Making sense & nonsense of Data: Successes & Failures Using Big Data in Magazine Marketing

Pegg Nadler, award-winning magazine marketing guru from New York spoke at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, about magazine business in US. She has made a survey, where she interviewed 15 major magazine publishers representing 250 titles in US.

"Big data is like teenage sex: everyone is talking about it, but nobody knows how to do it."

What is then Big Data? Traditional data is now being called in US as "big data", because there is so much of it and it has been collected for ages from multiple sources. It's also a combination of offline and online data.

The old data concerns are gone, because data storage is no longer an issue, data is easily attainable, data processing quicker than ever and data matching and cleansing improving.

The "true" big data means detailed data collected from websites and digital sources. Big is not just volume, but velocity and increased activity, size and scales not important, but the most important is, what you do with it.

How to use big data:
1. Start with determining the business needs and questions to be answered. What is troubling?
2. Identify the data that will get you there.
3.  Build the technology around the information to get it.

Different business needs and questions: customer buying cycle, cross channel view, customer purchase path & preferences, customer research modes, feedback behavious, next best offer.

Data assessment: What data do you have? What data do you need? What other data is available?

After that you have to identify the different data resources and start analyzing, e.g. customer segmentation.

The future is obscure and we can't predict it from the history, so we can use big data to draw the future.

Problems

How BuzzFeed is leveraging Big Data to target its readers in real time

Ky Harlin, Director of Data Research and Development at BuzzFeed, New York, focuses his work on identifying, organizing and accelerating the audience for viral content on Buzzfeed.com

In his presentation to the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, Harlin said his department measures the "social reproduction rate" of media and is a measurement metric that exists in real-time on a variety of content, from news to animal videos to food recipes and preparation.

Harlin mentioned the case of an article: "60 Thanksgiving Side Dishes To Make Absolutely Everyone Happy" (http://www.buzzfeed.com/rachelysanders/60-thanksgiving-side-dishes-to-make-absolutely-eve). "We publish pinnable holiday content a couple of weeks in advance. Then, promote on Facebook closer to the actual day," Harlin said.

Key takeaways of Harlin's presentation were:
- Data can be used to optimize content for sharing throughout the life of an article.
- Before publishing, data can help determine what to write about and how to present it.
- After publishing, data can help optimize the article's promotion as well as the article itself.

Rethinking Mobile Journalism Via Data and Technology

Arsenio Santos, CTO and co-founder of Cir.ca, spoke at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, about mobile journalism and using data for building a new kind of news service. Circa is a 2.5-year-old mobile app based in the United States, and soon to expand worldwide.

Cir.ca has focused its app on solving the problem with mobile news: Most people want basic information and not a long narrative on their mobile phones. The focus, then, of Cir.ca, was to develop atomized news "cards."

The mobile news stories are collections of news items written by journalists, and not generated by machines, as many news aggregators are now doing. The items are nuggets of news items and are interconnected by links for easy access to more information if the reader wants it. As the story develops, Cir.ca only pushes new information, not the background that has already been published. Big Data on registered Cir.ca users are employed to send users only information they have requested to follow, and only information they have not yet seen. 

Santos also explored a current debate on whether journalists should have access to metrics. His view is yes, give journalists access to everything (users, pages, sources, system and editorials).

The metrics collected my Cir.ca that "matter," according to Santos, are explicit user behavior and implicit user behavior. Explicit user behavior includes story reads (good), social sharing (better), story follows (best). Implicit user behavior includes: interest graph (decent), heat maps (problematic), card reads (awesome).

Santos also identified breaking news as an area of journalism to re-think, showing examples of failed breaking news attempts, particularly online and on TV. Santos said breaking news is 
an escalating arms race for urgency and attention" but that it is overused by media companies in order to earn traffic and audience. Instead of breaking news on mobile indiscriminately, Cir.ca breaks news on the mobile app sparingly. Users opt-in to the ongoing news feed on any particular topic, and Cir.ca doesn't want to turn off readers with meaningless breaking news alerts. 

As an example, the story of the missing Malaysian Air flight MH370 by numbers:

A narrative spanning 3 stories over 9 weeks (so far). Comparatively, other news outlets have reported breaking news about the missing plane's investigation without any real news about the plane's whereabouts.

Cir.ca's stories:

Malaysian Plane Loses Contact

Malaysia Airlines Flight 370 Investigation

Malaysia Airlines MH370 Disappearance Aftermath

85 points, 7050 words.

Santos concluded that in order to publish news tailored for the mobile device, publishers mud:

* Redesign your content for the medium.

* Empower your editorial team with thoughtful transparency.

* Find your best implicit metrics.

* Watch your best explicit metrics.

Big Data Opportunities and Strategies in Channel 4

Sanjeevan Bala, Head of Data Planning & Analytics from Channel 4, is a big believer in using data to innovate the broadcast business model.

Bala spoke at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, saying Channel 4 has millions of registered viewers, which they have connected to their viewing history on C4.com. This allows the company to segment viewers into groups, so that they can offer personalized emails and suggest tailored content recommendations and serve relevant advertising and gain revenues.

How essential customer insights and engagement are made possible by Big Data

Jonathan Lyon, Global Director of Strategic Insight in DigitasLBi, a global marketing and technology agency, is focused on how to better engage with the consumers.

The new imperatives of publishing and marketing are to know your audience better beyond measuring and counting, he said at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism,

So how can we level the playing field? Understanding people individually, in depth and at scale, he says.

Lyon's company has mapped 1.3 billion individuals' engagements with 50 000 brands, TV shows, media, celebrities, musicians, bands in near real time of the last two years with People Pulse survey software.

People engage differently in different platforms, so it's important to survey all the different platforms.

Some case studies from DigitalLBi:
1. Volvo and creating context, showing Volvo ads with suitable content after mapping the consumers interests.
2. eBay, matching email and social media using habits. Showing the customer the products, which he is interested.
3. TV engagement by brand by program within broadcast content clustered by affinity and engagement between brand and TV show.
4. Movie taste is broadly reflective with the people's gaming habits.

"Digital information is just people in disguise", says Lyon.

How Twitter, Facebook, Instagram, Snapchat usage are changing the face of journalism and business development

Dr Farida Vis, acclaimed social media data scientist and occasional Guardian Twitter analyst, University of Sheffield, spoke about the importance of social media platforms and apps at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, 

Among the most notable social media landscape changes are the rise of platforms and apps focused on visual content on a variety of social platforms including:
- Pinterest
- Tumblr
- Instagram
- Vine
- Snapchat

Importantly, visual content drives retweeting, Vis says. "Photos give verified users a 35 percent bump in retweets. Images important for multiple reasons."

However, the rise of social media has given rise to inaccuracies and the need for a greater degree of journalistic verification. "The price for inaccuracy has never been higher," Vis quoted author Craig Silverman saying in his 2012 book, "Verification Handbook."

Vis analyzed the case of fake pictures of Hurricane Sandy images but also mentioning post Boston-marathon bombing.  "Twitter does it best work in the first five minutes and its worst in twelve hours after that," wrote Matt Roller a writer based in LA tweeted.

Vis mentioned Izitru (http://www.izitru.com/), a new image hosting service that helps you verify the authenticity of an image.

The changing social media landscape

Jason Mander, Head of Trends for Global Web Index, a London-based global research firm, laid the groundwork about the social media landscape at the Big Data for Media conference in London on May 8, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, 

GWI's social media research surveys 170 million Internet users four times per year in 32 countries, or 89 percent of the Internet population.

The study has found that Indonesia is the world's most active active Facebook and Twitter market, while China is the world's most active e-commerce market.

"Often countries with lowest internet penetration shouldn't be disregarded. Often their audience is the most engaged one," Mander says, adding that these markets represent a huge opportunity for marketers and publishers online.

Despite its much-publicized recent decreases in active usership, especially among the young, FB is still on the top of the charts: more members than anyone else, more visitors than anyone else, he says.

Those social media platforms with rising active usership in the past six months include Changes in Instagram, Tumbler, and Pinterest, he said.

While it unclear exactly the reasons for the social media landscape changes, Mander says there is a sense of social media fatigue, and social media users are becoming more passive, using Facebook as more of an email box rather than a place to spend time, hang out, converse and share. Part of this phenomenon is caused by the shift from desktop/laptop social media use to mobile. Less time is spent on Facebook sessions on mobile, likely because of the size of the screen, and the places where users are posting...during a free moment, for example.

Some of the fastest growing social platforms include Snapchat (37 million users) in the past six months. Best results in westernized most developed markets: Australia, US, UK. However, Snapchat pales by comparison to Wechat (408 million) much bigger in terms of users (China). Mander pointed out that western media tends to get more publicity, but cautions that other chat platforms, especially in India and China, has far more reach.

Mander pointed to some challenges in tabulating social media trends, including:
- The growth of device sharing per household. This makes it difficult to identify just how many social media platform users there are, as these users are identified by one device.
- The proliferation of VPNs (virtual private networks), as a major social gateway. 416 m (28 percent) of the social platform-using population use VPNs, which allows them to stay anonymous, and get you to website that are blocked in your own country (YouTube, Facebook, Netflix, Guardian, BBC website) in countries like China where accessing these sites is not allowed.

How Tesco and dunnhumby have led the way in the business approach to Big Data in customer and efficiency strategies for 15 years.

Matt Keylock, Global Head of Data for dunnhumby, spoke at the Big Data for Media conference about his company's and Tesco's work in developing a customer strategy behind Tesco's club card. dunnhumby is a global customer science company owned by Tesco.

Traditionally there were financial metrics for keeping scores and population profiles for marketing in the retail business, but it gave only very superficial view to the customer, he says at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism,

Customer behavior is the key and knowing the best customers, because 4 percent of customers bring in 41 percent of the sales. So one super premium customer corresponds 351 uncommitted customers.

"If you don't engage with your best customers, they stop being your customers," Matt Keylock says.

The second thing to do is understanding the customer's needs, which dunnhumby has been doing for 15 years for their retail customers. The customer can be profiled through different attributes, e.g. price sensitive, promotion sensitive, adventurous, pre-packed, convenience, fresh, kids, fuel, calorie counter, coupon redeemer, finest and luxury, foreign, fair trade, long life, early adopter, big box and own brand label.

When you know your customers needs, you can build better strategy and create better measures.
The consumer data enables smarter actions than just analyzing the traditional sales figures. E.g. customer's product loyalty compared to product sales.

Customer data powers also personalized communications. E.g. the overall response for the Tesco's personalized campaign 4 times a year is 70% and the sales goes up during it.

When Tesco launched club card, it was the second in the retail business after Sainsbury's with 15% share, now it has a 30% share.

Digital is changing also the retail business and the shopping trip is becoming iterative and non-linear, where there are many different actors influencing each other.

The Big Data journey begins from many different data silos, where the information is separate, but we need to enrich that data and insights.

"Data silos create more complexity, but retailers need to connect all that data. Big data allows for a 360 degree view about behavior, motivations and influences and enables the step to fully personalized service," Keylock says.

To know and have the perfect customer data does not mean, that you should always use it, e.g. when it is a very personal issue.

Keylock says that doing this kind of customer analysis is also possible in media business.

Some important questions for the media companies:
1. Do I know my audience as individuals?
2. Is my media measured right? What's the ROI?
3. Which content attracts niche customer groups? How should I price and sell my channel better?
4. Why is my media not personalized yet? etc.

The process starts by measuring the effectiveness of each media channel.

Major opportunities for media industry:
1. Loyalty approach for your own audience
2. Benefit from other data
3. Drive targeting and measurement
4. Use your data to influence beyond your existing business

Stop leakage to third party competitors.

Paul Hood, Archant Media’s Digital Director talked why leaking data to third party competitors, like Google is lethal for traditional publishers. Archant Media is the UK's largest regional media company, which has 4 daily regional newspapers, 50 weekly titles, 80 regional and lifestyle magazines and about 200 digital properties.

Hood spoke at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism,  and said that the challenge in today’s media business is to put the customer at the center of the strategy

“Premium content creates actionable data”, says Paul Hood. It means, that relevant content helps to deliver high impact ads to the readers. That is why, the media houses have to end leaking the user data to the third party sources, which can use that data with much more lower prices.

Much of the web’s content is inactionable data and unrelevant, but the media publishers produce actionable content, which has high value.

Online display ad revenues are growing every year, but the revenues are not going to the traditional publishers. Today’s market is ultra complex to tacle for the publishers. Also the advertising customers buy, where they get the most cheaply and efficiently. The publishers have hard times in replacing newspaper dollars with digital pennies. Third parties have also better tools and reach than the traditional publishers.

What to do then about it?

“Don’t let confusion paralyse you. Don’t wait, but start something right away”, Hood says. “You have to move forward or there is a risk to die.”

The steps for minimizing data leakage:
* Reclaim your online display inventory
* Stop hemorrhaging data into the remnant ad networks
* Embed native ad formats into your sites
* Audit third-party tags on your sites
* Read the fine print on the Terms of Use of third-party ad networks

Minimizing data leakage is the first step to reclaiming fair market share. Big data isn’t about the data, but about the analytics and metrics that matter e.g. demographics, social, contextual, friends, trends, intent, family and interests. The basic data, like visits, is also useful, but do not tell a lot about the customer.

“Premium content creates premium audiences. The value lies in understanding those audiences. Don’t give audience data away for free. Act now and do every month something new, even baby steps are acceptable”, Hood concludes.

How Sacramento Bee is Using Data Visualization to Extend the Utility of Big Data Sets?

Darrel Kunken, Director of Marketing, Sacramento Bee, talked about data visualization for advertising and marketing at the Big Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism,

The Sacramento Bee is part of The McClatchy Company, which is a 157- year old newsmedia chain in the United States. The company is on the digital transition road for positioning itself to the future. The company uses data visualizations to reveal more and quicker from the existing data sets in advertising, circulation and corporate level, Kunken said.

One good example in using data visualization in advertising is ‘auto dealer share of sales’ shown in visual charts, which are based on different filters (multiple years, dealership, make/model, unit sales). The advertising reps use the visual dashboards, when negotiating with the ad clients e.g. they talk and show ‘needs analysis to consumer path-to-purchase tool’. The tool shows who is consuming the content and where and what kind of needs the reader has, which is valuable information to the ad client.

Circulation department uses data visualization, when e.g. making survey about the former subscribers.

“Data visualization helps to move the legacy staff to digital," Kunken says.

The Sacramento Bee asked, what readers want from them. The readers answered, that they want to be informed. Sacramento Bee wants the readers to come back and engage, so they have created an individual content profiling and tracking system in project with Stanford University. The system helps to customize the user experience. The system ads metadata to the content and constructs a user profile to recommend content to the reader.

Setting up for learn and actions in using data:

1.     Data sourcing and integration

2.     Analyst and marketing talent, the staff skills

3.     Utilizing analytics, how to use the data?

The Sacramento Bee has programmers inside the newsroom.  Research analysts are becoming data miners and the company is opening up Beelabs and beginning a Linkedin group.

Transparency is one of the cornerstone components of collecting data, e.g. readers can manage their own profiles.

The company is also having innovation projects, like a high school sportshow build in partnership with Concast.net. There are also plans to engage the schools, which could sell the subscriptions for the channel and get funds for their sport programs.

Next generation journalists need to be savvy in databases, multimedia, do some coding and engage readers.

Making Sense and Nonsense of Data: Successes and failures Using Big Data in Magazine Marketing

Pegg Nadler, award-winning magazine marketing guru from New York spoke about magazine business in US. She has made a survey, where she interviewed 15 major magazine publishers representing 250 titles in US.

At the sBig Data for Media conference on May 8 in London, sponsored by World Newsmedia Network and Reuters Institute for the Study of Journalism, said that in 2014, there are 2,675 consumer magazines in United States. Consumer magazine circulation revenue declined from US$9.8 billion in 2008 to $7.5 billion in 2013. Newsstand sales continues to drop. The entire industry is in a flux and hoping that big data will save the industry, she said.

What is then Big Data? Traditional data is now being called in US as "big data", because there is so much of it and it has been collected for ages from multiple sources. It's also a combination of offline and online data.

The old data concerns are gone, because data storage is no longer an issue, data is easily attainable, data processing quicker than ever and data matching and cleansing improving. 

The "true" big data means detailed data collected from websites and digital sources. Big is not just volume, but velocity and increased activity, size and scales not important, but the most important is, what you do with it.

How to use big data: 

1. Start with determining the business needs and questions to be answered. What is troubling?

2. Identify the data that will get you there.

3.  Build the technology around the information to get it.

Different business needs and questions: customer buying cycle, cross channel view, customer purchase path & preferences, customer research modes, feedback behaviours, next best offer.

Data assessment: What data do you have? What data do you need? What other data is available?

After that you have to identify the different data resources and start analyzing, e.g. do customer segmentation.

“The traditional plan was to use data to understand the past to predict the future. Now we are going into uncharted territory, where the past does not excist”, says Charles Swift from Hearst Magazines.

Hearst succeeded in using big data for multi-channel customer understanding. They discovered that the database needs to be rebuilt to succeed in the data-driven analytics environment.

Forbes discovered by analyzing the data, that timing is everything, when publishing content.

The amount of available technology for data analyzing has grown dramatically.  

Magazine publishers are at all different stages in learning, using and understanding big data. Marketing needs will drive advances in database systems and solutions to track, communicate and monetize customers. Big data usage involves a blend of technology, people and process to ensure success. Big data spans all types of data (including traditional) to maximize using information and analytics.