The Deep Customer Insight work package focuses on gaining information about the true customer value of potential services, product features, and other possible aspects of user interaction with a service or a product. The goal is to enable data-informed design decisions through the specification, development, and evaluation of information gathering methods, techniques and tools. A deep understanding of customers, product, and rapid feedback is garnered by gathering data continuously from the live use of the products.

The business hit rate—that is, sales volume and revenue— will be significantly improved by linking product-, service- or idea-related activities to the business operation, using live customer feedback and validated knowledge about customer value (trends, markets, etc.).

The work package

• provides multi-disciplinary mechanisms that companies can use proactively and early for validating innovative idea, feature, product, service and business models as well as future customer needs
• provides support to organizations for making data-informed design decisions for developing new software-based products and services
• builds robust, automated feedback systems, for both new and existing products and services
• develops a tool-based infrastructure for continuous experimentation and live customer feedback that can be used efficiently by organizations in different domains
• cost efficiently executes massive scale experiments with live users, together with the analytics tools necessary for studying the real use of products and services
• explores advanced mechanisms, such as recommendation systems, to predict customer behavior using existing data, or mimic real customers based on validated models.

Research areas

• data collection, real-time feedback from real customers and environment
• data analysis, visualization, interpretation and business integration
• continuous experimentation

Data Collection, Real-time Feedback from Real Customers & Environment

The focus of this research area is to create the mechanisms, tools and methodologies to collect usage and behavioral data and feedback, as well as different trend, market, technical, competition, weak signals, services data, from different sources. It is important to gain an understanding of what data should be collected. For example, is the primary interest the usage of a single feature, the particular combination of features, or an entire new service or product? Appropriate measures need to be derived from measurement goals, including business-related measures such as innovation accounting (i.e., what has been learned about the customer value).

The main focus is to collect usage data or other value-related data without interfering with actual use. However, other, complementary means of collecting user feedback or other information should be utilized. Such means include, among others, traditional usability tests, interviewing and observing users, user surveys, active user feedback directly from applications, and different social media channels where users discuss, evaluate, and critique the services and products they use. Data collection should be based on rigorous scientific principles of data collection and experimentation, clearly adapted to the needs of industrial practice. For example, data collection should not distract from other product development activities.

Data Analysis, Visualization, Interpretation & Business Integration

The primary purpose of experimentation and data collection is to provide a factual basis for decision making. This requires that the collected data is properly analyzed and correctly interpreted by the respective stakeholders. The proper methods and tools need to be readily available and built into the overall experimentation setup. This includes the availability of open source software for analytics and other tool support built onto OpenStack and other open source software platforms. In the interpretation of data, potential differences in the context of usage must be understood and taken into account, in order to avoid the dangers of serious misinterpretation and consequent wrong decisions. Finally, the data needs to be graphically represented, and must provide a clear link to customer value and business goals, with clearly interpretable answers, in order to support good decision making.

Continuous Experimentation

The focus of this research area is to create an experimental infrastructure consisting of the mechanisms, tools, processes, and methods for efficiently running real-life fast experiments. For example, innovative features that are not familiar to customers need to be tested as early as possible so that their real value can be understood. The experimental infrastructure needs to be easily adaptable to the needs of different organizations so that they can customize it effectively for their purposes. It is first necessary to analyze which experimental designs and testing methods are appropriate for different goals and contexts. Based on this analysis, decision support for selecting appropriate designs should be provided. In addition, the selection of experimental subjects needs to be better understood and guidance needs to be developed.

Given the overall goal of rapidly getting customer feedback through continuous experimentation, the experimental setup—including measurement instrumentation—needs to be highly automated. Appropriate mechanisms and tools for this kind of automation need to be identified or developed. Concepts for storing and packaging the results from conducted experiments will be part of the experimental infrastructure. Experimental results could, for instance, be stored in a company-specific “experience base” that captures the knowledge and supports advanced analyses of customer behavior and customer expectations. Facilities and laboratories such as Software Factories could be established to create experimental objects (such as prototypes or minimum viable products).

← N4S Treasure Chest – How to make software business in real-time