Applied Big Data Analytics Using Listener Local Media Association
Mather Economics Firm Overview 2 Founded In 2002 in Atlanta, Georgia Extensive experience in yield management for subscription services Over $5 Billion in subscription revenue under management Data on over 50 million subscribers received weekly Analytics services: Customer yield management & lifetime value Predictive modeling & forecasting Digital and print advertising pricing and yield management Digital analysis: meter optimization, pay wall strategies, web advertising yield Big data capture, storage, preparation, visualization A/B testing 30 Employees and 10 academic affiliates
Mather Economics Media Experience (current and former) 3 Publishing Holding Companies Individual Publishers Advance (8 publications) Belo (3) BH Media (30) Cox (4) Gannett (81) GateHouse (80) Hearst (6) Journal Register (16) Landmark (4) Lee Enterprises (30) McClatchy (30) Media News Group (55) Morris Communications (12) Post Media (10) Sun Media (30) Scripps (13) Swift (9) Tribune (6) Many others Wall Street Journal New York Times Columbus Dispatch Houston Chronicle Little Rock Democrat Gazette Chicago Sun-Times New York Daily News Minneapolis Star Tribune Orange County Register San Diego Union Tribune Tampa Bay Times Toronto Star Many others.
Mather Digital Services Overview 4
Mather s Point of View: Data, Strategy 5 First comes the data Print Advertising Audience Digital Strategic revenue model Diversified Balanced Advertising Audience Decisions keyed by analytics Copyright 2015 Mather Economics LLC. All rights reserved.
Data is Oil Publishers Need Gasoline 6 Data is the new oil it s very valuable to the companies that have it, but only after it has been mined and processed. The analogy makes some sense, but it ignores the fact that people and companies don t have the means to collect the data they need or the ability to process it once they have it. A lot of us just need gasoline. Derrick Harris, Gigaom.com, March 4, 2015 Copyright 2015 Mather Economics LLC. All rights reserved.
Listener captures, prepares, and analyzes data for implementation of integrated analytics across the enterprise 7 DFP, OpenX, Rubicon, AdX Acxiom Claritas Census Blue Kai Demo/Seg ments NewsCycle Circulation System Digital Ad Server Email Campaigns Exact Target Registration System MG2 Burt Desktop ios / Android Mobile Web E-edition (Olive) Preprint Inserts/FSI Digital Engagement Data Analytics Content Mgmt. System OpenCMS Escenic Polopoly Ad Targeting Krux Cxense Lotame Centralized Big Data & Analytics: Audience Advertising Content Video Player CNN Storyful A/B Testing Paywall / Payments Direct Mail Big Door Digital Direct Piano Syncronex MPP Tinypass Copyright 2015 Mather Economics LLC. All rights reserved.
Mather Listener TM combines data capture, storage, preparation, visualization, analytics, implementation, and reporting 8 Data capture JavaScript and HTML 5.0 tags placed within website and other operational systems to capture data on digital activities Data storage Hadoop stack within the Mather data center; SQL server instances used for visualization layer/dashboard Data preparation Data combined from multiple online and offline systems to provide a single view of the customer through an entity resolution merging process Data visualization Mather has developed its own custom dashboard to offer a singlesource of insights on all parts of the operation: audience/traffic, advertising, and content Analytics Mather has a team of 30 economists and analysts and technicians that work with the data and develop predictive models and recommendations Implementation Recommendations for tactical implementation of the data-driven strategy are sent to each operational system; such as meter levels, targeted pricing by account, acquisition lists, and advertising pricing Reporting Mather provides weekly performance reporting using A/B results and other metrics to identify successful strategies and tactics Copyright 2015 Mather Economics LLC. All rights reserved.
Mather Digital Opportunity Overview/Case Studies 9
Optimizing performance across the organization requires a holistic understanding of revenue and customer segments: Selling audience 10 DFP, OpenX, Rubicon, AdX Acxiom Claritas Census Blue Kai Demo/Seg ments NewsCycle Circulation System Digital Ad Server Based on the audience and advertising value, content can be set to monetize via paywall or advertising by adjusting the meter or % paid content Desktop ios / Android Mobile Web E-edition (Olive) Digital Engagement Data Analytics Content Mgmt. System OpenCMS Escenic Polopoly Copyright 2015 Mather Economics LLC. All rights reserved. Ad Targeting Krux Cxense Lotame Centralized Big Data & Analytics: Audience Advertising Content A/B Testing Big Door Online and offline data can be linked with advertising DMP s to refine and properly value inventory Paywall / Payments Piano Syncronex MPP Tinypass
Optimizing performance across the organization requires a holistic understanding of revenue and customer segments: Programming/Product 11 DFP, OpenX, Rubicon, AdX Acxiom Claritas Census Blue Kai Demo/Seg ments NewsCycle Circulation System Digital Ad Server Online user behavior (engagement) is combined with revenue metrics to determine the value of each piece of content and how users are segmented by content type/section Desktop ios / Android Mobile Web E-edition (Olive) Editorial management can program based on data collected and be more efficient about timing of publishing based on analysis Digital Engagement Content Mgmt. System OpenCMS Escenic Polopoly Copyright 2015 Mather Economics LLC. All rights reserved. Ad Targeting Krux Cxense Lotame Centralized Big Data & Analytics: Audience Advertising Content A/B Testing Big Door Ad sales management can be more opportunistic about pricing inventory based on value/demand. DTC products also more efficiently promoted Paywall / Payments Piano Syncronex MPP Tinypass The meter is set differently for different types of customers based on the value they bring to the company Data Analytics
Optimizing performance across the organization requires a holistic understanding of revenue and customer segments: Targeted offers 12 NewsCycle Acxiom Claritas Census Blue Kai Burt Demo/Seg ments Preprint Inserts/FSI Circulation System Audience data is captured and combined with demographics, advertising revenue, and cost data Centralized Big Data & Analytics: Audience Advertising Content Email Campaigns Xact Target Targeted acquisition offers by household are sent to sales channels and systems Direct Mail Digital Direct Data Analytics Olive E-Edition Data is prepared & analyzed. Customer segments are determined, advertising revenue by category forecasted, conversion rates by segment are estimated Copyright 2015 Mather Economics LLC. All rights reserved.
Optimizing performance across the organization requires a holistic understanding of revenue and customer segments: Paywall example 13 DFP, OpenX, Rubicon, AdX Acxiom Claritas Census Blue Kai Demo/Seg ments NewsCycle Circulation System Digital Ad Server Data is prepared & analyzed. Customer segments are determined, advertising revenue by category forecasted, conversion rates by segment are estimated Desktop ios / Android Mobile Web E-edition (Olive) Digital Engagement Digital audience and advertising data is captured using Listener JavaScript tags or other tagging system Ad Targeting Krux Cxense Lotame Centralized Big Data & Analytics: Audience Advertising Content Paywall / Payments The revenuemaximizing meter setting by segment and content category is delivered to meter Piano Syncronex MPP Tinypass Data Analytics Copyright 2015 Mather Economics LLC. All rights reserved.
Optimizing performance across the organization requires a holistic understanding of revenue and customer segments: Renewal Pricing 14 DFP, OpenX, Rubicon. AdX Acxiom Claritas Census Blue Kai Demo/Seg ments NewsCycle Circulation System Digital Ad Server Recommended renewal rate codes by account are sent to the circulation system with A/B tests incorporated Burt Desktop ios / Android Mobile Web E-edition (Olive) Preprint Inserts/FSI Digital Engagement OpenCMS Escenic Polopoly Data Analytics Content Mgmt. System Centralized Big Data & Analytics: Audience Advertising Content Accounting System Long-run optimal pricing for each account is determined based on operating margins and price elasticity Data on customers is collected from the circulation system and combined with demographics, variable direct costs, online behavior, e-edition engagement, and advertising revenue Copyright 2015 Mather Economics LLC. All rights reserved.
Finding and using the Data Truth by starting with the end in mind 15 Digital Customer Lifetime Value (CLV) Targeted acquisition price points, acquisition campaigns, content bundling, renewal price points, etc. Expected revenue streams generated by subscriber retention and monetization (advertising and subscription revenues) Dynamic meter recommendations Determine when content should be monetized based on potential advertising at risk vs. potential subscriber revenue Targeted recommendations by site section, seasonality, geography, platforms, etc. Dynamic advertising pricing recommendations Pricing based on user demand, user characteristics and behavior, etc. Pricing based on advertiser demand, line of business, and inventory levels, etc. Digital inventory forecasting Optimize programming and product offerings How often and when should new articles be published to maximize lift in traffic and advertising or maximize total time on site Which content should be packaged together as products to sell directly? Copyright 2014 Mather Economics LLC. All rights reserved.
Listener Development Pipeline 16
Existing digital data capture by different systems creates silos 17 Copyright 2014 Mather Economics LLC. All rights reserved.
Matching disparate data limits intelligence and comprehensive view 18 This process presents challenges due to lack of access to data systems and vendors in a timely manner There can also be issues with crossreferencing the data due to post processing by the data provider Copyright 2014 Mather Economics LLC. All rights reserved.
Comprehensive data gathering with Listener intercepts data across online and offline vendors and provides its own tracking function 19 Listener Copyright 2014 Mather Economics LLC. All rights reserved. 1
Comprehensive data gathering with Listener 20 Using Listener, all tracking data is collected, crossreferenced and processed at the same time This eliminates latency, the risk of losing data and any impact from data processing by the service provider Listener Copyright 2014 Mather Economics LLC. All rights reserved.
How does Listener work? 21 Listener code is installed on your webserver similar to Google Analytics or Omniture Listener runs parallel without impact to any other web analytics tracking system Listener gathers data across all services at the same time and provides its own tracking function Website performance is not impacted Offline entity resolution further brings together all systems used by the publisher Copyright 2014 Mather Economics LLC. All rights reserved.
Listener technology infrastructure diagram 22 Amazon Web Services Mather Data Center Hadoop Cluster CloudFront S-3 Buckets a1 a3 Elastic Load EC-3 Instances Collector a2 a4 Internet Collector Data Data Visualization EC-3 Instance INDS Web Service Data Ingestion Revenue Customer Server FTP Server Offline Copyright 2014 Mather Economics LLC. All rights reserved.
Online Dashboard 23
Single log-in for comprehensive online data and reports 24 Copyright 2014 Mather Economics LLC. All rights reserved.
Online dashboard Overview showing conversion funnel, site traffic and advertising metrics in one location with data filtering and exporting 25 Copyright 2014 Mather Economics LLC. All rights reserved.
Online dashboard Advertising metrics & revenue; data can be filtered using 26 multiple criteria; Impressions, click-through, and revenue shown together Copyright 2014 Mather Economics LLC. All rights reserved.
Online dashboard Traffic metrics, time on site, referral domains, unique visitors; data can be filtered using multiple criteria and exported to Excel 27 Copyright 2014 Mather Economics LLC. All rights reserved.
Collaboration and Integration 28
A sample of some partners and systems that can be included with Listener 29 Copyright 2013 Mather Economics LLC. All rights reserved.
Appendix 30
ROI Case Study Digital Engagement s Effects on Price Changes 31
Retention for Subscribers that access digital content weekly exceeds that of occasional readers and those accounts that have not registered digitally 32 Copyright 2014 Mather Economics LLC. All rights reserved.
Following pricing change Registered subscribers had 1.4% incremental stops 33 & Unregistered subscribers had 2.5% incremental stops The Registered subscribers that had received a price change had lower churn than Unregistered subscribers that had not received a price change. Copyright 2014 Mather Economics LLC. All rights reserved.
Following pricing change Subscribers with weekly digital use had an incremental stop rate of 0.8%; Notice the targets retained at 90% 34 Copyright 2014 Mather Economics LLC. All rights reserved.
Additional revenue potential from digital engagement Reduced price elasticity provides opportunity from additional revenue from existing subs 35 These data show the different price elasticity observed from targeted pricing across groups. The digitally engaged users have a 71% lower price elasticity than the unregistered subscribers If all groups received the same increase, the incremental revenue from digitally engaged subscribers would be 18% higher than the unregistered group and 5.2% greater than less engaged digital subscribers. If the increases were adjusted to balance incremental stops across groups, the digitally engaged subscribers would yield almost 3X the net incremental revenue from the pricing change. Listener data enables Publishers to measure engagement by Subscriber. Incorporating this data into subscription pricing decisions will increase the net incremental revenue yielded from these price changes on targeted segments up to 17%. Copyright 2014 Mather Economics LLC. All rights reserved.
ROI Case Study Using Content Value/Audience Behavior to Build Revenue Through Programing/Products 36
Case Study: Publisher Facing Challenges with Growing Digital Revenue 37 Pre-Mather Strategy: Digital paywall set at five PV/mo. Digital Subscription = bundle Result: Ad sales numbers continue to decline Little Digital subscription penetration Mather s Assignment: Combine digital metrics (traffic, advertising sales, subscription conversion) to help build digital revenue yield Copyright 2015 Mather Economics LLC. All rights reserved.
Case Study: Mather Focuses on Value of Customer, Each Piece of Content 38 Digital Engagement Key Discoveries: A majority of traffic comes from outside newspaper s home market Certain sections (sports, news) create much heavier engagement than others (entertainment, business) Ad Sales Key Discovery: Out-of-Market ad sales have very low CPMs Copyright 2015 Mather Economics LLC. All rights reserved.
Listener data enables Publishers to identify customer segments; Here two customer segments are profiled using anchor content and other topics they read too 39 Unique Visitors 25,000 20,000 15,000 10,000 5,000 0 news_local suburbs news news_nationworld business_breaking news_opinion entertainment news_local_politics sports_football news_columnists business sports_breaking entertainment_dining sports_baseball_cubs sports_college classified sports These subscribers primarily read football coverage. Other topics are far smaller for this group. Obvious target for sports/football product Unique Visitors 16,000 14,000 12,000 10,000 Anchor 8,000 6,000 4,000 2,000 0 These subscribers primarily read local news; Other topics are shown on chart. Revenue associated with this traffic can be added to the analysis sports_football sports_breaking news_local sports_baseball_cubs sports_rosenblog sports_basketball sports_smackblog suburbs news_nationworld news business_breaking sports_columnists sports sports_college sports_baseball_whitesox news_opinion entertainment Anchor Copyright 2013 Mather Economics LLC. All rights reserved.
Case Study: Digital Strategy Focused on Audience Value 40 Based on audience value: Two strategies Local CPMs high very loose meter created Out-of-market CPMs low traffic funneled into topic-specific subscription products Local audience Monetized through high-value impressions Most-engaged users eventually funneled into appropriate paywall products Out-of-Market audience Pushed quickly into pay products Low risk of lost advertising revenue Copyright 2015 Mather Economics LLC. All rights reserved.
Four P s: Case Study Actionable Recommendations 41 Programming Focus timing of content production and publication around peak traffic periods Implement targeted ad pricing policies that place premium on heavy audience value periods Products Build products that resonate with high-value audiences Engagement Pricing Expanded audience from multiple subscription options Print-to-digital engagement provides opportunity Promotion Relevant messaging to distinct audiences One-size-fits all a mistake in digital Copyright 2015 Mather Economics LLC. All rights reserved.
Listener Data supports digital Customer Lifetime Value (CLV) A sample analysis is presented below showing customer CLV from content audiences 42 All content to the left of sports attracts current subscribers with a high customer lifetime value CLV Score $1,400 $1,200 $1,000 $800 $600 $400 Content categories with values above the yellow line (average) attract valuable subscribers Lifestyles:Home is the median point of total page views 100% 90% 80% 70% 60% 50% 40% 30% Page View Proportion $200 20% 10% $0 0% CLV C.Page View Proportion Copyright 2014 Mather Economics LLC. All rights reserved.
Contact information 43 Matt Lindsay President Mather Economics LLC 1215 Hightower Trail Building A, Suite 100 Atlanta, GA 30350 770-993-4111 main 678-585-4101 direct 404-395-4937 mobile matt@mathereconomics.com www.mathereconomics.com Copyright 2015 Mather Economics LLC. All rights reserved.