Key Aspects to Consider for an Effective Use of Data
Wahab Moradi & Leonie Gehrmann
Nowadays, developments such as the Internet of Things (IoT), user-generated content, and the tracking and connectivity of devices have revolutionized the world of data. What is popularly referred to as big data is characterized by the three V’s: volume, variety, and velocity (Tonidandel, King and Cortina, 2016). These large data sets bear great potential. The sheer volume can be leveraged as a competitive advantage, assuming more data provides more accurate insights on customers and the market environment (Lycett, 2013). With an understanding of which data can be discarded, firms are able to speed up data processing, generating information faster and allowing for quicker decisions and actions (Lycett, 2013). And while traditional data sets are typically structured, big data is generally obtained from a number of different sources, resulting in unstructured data sets with varying data types (Tonidandel, King and Cortina, 2016). The practice of using data to predict scenarios, trends, or even changes in customer needs is not novel, internal transaction data is often used for decision-making or forecasting. However, big data changes the requirements and tasks of companies working with data. Instead of merely falling prey to the hype, a deliberate use of (big) data should be the focus.
Leading companies recognize that data is a strategic asset that needs to be gathered, adjusted, and analyzed to gain meaningful and actionable insights (Lukosius and Hymann, 2019). If successfully managed, the strategic use of data can result in a competitive advantage. However, this requires the development of a clear, consistent, company-wide and well-designed data strategy. Companies must ask themselves what the intended goal and scope of the data strategy is and consider the tasks of data collection, storage, and processing, as well as the potential insights and generated value (Erevelles, Fukawa and Swayne, 2015). For example, marketing departments stand to benefit from a strategic use of (big) data by analyzing market data und consumer behavior to generate measurable value (Kumar et al., 2013). In general, there is no single universally applicable data strategy, but this post provides first key components to consider.
Scope, purpose and mindset
First, the data strategy should reflect the company’s mission and vision regarding data handling, ensuring that the data strategy has an identity and role within the company and is in line with the overall strategic goals (Thompson et al. 1999). Only once clear objectives for the data strategy have been defined, can the scope of the necessary and relevant activities and data sources be considered. Thereby, a reflection of the potential use of each instance should ensure that the company’s resources are used efficiently. Furthermore, the definition of key performance indicators that measure the success of the data strategy enables the discovery of optimization opportunities (Thompson et al. 1999).
So far, statistical models that make assumptions about the underlying relationships between variables have been used in data-driven decision-making processes (Breiman, 2001). However, the unfeasibility of processing unstructured data sets via predefined stochastic models requires a change in mentality and culture, a great challenge for many companies. When defining the scope and purpose of the data strategy, managers must also assess how data-driven the corporate strategy and employees are. Since analysis tools for (big) data are under constant development, flexibility and eagerness to learn are required.
Data collection, storing and processing
These general considerations guide each of the following steps in the use of data. The collection and storage is particularly important to ensure sufficient data quality for further processing (Simsek et al., 2019). However, especially the volume of unstructured big data renders traditional hardware and software solutions inadequate to acquire and store the necessary data, requiring an investment in new solutions (Bharadwaj et al., 2013). For example, very large and unstructured data bits could be pre-processed and stored in an aggregated manner. In addition to ensuring that the data collected fits the scope and objectives of the data strategy, it also must be cleaned before storage, thereby removing redundant or meaningless information from the data set (Lycett, 2013).
Beyond an adequate IT solution, companies must also consider what type of infrastructure to use for the collection and storage of data. This decision depends on the future uses of the data. Legal regulations of data privacy must also be respected and employee access rights to data within the company must be reflected upon. Weinberger (2007) differs between a clearly structured data architecture or “sorting in the way in” type and an unstructured data architecture or “sorting in the way out” type. For the first type, data is sorted and its location determined definitively once it enters the system or data infrastructure (Constantiou and Kallinikos, 2015). For the second type, unstructured data and collected or generated information are sensibly categorized and compiled and can be used for different purposes in the future (Constantiou and Kallinikos, 2015).
Similarly, data lakes, large data sets stored company-wide in their native format, are another possibility for data storage (Kitchens et al., 2018). The elimination of barriers and preliminary costs for data exchange within the company are some of its advantages (Kitchens et al., 2018). However, the lack of standardization and integration of the data requires relevant data to be fished out of the lake, reorganized, and assigned to the holistic customer view for each analysis (Kitchens et al., 2018). In the worst case, if the data has not been collected appropriately to allow a link to other data sources, it might not be useful at all (Kitchens et al., 2018).
Data processing leads to the extraction of insights. Taking a look at the marketing department for example, individually collected and stored data bits themselves do not provide high value for companies, since they contain too little information as a single data element. Only through processing and aggregation can the separate data elements be placed into the correct context and consumer behavior patterns recognized (Chen, Chiang and Storey, 2012). Since external data is not exclusive to a single company and can also be analyzed by competitors, a combination of internal and external data is likely to provide more sophisticated and unique insights (Grover et al., 2018).
The nature of unstructured data provides a new challenge for data processing since now text, image, and sound must be analyzed (Hu et al., 2019). Furthermore, when processing large data sets, less traditional models, software and hardware can be applied and more alternative methods such as artificial intelligence (AI) and machine learning (ML) are necessary and can provide real-time analyses (Adamopoulos, Ghose and Todri, 2018). However, to build up the required competencies, not only financial resources but also notably human capital are needed. According to a Bain & Company survey, 56% of executives report their company lacks the skills to develop deep data-driven processes (Grover et al., 2018). So when designing the data analytic processes, managers need to define what insights are necessary and what procedures, products, and services should be improved. A review of the existing and potential assets and capabilities will determine whether sufficient human capital lies in the company or must be purchased externally.
Generating insights and creating value
In general, there are an abundance of opportunities for firms to create value using big data. Focusing only on marketing instruments, insights of a clear data strategy and the associated development of processes and structures for the utilization of data can create added value for customers in many ways. For example, products and services can be improved, customer experience can be enhanced, and internal processes can be optimized to improve response speed to customer inquiries (Grover et al., 2018).
Companies such as Netflix, Spotify, and Uber have revolutionized their industries by using big data to innovate products and services and generate a competitive advantage. Their ability to use data more efficiently and effectively for innovation processes than traditional competitors is related to their structure, culture, and internal processes (Troilo, De Luca and Guenzi, 2017). These companies minimize the uncertainty and risk associated with innovation by proactively using big data to analyze customer needs and feedback in the form of consumption patterns or social media postings to support the market success of new products or identify potential defects. Furthermore, big data opens the path to entirely new business models such as is the case for the insurance industry and novel approaches of the measurement and collection of data. By incorporating the information from various sensors capturing things such as speed or braking habits, companies can offer new insurance products dependent on the individual driving behavior (Varian, 2010).
More companies recognize the importance of customer-centricity. It is difficult for companies to differentiate themselves significantly from their competitors in a connected and highly transparent world. Therefore, marketing departments in particular are called upon to offer a fulfilling customer experience (Troilo, De Luca and Guenzi, 2017). Insights extracted from big data can aide in the analysis of customer touchpoints, determining their amount and quality, as well as necessary improvements and where customers do (not) wish to encounter them (Troilo, De Luca and Guenzi, 2017).
Similarly, nowadays customers can be divided into much smaller and more detailed segments based on preferences, characteristics, and personality, allowing companies to meet the trend of consumers’ desired personalization. Recommender systems are exemplary value-generating personalization services, suggesting products or services that match the customer’s preferences or past purchases. Additionally, personalized advertising lets customers feel directly addressed and considered, strengthening their relationship with the company (Varian, 2010).
Discussion
While clear potentials of big data for value creation were mentioned, it is important to keep in mind that a blind jump into the vastness of big data analytics is likely to be unsuccessful. Rather, companies must conduct research on AI applications in order to analyze the large amount of different data types in a reasonable timeframe to draw meaningful insights. There still exists potential in the development of methods e.g. to analyze non-textual data. Additionally, novel collection and storage solutions must be determined with fewer disadvantages and more regard for data governance and security topics than initial approaches. Current trends such as autonomous vehicles, increasingly connected customers, and IoT will further multiply the amount of generated data. Furthermore, the use of big data requires ethical considerations and companies to reflect on the ensuing privacy restriction of consumers and societal transparency.
As indicated, there is no universally applicable data strategy. Managers wishing to implement an effective use of (big) data in their company must consider each of the mentioned components when designing an approach tailored to their company, industry, and goals. Honoring an efficient and effective use of a company’s finite resources, a successful data strategy collects and stores meaningful data while disregarding meaningless data. Furthermore, the rise of big data external to the company does not diminish the importance of internal data generated through interactions among employees, customers, and suppliers and available only to the company. In fact, an appropriate mix of external and internal data is another key to success and the basic considerations mentioned here generally apply to all data-driven activities in a company – regardless of the size of the data set.
Concluding, managers face three main important challenges in the development of a clear data strategy. First, the corporate culture must reflect a data-driven and customer-centric way of thinking and an internalization of the importance of big data. Second, companies must find talented human capital specializing in data analytics and AI. Third, increasingly stringent data protection and security regulations impact companies’ data collection and storage processes and must be considered in the strategy development process.