1 Big Data Efficiencies That Will Transform Media Company Businesses TV, digital and print media companies are getting ever-smarter about how to serve the diverse needs of viewers who consume content across a range of platforms on many different devices. Every video stream they deliver generates great data that spans everything from interaction with devices and apps to conversations about the content on social networks. It s now imperative to collect and efficiently put the volumes of data to work to increase engagement, loyalty and monetization. IBB s many client engagements helping media companies transform their big data initiatives has taught us that fully accomplishing this vision means combining the right big data vision and strategy, the most effective big data technology, and organizational excellence. By Jonathan Weitz, Miles Johnson and Justin Schutt At the same time TV and digital platforms deliver the highest reach channels and content, there is continued growth in long-tail niche content, binge viewing and social sharing of content consider the avalanche of tweets that accompanied the season two premiere of House of Cards on Netflix or the series finale of AMC s Breaking Bad. There s also a greater variation across segments, with youth behavior diverging from baby boomers in their preferred viewing platforms, and how different genres of content are promoted and aired to those
2 Navigating the Ocean Of Data Big Data Strategy and Governance audiences. Traditional release windows for programming are changing too. With advances in broadband and mobile access, social media and digital outlets, we're also seeing a rise in the need for strategies that can take advantage of all this new consumer data. Rapid testing of various approaches is also taking place given the low cost presented by elastic cloud infrastructure solutions. The rise in new tools and analytics, such as cohort and content loyalty, are helping drive a better understanding of audience behaviors. This includes both individual and household-based understandings of behavior on a range of platforms and devices. The potential is here to obtain incredibly rich information for marketing content, targeting ads and increasing cross-platform engagement. These developments have resulted in big changes for the technology departments of media companies. More than just an increase in the volume of data, media companies are now getting direct access to their customers and information about the products they use. The data sources include consumer devices, set-top boxes, social platforms like Twitter and Facebook, apps and websites each more varied than before, generating data at higher volumes and velocities. Research and analytics departments at media companies also face new and unique challenges. They need to answer increasingly complex questions about consumer behavior that requires cross-functional data analysis, all while accounting for sensitivities around personallyidentifiable data. Today s business environment requires flexibility and agility, as insight must be delivered with speed and accuracy. Many media companies are moving quickly to respond to these challenges. These leaders are focusing heavily on the big data opportunities in their consumer, advertiser and partner relationships and are making considerable investments in data capabilities, focusing on the key areas that need to be addressed to ensure long-term success: governance, architecture and the shape of the organization. Data governance should be done with a focus on simplicity and flexibility of data processes. Smart strategies provide a foundation for streamlined operations, facilitating cross-functional data management. It brings
3 clarity to responsibilities and ownership and alleviates the risks associated with data movement throughout an organization. A common aspect of data governance is to create a forum that facilitates rapid integration of cross-functional data sources and manages downstream impacts resulting from changes to data in the source systems. For example, if cross-functional analytics teams are trying to correlate mobile video app usage with ad sales information to enrich targeting, changes to how usage information is represented in the product will have downstream impacts on the targeting models. Good governance allows such changes to be handled gracefully and efficiently. The greatest governance challenge we typically see is how to represent a single identity for a user or product across multiple systems. Different code sets often proliferate, creating headaches for analytics teams that must play catch-up to bind different data together with some meaning. Tackling this will position an organization for much simpler integration of future data sources. Because governance works across departments, such initiatives require high-level cross-functional sponsorship from the organization s leadership. Federated approaches that balance this decision making with bottom up analysis of data sources can often be the most effective. In fact, if done properly, good data governance can reduce overall bureaucracy through a set of principles rather than extensive processes. Scalable and Flexible Big Data Architecture When it comes to evolving the data warehouse architecture, companies have to evaluate a range of technologies to meet differing needs. This includes everything from traditional enterprise data warehouses for interactive analytics to Hadoop solutions that support cost-effective scaling of more diluted data. Additionally, most companies will already have existing relational data stores, used for things such as Master Data Management, long-term storage of aggregated data, and existing structured data sources. These will continue to offer great value to the business. Integrating new technologies with these solutions can also provide a means to shield existing users from any underlying platform changes and transition at a pace that suits the business, rather than the technology need. Companies with existing large data warehouse implementations will benefit from enhancements to their architectures that enable rapidly
4 Big Data and The Media Organization adding new data sources, for example to combine anonymous user data with personally identifiable information and enriching user models with 3 rd party data for audience analysis and ad sales. As data volume, velocity and variety grows, technologies like Hadoop will become more and more important. Designed from the ground up to handle processing of high volumes of unstructured data on commodity hardware, it can be used to great efficiency. For example, we ve seen a recent trend where media organizations want to identify and communicate with their super fans users that might binge-watch a series, then buy the DVD on Amazon and also comment frequently on Twitter. This can require an intense data enrichment and processing capability. Being able to offload this work to Hadoop and commodity hardware-based clusters rather than the typical proprietary ETL tools, will greatly help overall data warehouse performance It should be noted however, that while Hadoop is a very promising technology and is allowing media organizations to capture more data cost effectively, it was not historically well suited for a range of real-time analytics, and a new crop of technologies has grown recently to address real-time analytical needs. A key challenge faced today is often tightening the connection between the data analytics teams and the big data platform operations. Media companies typically have existing and mature enterprise data warehouse capabilities. But these can be reliant on strict methodical process for reporting and analytics. Placing focus on streamlining communications and making teams more agile can yield huge returns. Simple changes like co-location of these teams can greatly improve productivity and build trust. The skillset of the current business intelligence organization and the day to day analytics tools in use or being considered will also play a key role in architecture decisions. It's not a matter of choosing one technology over another, but evaluating the tradeoffs between existing capabilities, skillsets, and costs to determine the best architecture for an organization. This can often mean a hybrid or federated solution is ultimately required.
5 Additionally- a semantic abstraction layer can be vital for facilitating communication around data between different types of users and consumers of that data. Enabling self-service, high-performance workspaces that use data partitions and workload management can also greatly mitigate the need for IT processes, enabling rapid ad-hoc analytics, while still protecting system integrity. With the second screen app revolution bringing in new data, new groups are standing up quick-and-dirty analytics solutions with open source software, and then using agile practice to iterate and improve can be worthwhile. Bringing these worlds together later will increasingly become a technology and cultural challenge. In the future, growing the organization s core strengths in managing data and analytics will be a key challenge. It is unclear if the supply of talent will be able to keep up with the rapid industry expansion of data science. Ideally, candidates should have the blend of skills needed to meet the need. If those candidates come internally, so much the better, since real-world experience with the company's actual data is invaluable. By using the best ideas, whether from leading Web companies or homegrown innovation, and coupling them with unique, deep institutional knowledge, media companies can leverage the incredible data and resources at their disposal and extract new business insights that will drive the next wave of media services delivery and the next generation of customers. For more information about IBB s big data experience with media companies, please contact