5.2.1. Volume

Big Data is a concept that we have regularly come across in recent years. We can even say that it is almost impossible not to have heard the term before. Moreover, a simple search on Google Trends will allow you to see how omnipresent this term has become in the various conversations around the world.

What seems more interesting is to understand what Big Data is. You would probably like to understand this concept, its characteristics, its opportunities and promises, and everything that has made Big Data so “big”. But the most important thing is to know where we are going to start.

Think, for example, of a Boeing jet engine that creates about ten terabytes of data every thirty minutes (10×1012 bytes). Or, a Jumbo jet plane travelling across the Atlantic Ocean, whose four engines can generate approximately 640 terabytes of data (640×1012 bytes) (Rogers 2011). Now, with an average of more than 100,000 flights per day, can you imagine the amount of data produced per day by all the planes that have flown across the sky?

Facebook, for example, stores photos. This simple statement may not seem impressive to you, right? But once you realized that Facebook registered more than 2 billion users in 2018 (more than the population of China), and that in total, more than 250 billion photos were stored, this may change your mind.

Yes, 250 billion photos is already a lot, but you haven’t seen anything yet. In fact, more than 250,000 photos were uploaded every minute on Facebook in the same year. What? You think one minute isn’t a lot? Alright, come and take a look at Table 5.1.

Thus, every minute in the world, more than 700 people use an Uber, 18 million text messages are sent and 2.4 million snaps are created, 1.1 million profiles are swiped on Tinder, 174,000 images are scrolled through on Instagram, $862,823 are spent online and more (Sedkaoui 2018b).

Do you now see what happens in a minute on the Internet? Do you still believe that one minute is not a lot?

In this case, you can imagine how many thousands of personal and professional transactions are made every minute. We use applications to measure our sports performance, our cars transmit data, just like planes, trains, etc. As such, many industrial and commercial processes are controlled by connected devices. This means that data is transferred from several computers, mobile devices and connected objects around the world.

Big Data therefore involves a huge volume of data. This is probably the first thing that comes to mind when we hear of the concept of Big Data.

5.2.2. The variety

Among all the statistics given in the previous examples, you may have noticed that we talked about photos, tweets, videos, SMS, etc. Various types of data now exist, and are very different from each other. Big Data is therefore much more than just a “huge amount of data”.

In other words, we cannot understand Big Data without understanding its variety, linked to these two notions:

• – the variety of data types;
• – the data structure.

Yes! Photos, tweets, videos, emails, etc. are also data, but in different forms. In addition, the data in Big Data is of a largely unstructured nature, and cannot be collected, stored and analyzed in the old row/column database format. This means that each form of data requires a specific type of analysis that differs from the other forms.

If you take an e-mail, for example, you will realize that no content in an e-mail is identical to any of the thousands or even millions of other e-mails sent. Each e-mail consists of different addresses of the sender, sent to different destinations at different times. Each contains a message (text), and possibly, attachments (documents, photos, etc.).

This diversity doesn’t just involve the variety of forms of data produced, but also the variety of sources from which these data come from. This variety, as you can see, represents the second property that joins the first one, in order to help you understand how different the nature of data in Big Data is from the one we knew.

5.2.3. Velocity

Volume and variety are important, but if you really want to understand the context of Big Data, you must pay particular attention to the velocity aspect, which is just as important as the two previous ones. Just the same as data collection, processing and analysis must be carried out in real-time.

The Internet and the various connected objects considerably accelerate the generation of large amounts of data (Maheshwari 2019), from e-mails to shared photos on social media, the number of videos viewed or downloaded, and of course, the data from geolocation systems. Data can move quickly.

Just think of our Facebook example and the 250 billion photos stored last year. Remember that it was clearly indicated that in 2018, every minute, Facebook users posted more than 250,000 photos online. Yes! Every minute. This means that Facebook must manage more than 15 million photos every hour, which means that it must process and classify more than 360 million photos per day to facilitate their recovery and reuse.

In this context, velocity therefore consists of measuring the speed with which data are produced, analyzed and stored. However, it should be noted that there is no absolute rule, such as the number of bytes, for example, that we could consider as a threshold to define the velocity of data.

5.2.4. What else?

When we hear the word Big Data, we can only think of the three characteristics, often used to define this phenomenon. In other words, the 3 Vs: volume, velocity and variety. Its definition is most often based on these characteristics. But, while this model is certainly important and correct, it is now time to add other crucial factors. And you know what? They all start with the letter V.

Definitions given by words starting with Vs have become so classic that practitioners want to explain every aspect of Big Data with a new V, such as:

• – value: which refers to the benefits of Big Data, that can be obtained through appropriate analysis;
• – veracity: the reliability of the data source, its context and its importance for the resulting analysis;
• – variability: or the presence of inconsistencies in the data;
• – validity: or the extent to which the data are accurate and correct;
• – volatility: describes how long it will take for data to be available and how long it should be stored;
• – visualization: the manner in which the results of the data processing (information) are presented, in order to ensure greater clarity;
• – viability, vulnerability, and many others.

Thus, there are several characteristics to help us make data useful and generate value. Each characteristic plays an important role in enriching the context of Big Data. All these additional Vs illustrate other challenges in extracting value from unconventional data-sets.

Veracity, variability and volatility, for example, refer to the problem that Big Data often consists of unguaranteed accurate data, irregularities in data-sets, an uneven data flow and complicated navigation between components.

Veracity, which often refers to data quality, is an important aspect of Big Data, because not only does data come from everywhere, but it also belongs to everyone. Visualizations are also necessary to make sense of the results.

With all these characteristics, defining Big Data is not so easy, because the term itself refers to many aspects and new features. Certainly, these features are all important. But the most important thing is to understand how to generate value that corresponds to the key of obtaining useful information and better conduct the decision-making process. This value is only possible by analyzing huge amounts of data (volume) from different sources (variety) in real-time (velocity).

5.3.1. Big Data versus Big Computing

Being able to process and analyze the large volume of data, available in different formats and come from different sources, is another answer to the previous questions. This is mainly due to the increase in computing power. Big Data therefore refers to data-sets that are so large and complex that traditional data processing tools cannot support them.

Understanding the 3 Vs – volume, variety and velocity – is an essential element in understanding the Big Data universe, but it is far from over. We do agree that when a size of data is produced in real-time and arrives in continuous flow from multiple sources, it qualifies as Big Data.

However, to transform this quantity into value, we need large amount of computing and at a lower cost, so as to examine and process this data.

So “big” is, of course, a term relating to the volume, variety and velocity of data, but more importantly, it is also a term relating to the IT infrastructure that is in place. Because this “big” volume also induces great analyses.

The principles of statistics, forecasting, modeling and optimization remain the same. It is this computing capacity that has the potential to monetize data. Today, we can run billions of simulations thanks to the advancements and progress of IT tools, which tend to focus on Big Data technologies. There are now several tools available for solving problems, determining models and identifying opportunities.

Thus, the growth in computing power is the true change that has opened the door to great opportunities.

5.3.2. Big Data storage

Nowadays, companies focus so much on data analysis and processing that they often forget the need for data storage solutions. They are mainly interested in how they will be able to transform all the data collected into value.

However, the data accumulated does need a space where they will be stored. This means a necessary infrastructure that allows the storage of large amounts of data.

In 2011, the McKinsey Institute proposed its own version of the term Big Data. This term is defined by this institute as:

Data whose scale, diversity and temporal distribution require new technical architectures and more advanced analyses in order to extract knowledge that represents a new source of value. (Manyika et al. 2011)

If you look at the large number in a single zettabyte, you will realize that growth is not only in computing capacity, but also in storage capacity (Table 5.2).

Technology is constantly evolving and, as a result, machines and connected objects are consuming more and more data. Table 5.2 shows that in 2018, 50 000 x 10⁹ bytes of data were created per second, thus billions of bytes were created daily. The volume of data will only continue to grow very significantly.

It should also be noted that the quantity mentioned in the previous definition may vary from one company to another (small or large) and from one business sector to another (trade, industry, service). Of course, it can also vary depending on the analytical technologies used and the size of the databases available in a particular company or sector.

This is why, in many companies, or in many sectors, the amount of data ranges from a few dozen terabytes to several petabytes, which represent thousands of terabytes (10¹⁵ bytes).

Big Data generally introduces data-sets that have a volume that exceeds the capacity of the IT tools commonly used to capture, manage and process said volume in real-time. Therefore, it is essential that large data storage systems evolve.

5.3.3. Updating Moore’s Law

A simple computation of all the data on our physical activities that are recorded through connected objects (smartphones, smart watches, etc.), such as our heart rates for example, can not only save our lives, but also those of thousands of people. Here is just a simple example of the power of data.

The data generated on a person who jogs for 20 to 30 minutes in one day may seem like little to you. But imagine the amount of data that will be generated by calculating the pulses of a billion people who own smart watches or applications on smartphones? Now imagine the quantity produced per week, per month and per year.

Every sharing or comment on Facebook or Instagram, every video you watch on YouTube, is data. With billions of subscribers, data will continue to grow exponentially. This exponential growth was predicted more than 50 years ago by Gordon Moore.

In his 1965 article “Cramming More Components onto Integrated Circuits”, Gordon Moore indicated that: “The number of transistors per circuit of the same size doubled every eighteen months, with no increase in costs” (Moore 1965).

In his article, Moore noticed that the number of transistors on a chip had almost doubled each year from 1959 to 1965. The article also stated Moore’s law for the first time, according to which the number of transistors on a chip would roughly double every two years.

Table 5.2 shows that processor performance is increasing exponentially, as if hard disks were applying their own version of Moore’s law. This is the case because computers and various connected objects nowadays are much more powerful and smaller than those used 10 or 20 years ago.

This law certainly has an effect on data because the placement of sensors and smart interconnected objects everywhere increases their dissemination. And the bigger this dissemination is, the more the quantity of data produced by these objects increases. As the amount of data increases, so does the need to improve computing and storage capacity at a low cost. The lower the price, the more the dissemination of smart objects and sensors increases, and so on.

Do you see? It is an endless loop, which leads to the motivation of using and improving machine capacities. We can then say that Big Data pushes the limits of Moore’s law. Big Data has brought new ways of doing things and new methods of collecting, analyzing, integrating and visualizing data.

As processing power increases and storage capacity increases, it will be necessary to examine new computing fields and rely on a large number of essential resources, including the cloud, which is one of the main means by which computing can continue to grow.

Cloud computing, which has attracted considerable attention in recent years, is seen as an infrastructure that has revolutionized data storage and computing capabilities. These capabilities have led to advancements in the field of computing, which have improved the power of processing and analysis. This has enabled companies to be more efficient and introduce innovative solutions.

In parallel with these advances, Big Data technologies, which heavily rely on cloud platforms for storing and processing data flows, have been widely developed. They are among the most frequently used technologies in the development of applications or services in various sectors (health, education, energy, web, etc.).

5.4.1. The decision-making process and the dynamics of value creation

The different Vs of Big Data challenge the fundamental principles of existing technical approaches and require, as already mentioned, new forms of processing that promote decision-making, knowledge extraction and operations optimization (Curry 2016; Sedkaoui 2018b).

It is well known that the decision-making process is often based on the model of limited rationality: intelligence, modeling, choice and control, forged by Herbert Simon in 1977. With the varying amounts of data produced every second, this model becomes complex and needs to be improved (Sedkaoui 2018a). Indeed, decision-making, strategy development and anticipation of change have always been dependent on the availability and quality of data. Information technology experts create algorithms to better manipulate and organize data. These experts are supported by data science experts who are responsible for the initiation, development and application of quantitative methodology.

They collect, store and analyze data, using software and programs, applying the necessary analyses (data analysis algorithms) to generate models or information that subject matter experts, or decision-makers, interpret in order to make strategic decisions and create value.

All these steps will be part of a data value chain, from collection to decision-making. With, of course, the contribution of the various stakeholders supported by the technologies used. The value chain therefore depends on the quantity and quality of the data that will be put to analysis.

The decision-making process depends on the process of creating meaning, in other words, the process of creating knowledge. In this respect, decisions are data-driven, that is to say, they contain an important analytical phase, based on the processing of structured or unstructured data, be it from internal (company databases) or external (web, social networks, etc.) sources.

A data-driven decision-making process is therefore a process that includes an analytics aspect. This aspect can help to identify decision-making opportunities in the intelligence phase of Simon’s model, where the term “intelligence” refers to the discovery and extraction of knowledge. In this particular case, this phase consists of identifying the opportunities for which a decision must be made (Simon 1997).

From the decision-maker’s perspective, the importance of the data analysis process lies in its ability to provide valuable information on which to base decisions.

It allows decision-makers to take advantage of opportunities resulting from the wealth of data generated by supply chains, production processes, customer behavior, etc. (Frankel and Reid 2008).

This process includes several distinct phases, which we will discuss in more detail in the next chapter.

5.4.2. The emergence of new data-driven business models

When technologies become cheaper and easier to use, they transform businesses. This is the case with Big Data technologies, with a substantial reduction in data processing and storage costs.

The range of opportunities offered by these technologies allows companies to transform their business model through new visions of the value chain into a vast, fully digitalized ecosystem. This involves new ways of operating and has led to a reassessment of the foundations of business and an interpretation of new ways of creating value.

This is what Sergey Brin and Larry Page of Google, Jeff Bezos of Amazon, Jack Ma of Alibaba.com, Travis Kalanick and Garrett Camp of Uber, Reed Hastings of Netflix, Brian Chesky, Joe Gebbia and Nathan Blecharczyk of Airbnb, Mazzella of BlaBlaCar, and many others, have done.

You may be wondering how they did it. The answer is simple, they took full advantage of the opportunities of Big Data analytics by deciphering the underlying messages revealed by each data byte. Of course, these innovative models have also understood the importance of the digital ecosystem and digital platforms.

Their success depends on their resistance to change and their ability to monetize it in their “ROI” (Return on Investment). They have exploited the potential of analytics, not just to differentiate their business models, but also to innovate. And innovating in a business model means exploring new ideas, creating new value propositions and setting up new value chains. It is therefore a question of innovating, but in a different manner.

As Drucker (1994) indicated, it is about how to make a difference. This is what the business model concept itself means. The concept refers to the technique that companies adopt in the face of competition and change, based on the skills and resources available (McKelvey and Zhu 2013). The business model therefore describes how a company can create, deliver and capitalize value.

Value creation, for the benefit of all stakeholders in the business context, is the main element of a company’s success. However, in a generalized digitization ecosystem, be aware that evaluating and evolving a business model is not an option, because everything changes quickly and only the most agile companies will resist.

But when a company makes a strategic decision to develop its project, data is always useful. The experiences cited best illustrate this point. If these companies, among others, managed to surprise us, it is because they set up a series of innovative business models oriented and based on data and analytics, or what is called the data-driven business model.

Creating a model that is oriented and driven by data means making decisions based on the analysis of that data. For each company, or start-up, that designs and implements what constitutes a data-driven business model, which has changed the business landscape, there are several experiences that can serve as a reference to capture new sources of revenue.