Introduction: A Value Shift from Tangible to Intangible Properties

Throughout history, “mankind has wanted to know to be able to act” (4) and as such businesses which had or have the right knowledge – to make better decisions leading to rational actions based on facts, rather than on intuition – have a real competitive advantage.

Data are particles of information (4) something that can be recorded,  analyzed,  and  reorganized  (5). For decades to stay competitive businesses have exploited data to keep up with a continuously faster-pacing world.

In the 20th century, value has shifted from the physical, like factories or land, to intangible properties, such as a brand or intellectual property. This now encompasses data, a vital asset, and the basis for new business models, innovations, and economic value. Data is seen as indispensable for our “information society” to learn from and hence its use is expanding vastly to a point of impacting all aspects of our human endeavors. What Mayer-Schöneberger and Cukier (5) call “big data consciousness” is the presumption of a quantitative component in everything we do. What the authors term “datafication” means recording and quantifying information of everything under the sun: from natural phenomena to essential acts of living (like a person’s location), now accelerated through embedded sensors in everyday objects (Internet of Things). By applying math and inferring probabilities from information, the long-term goal is prediction. This aspect is characterized through the use of advanced technologies like artificial intelligence (AI) or more specifically, machine learning, and supersedes the mere definition of “big data”. Information technology is still linked to people transforming information into knowledge by applying  context and meaning (6), however, it can be observed that human judgement can be replaced by computer systems (7). 

The following sections aim to give an overview of how big data has evolved – what came before it and what comes after it – and what role human beings play in the future of big data. Furthermore, it will be set out why specifically human judgement is the new competitive advantage in the world of big data.

The Growth of Data before Big Data:
E-commerce and Better Marketing

As mentioned, the datafication of all aspects of life goes hand in hand with the advent of big data (5). Even before big data, humans have always attempted to quantify the world (like population censuses) as well as to measure and track production and business transactions (ibid).

An increasing recording and datafication of everyday life created the promise of a better understanding of how humans think, what they do and what is shaping their behavior (8). This was a real value and differentiation to organizations, to better understand their customers, to create better strategies – better services, products and marketing (5,8). What is now known as business intelligence (methods for managing business data to improve competitiveness, monitoring, etc) used to be called decision support systems at the time before e-commerce (4).

According to Kurzer (9), the first gathering of data started with e-commerce.  Merchants in stores would observe details of  customers, what they bought, and when, which would then form the basis of marketing plans. Although one could correlate between advertisements and customer purchases, this was nevertheless speculative.  Once it was possible to record information of customers electronically, businesses were able to create more personalized marketing. In the 1980s, businesses would create direct mail campaigns based on personalization, and through digitalization in the 1990s, a download of an e-mail coupon could be correlated with in-store behavior. (ibid.)  Marketing expert Don Peppers (10) wrote in a Forbes article in
1995 that emerging computer technology would transform traditional marketing to mass-customized marketing. Based on past purchases, a business could send personalized offers to customers. A flower shop, where you once bought a Mother’s Day gift, could record your transactions and send you a personalized offer the following year, asking if they should send the same bouquet to your mother for the coming Mother’s Day. If companies can track customers individually, they are going to focus on their share-of-customer - trying to sell them as many products and services as possible – as opposed to their market share.(ibid.)

After Big Data: Smart Data & Data Analytics

When Kitchin (15) talks about the “Data Revolution” in 2014, he notes that while it had a substantial impact in a short amount of time, this is still only in its infancy. In 2020, big data forms a part of any data-driven enterprise and, at the same time, connected devices generate massive amounts of data every day (18). The three formerly described “V’s” of big data now graduate to also include value (4) and veracity (18). This can be seen as the next step of differentiation for a company, where the new winners are the businesses that process their messy data on time, in order to generate high-quality data, leading to useful insights (ibid.). So, while the first chapter of big data can be seen as the age of gathering and storing data, the second chapter is that of analysis. What terminology would replace big data, however, is not clear (19). De Goes (16) expressed in 2013 that terms from “Smart Data” to “Data Science” and “Predictive Analysis” have common attributes with the aim of applying advanced techniques – from processing to artificial intelligence (AI) - to generate further insights from big data.

​How Smart is Smart Data? What AI and Automation Can’t Do

​Big data, AI, machine learning, and algorithms work well for optimizing business processes (27) by seeing through the chaos of a company’s many data transactions and supplying new correlations – this is something no human can do. In the last ten years, companies invested in these new technologies with the quest to become better (28,29). However, they are not experiencing better decisions (29). Instead, companies are becoming more bureaucratic: for example, AI decides based on rules – inferred from past data (28). Additionally, the reliance on data gives the illusion and comfort of control (29,30), while not really knowing anything. Big data is framed as accurate and objective, but it is only a selection of possible available data, making it inherently subjective (15). Judging people on the basis of data rather than their actual behavior is a risk that comes with the increasing sophistication of predictions (5). Where applications for universities (27), jobs, or credits are now carried out by algorithms and AI, a bias in past data can result in the exclusion of certain ethnicities or postal codes because they are seen as less skilled or not creditworthy (7,27,28). The argument that these decisions are objective because all rules apply equally to everyone (28) creates an echo-chamber of “good candidates”, where the organization may not even be informed about this (7). Automating a brand’s social media interaction is a risk as well, because a machine can misread emotions or languages, where a formulaic response can make a customer feel unappreciated (31). While purchase predictions and product recommendations may be an easily automatable process, a machine cannot understand a context – resulting in further purchase recommendations for burial urns (28). The greater amount of companies is dissatisfied with the investment they have made in big data and its technologies, the higher the likelihood of a slower adaptation of big data technologies by companies (28,29,32). Understanding that technologies like big data remain only a tool and do not offer ultimate answers (5), is the first step to betterment. 18% of the companies that focus 80% of their initiatives on the efficiency of their decision-making (instead of trying to save money) experience growth (28). With the question of who is accountable for the decisions taken by algorithms (7,28,33), new challenges arise. Inferring from the difficulties that machines have in the complex environment of human behavior, the need for a more human component in our work arises.

Humanized Data Requires Organizational Changes in Companies

​​How is differentiation achieved? Companies have to invest time and resources to make the cross-disciplinary collaboration between business, data science, marketing, and human resource teams possible. This means deciding on a protocol of how decisions are made based on data prediction models (28) and asking the right (second-order) questions (8) based on data: Where did it come from? In what context, and what kind of analysis was conducted? (36) This means a new culture of decision making, where important decisions are not taken by the highest-paid executive (35,36), who may be subject to bias, lack the cognitive ability, and discard valuable information (8). This means training employees in big data literacy to make automated processes transparent in order to challenge any biases in prediction models (7). As data scientist jobs are becoming increasingly replaced by automated machine learning tools (37) training computational (38), statistical (39), and skeptical thinking (40) is necessary to ensure that all employees can see through the ‘black box’ of algorithms (7). Organizations that democratize and decentralize data access, tools, and judgement power will have a sustainable competitive advantage (35). As real-time data gathering and modeling picks up speed, requiring the approval of decisions of many levels can boggle down any advancements made (41). As such, enabling more people with access to data management tools and decision-power can result in the right decisions for a particular context on time (35). An example of this is the credit union Affinity (ibid.), which enables its employees to decide freely - with or without the basis of data models and company frameworks – as long as it is in the best interest of the customer in that particular context. This resulted in a reduction of fees by 50%. Where one would assume the contrary, machines can lead to more empathy: Where a job requires emotional labor - like managing a brand’s social media – algorithms can interpret sentiments and help prioritize attention: which users are overly complaining? Which user did interact with the brand but never received a reply? Which user is linked to a broad network of followers? (31) The end-goal is not more data but to understand what the customer thinks and what the market demands (42). This means that in order to acquire the nuances behind quantitative data, direct communication with consumers has to be routinized. One example to achieve this is by reserving a certain amount of time for employees to regularly have direct conversations with their customers (43) – by working as a customer service representative for example (42). Data visualization and analysis has been a popular method for the basis of decision-making and analytical reasoning (15). This includes adding qualitative to quantitative data to not only show what is happening but why it is happening, much less how you can improve it. (42) The ability to visualize qualitative data (e.g. through journey analytics) leads to in-depth insights that can be handed down to more employees (ibid).