts • Interface for Household Based Field-Time Monitoring; Po. C Proxy for Android; User Identification for Tablets; Improvements for Scalability and Monitoring. Windows Phone/RT • Proof of Concept / Decision / Next steps Data Enrichment • Inference System to apply categorization schemes / rules to data files. i. OS and Android • • User Identification for Tablets (dependent from Po. C) Several features based on feasibility tests in the next months. © Gf. K 2012 | LEOtrace® Mobile | March 2013 56

Our Vision for 2014 Mobile is a key element in understanding media behavior at the moment. But it’s even more relevant to see the big picture: Consumer behavior is changing very quickly and the crucial point is to focus on cross-device and cross-media behavior. Gf. K nurago therefore is already focusing on cross-platform solutions rather than short-term hypes. As TV has been the biggest influencer we observe a rapid growing two screen media usage where TV drives mobile. © Gf. K 2012 | LEOtrace® Mobile | March 2013 57

LEOtrace® Mobile DMI Academy New York, 12. 06. 2013 Torsten Müller © Gf. K 2012 | LEOtrace® Mobile | March 2013 58