“Big Data” or “Media Companies Stranded on Their Data Journey?” Marc Schröder, 22. Oktober 2020
A Pragmatic Approach for a New Travel Plan
Five years ago, the realization that global platforms from the U.S. and China, controlling the customer interface and creating their competitive advantage based on their access to data, sent shock waves through several industries. The data economy was born. Since then, a great deal has happened – especially and also within the media sector.
On the one hand, examples such as Netflix and Disney demonstrate that even in the age of the data economy, traditional competitive advantages, e.g., access to talent and content quality, can play the decisive role. On the other hand, traditional media players have also asserted the point of view that data and the insights derived from it are the necessary requirements for superior decisions, products, and services.
Of course, it would be quite bold to claim that media decision-makers had not made data-based decisions in the past – one must only think of the comprehensive and extremely granular AGF-TV audience panel in Germany, which has provided the basis for programming decisions for decades. But up until now, data was frequently compiled for a clearly defined and isolated purpose, a specific analysis. Moreover, data was often the side product of core value added processes, which had to be collected for other reasons, e.g., to satisfy documentation obligations or legal requirements.
There have been many attempts to describe what is actually the new point of view on data in the data economy. And there are even more views on what it means for media companies as well as recommendations for how they should act specifically.
In our experience, it is crucial to approach data projects as pragmatically as possible. In developing data strategies with our clients, we have had the experience that completely new approaches in project management are essential in order to avoid typical pitfalls and develop strategies that have a realistic chance of being implemented.
Why Data Projects in Media Companies Fail
Even though it seems common understanding that the topic of data is relevant and important, and untapped potential is seen in this field, there are often misgivings about it at the top management level. Data is regarded as a technical matter and not a strategic one, which is frequently delegated to IT departments but not instigated by those responsible for the P&L, let alone from managing directors or the executive board. Why is that?
A possible reason is that many corporates and consulting firms approach the phenomenon of data with a traditional, strategic arsenal of tools. Illustration 1 shows the traditional approach.
As usual, at the outset of this traditional approach is the data strategy. Afterwards, the technological, organizational, and cultural framework is created and ultimately, concrete applications (use cases) are developed. So far, so problematic.
This approach contains a series of hazards. Here are a few examples:
- The strategy is often developed on the basis of optimistic assumptions about the availability and applicability of data. Technical, data-protection and legal obstacles become evident only later in the process and require at least a re-examination of the strategy.
- Due to a lack of human and organizational resources, the data strategy is entirely developed from outside the company. Very often, there is no in-house data champion, who can assess external recommendations, judge their feasibility, and ensure their implementation within the specific context of the company. Particularly in media companies that have benefited for an extended period from the experience and intuition of their creative decision-makers, this is quite common.
- Technology decisions are often made on the basis of being “fascinated by possibilities” and are not closely oriented along the requirements of subsequent use cases. Oftentimes, cost-intensive programs are initiated in order to "make the company ready for Big Data," although this is not even necessary in the first place. Some applications operate perfectly well with mySQL data bases and do not require a Hadoop cluster.
- When use cases are developed on the basis of a narrow strategical corridor and early fundamental decisions on technology, interesting opportunities are often missed. In this way, use cases are primarily developed with the focus on direct top- and bottom-line impact (see: “Examples of Traditional Data Use Cases in Media Companies” below). Completely new data-based business models are thereby overlooked, which may be off-strategy, but can generate value when regarded as diversification business or opportunistically brought into a partnership with a partner for whom this data offers a strategic competitive advantage.
The Pragmatic Approach – iterative and Agile Data Strategy
The above examples demonstrate that a traditional sequential waterfall approach to developing and implementing a data strategy contains considerable hazards.
We therefore recommend to execute individual project modules simultaneously — analyzing internal and external data sources for their availability and applicability, identifying the necessary technical and organizational prerequisites (infrastructure), as well as developing use cases in the core business and beyond. These activities are accompanied by an ongoing check to see if the results are fundamentally compatible with the business strategy. The data strategy is consciously not the point of departure for all deliberations but to some extent the result of them.
This simultaneous approach can be organized with traditional project management tools. However, the agile project method is particularly suitable. Illustration 2 shows this approach.
In different phases of the project – or as indicated here, sprints – the individual modules can have different priorities. Typically in early phases, the focus is on taking stock of the availability and applicability of internal data in order to broadly estimate which value pools the company can create with data. At the same time, early brainstorming sessions can be conducted about use cases. In later phases, strategic deliberations, which emerge based on initial internal findings, come to the fore. At the same time, technical requirements can also be discussed based on a preliminary list of use cases. The key is, that the results of different phases are summarized at the end of the project into a data strategy that is robust, well structured, and communicable in various levels of detail.
The data economy is here to stay. Most media companies have already embarked upon the data journey. Now is the time to revise the travel plan based on a pragmatic, iterative, and potentially agile process.
Excursion: Examples of Data Use Cases in Media Companies
Value Lever 1: Increasing Efficacy and Revenue
Improvement and personalization of product experience:
- Passive personalization of contents in digital media
- Recommendation function in streaming offers
- Dynamic optimization of spot placement in linear TV to minimize zapping
- (Proprietary) recommendation function for digital editorial contents
- Optimization of storytelling and/or casting for film and video productions
- New products based on facial/pattern recognition in images and videos
(Semi-)automation of pricing decisions:
- Automated A-/B-testing for special offers
- Price optimization for advertising impressions in programmatic and non-programmatic areas
- Dynamic pricing for print ads in the booking tool, depending on anticipated additional printing costs
Refinement of advertising impressions for targeting via (proprietary) data enrichment
Dynamic adjustment of marketing and sales channels
Lead scoring for outbound sales activities
Value Lever 2: Increasing Efficacy and Reducing Cost Basis
- Automation of management reporting in all corporate areas using visualization tools (e.g., Tableau, PowerBI)
- Partial automation of youth protection processes (text passages, video scenes are automatically marked as inappropriate for minors)
- Automated content generation: Editorial items (currently possible e.g., for financial and weather content), video clips, personalized newsletters/activation e-mails
- Optimization of processes in production management (for producers and content aggregators) by establishing automatic feedback loops and benchmarks
- Scoring models based on automated (text) analysis of scripts, concepts, and other creative products