7.2.1. Big Data on Walmart’s shelves

When we talk about big data, it’s hard not to think of the example (diapers and beer) often used to describe the importance of analyzing massive amounts of data. Of course, this is about Walmart and its experience in data analysis in the retail sector.

With tons of petabytes of usually unstructured and constantly generated data, Walmart was able, through the exploitation of the potential of Analytics, to optimize its distribution process. All Walmart decision-making processes rely on data extraction and the analysis of customer behavior and product inventories.

Whether internal or external, the analysis of petabytes of the company’s data has optimized its various operations and identified all sort of correlations in order to understand customers and anticipate their needs.

Let’s return to the example of diapers and beer. After learning that young men often buy diapers and beer between 5 pm and 7 pm, a simple reorganization of the stores led to an explosion in the sales of these products (Sedkaoui 2018b). This real-time correlation system, operating with a constant flow of data, has enabled Walmart to increase sales by simply placing these two products closer to each other.

Another approach that enabled Walmart to strengthen its decision-making process is that the e-commerce giant uses data from social networks like Facebook, Twitter, etc. Many of its decisions focus on the analysis of this data to generate useful information.

Walmart also exploits large quantities of data to support its recommendation system. For this, the company has developed an application that provides links to the products so that customers can have easy access. The app is called Shopycat, and it can recommend gifts for friends based on data from social networks extracted from their Facebook profiles.

In this context, Big Data analytics is an opportunity for retail to seize in order to provide an increasingly rich and innovative customer experience.

7.2.2. The Big Data behind Netflix’s success story

Why Netflix? Because this company, founded in 1997, analyzes everything we watch and interprets the results to derive trends for future productions.

It began when the BBC broadcast the series House of Cards in 1990. Wondering how Netflix has grown so fast? Simply by analyzing the tastes of subscribers who liked the first version of this series. Netflix found that subscribers also watched other films with Kevin Spacey and/or directed by David Fincher.

Netflix is largely based on a thorough knowledge of its subscribers as a constant source of service improvement. This knowledge enables it to offer real-time personalized recommendations, contrary to what was happening just a few years ago, when it was necessary to go out to buy a DVD to watch.

Netflix has tapped into the potential of analyzing data in different ways, and Table 7.1 summarizes some applications of Big Data.

Netflix has therefore upset the systems for personalization and recommendation, like Facebook, Amazon, and others. By clicking on “watch”, the data analysis process makes it possible to visualize the diversity of a subscriber’s choices and to later recommend movies or series.

Yes! This is because, in order to access content from Netflix, the subscriber must have a personal account that contains personal information. This data is used by the Netflix system. In addition to this data, each subscriber leaves behind tracks that this system can follow.

Everyone watches movies or series based on their taste, which is identifiable. But people also watch movies or series based on their availability or the date and time, and they will watch a certain number of times. Netflix collects this data and combines it with sociological characteristics and dozens of other criteria. Once the data is collected, it is analyzed and cross-referenced for groups of users (subscribers) with different characteristics in order to recommend movies or series to them.

7.2.3. The Amazon version of Big Data

If we only visit an e-commerce for a specific product, our interests cannot be clearly defined. Accordingly, data from news sites and other social networks will be analyzed to better understand what we are interested according to the information and research on the web, in order to subsequently recommend articles tailored to us.

That’s what Amazon does in order to suggest products to us. Like Walmart, Amazon uses data from Facebook, Twitter, Google, etc. to refine its targeting and to better understand our interests. Amazon is also another example that illustrates the importance and potential of the analysis of large amounts of data in real-time.

Thanks to the extremely advanced data analysis algorithms that Amazon has developed to anticipate our evolving needs, the company can suggest products that fit our needs by analyzing the traces we have left during our last online session.

Now in use for more than 20 years, the innovations achieved by Jeff Bezos’ company are strongly linked to the exploitation and analysis of data; and these innovations are still in progress, with Amazon Prime Now, launched in June 2016, and Amazon Pantry, launched in March 2017.

Amazon Web Services (AWS) is one example. AWS wants to leverage Amazon’s experience in analyzing advanced data, artificial intelligence, etc. to simplify the construction of models and to accelerate learning in order to make Machine Learning accessible. Amazon Echo is a step in this direction, and now no one has to move to shop, because a simple conversation with the Alexa’s virtual assistant will do the job.

As the foundation of the Echo intelligent home assistant, Alexa tends to become indispensable. This assistant is integrated directly into the functions of washing machines, refrigerators, vacuum cleaners, TVs, etc. These devices respond to more or less sophisticated voice commands for communicating directly with the Alexa assistant.

The objective of creating such an assistant is to facilitate the use of devices, such as starting or stopping a wash cycle, checking and setting the refrigerator temperature, starting a recording on the connected TV, etc. With this assistant, the customer will certainly be satisfied, since Amazon’s goal is symbolized by its logo: a smile.

7.2.4. Big data and social networks: the case of Facebook

Have you ever shared videos that Facebook has suggested to you? Do you know what we’re talking about? We’re referring to the “flashback”: the short preview of your posts, “likes”, pictures, etc. like those you may have received to celebrate your years on Facebook, or when Facebook suggests celebrating the anniversary of when you “became friends” with someone, etc.?

If you have ever had these experiences, you have participated in one of the examples of Big Data applications used by this social network. These examples, which retrace your activity since you registered on the network, illustrate one of the ways in which Facebook leverages the power of data.

Facebook’s data analysis is not limited only to data generated from your News Feed or your sharing activities, “likes”, etc. Rather, the data collection network can incorporate everything from your mailing address to the battery power on your smartphone. Data is transferred daily to our smartphones and other connected objects.

Are you having trouble imagining this? Nevertheless, Facebook can even follow you on the web using tracking cookies.

So, if you log into your Facebook account and you browse other sites (music, online sales, etc.) at the same time, know that Facebook follows you and knows which sites you’ve visited. Using this data, Facebook shows specific ads that may interest you.

Thus, Facebook can use the photos you’ve shared or liked to track you on the Internet and on other Facebook profiles. Using facial recognition capabilities and image processing, Facebook analyzes these photos, provided by sharing, to track you.

This is the Facebook version of Big Brother is watching you, the famous sentence from George Orwell’s 1984, published in 1949.

You’ll understand that using the example of Facebook should not be surprising, since this social network is considered one of the largest suppliers of Big Data, to which it owes its success in part. So even if the largest social network is essentially free for users, Facebook profits by analyzing its users’ data.

7.2.5. IBM and data analysis in the health sector

As in all areas, the health sector as a whole is gradually being overwhelmed by the 3 Vs phenomenon. Through the collection and analysis of large amounts of data, the use of different techniques for prevention, treatment, diagnosis, and monitoring of patients is changing rapidly (Sedkaoui 2018a).

In this sector, the analysis of large data-sets can also extract relevant information from the medical history of each patient. Big Data analytics facilitates the collection and analysis of both structured and unstructured data in real-time from a growing number of sources, including medical records, electronic surveillance systems, and others.

For this, IBM has developed a program, IBM Watson, that is able to store, understand, cross-reference, and exploit medical data to diagnose and analyze patients’ health. Looking through newspaper articles, tweets, and blog posts, “IBM Watson” seems to be able to provide powerful features to the health sector.

IBM was able to, as a result, extract the notes taken by doctors during consultations to diagnose heart failure, and to develop an algorithm that summarizes the text using a technique called Natural Language Processing (NLP).

Like a cardiologist, a computer can now read the doctor’s notes and determine if a patient has heart failure (Sedkaoui 2018a). Watson can also make custom recommendations by reading the patient’s genome.

Watson is positioned as a health cloud, aiming to create a real-time ecosystem for cross-referencing (anonymized) patient data.

Impressive, isn’t it? How have data analysis algorithms allowed Amazon and other companies achieve this success? After all, Big Data is just a large data-set from various sources, in different formats, which can be processed and analyzed appropriately.

That said, it’s easy to see that, since its creation in 1994, Amazon has built a culture based on data and its analysis. The key factor is knowing how this data can be correlated and formatted to make sense of it through analysis.

Emerging companies in the context of the sharing economy use digital platforms to connect people around the world and also to collect data and analyze it. In what follows, we will see how big data and data analysis capabilities helped create the context of this economy and boost its business models.

It is this question, which you may have been asking yourself since page one, that this book seeks to address through its various chapters, which began by immersing you in the context of the sharing economy, showing you the power of Big Data analytics, and now intends to help you understand the link between the two.

7.3.1. Data and platforms at the heart of the sharing economy

Because of the emphasis on sharing and collaboration, the activities of the sharing economy are often seen as very different from those provided by private and public channels. The different models in this economy have provided sources of additional or alternative income. Basically, the sharing economy is a specific business activity or a new type of business model (OECD 2016).

It is significant that companies like Airbnb and Uber were both founded, in 2008 and 2009 respectively, during the financial crisis. As such, this economy is portrayed as a way to present socioeconomic values through new forms of consumption, production, cooperation, participation, and sharing.

Indeed, the idea of sharing goods and services is not new; in fact, it is an ancient phenomenon. For example, community libraries allowed people in the eighteenth century to lend books to and borrow them from their neighbors. But the degree of sharing was very limited in the past, both because of the difficulty of matching supply and demand, but also because of the lack of trust between people.

However, the emergence of the Internet and social media and the rise of digital technologies and connected objects helped to fill these gaps. The term “sharing” has expanded (John 2017) and has become one of the fundamental cultural values of the digital environment (Stadler and Stülzl 2011). This has therefore launched a movement to make the most of these technologies, especially digital platforms, and to work towards the benefit of all.

Of course, to rent a room on Airbnb or to offer available seats on BlaBlaCar or to become an Uber driver, you need a digital platform. To connect the owner of a source (good or service) to a customer, you need a space that connects the two parts and that allows them to gain an advantage: everything must be “win/win”.

For example, the idea of renting your room, which is located in Paris, for a week to a person from London seems to be a very smart idea. Not only will the property be used for the intended purpose, since you’re offering up a property you’re not using to someone who actually needs it. But, furthermore, this use will generate additional income.

For that, the person from London who wants to rent a room needs a simple tool; namely, a digital platform that lists all available properties. Each property must list information about its defining characteristics. In other words, the exact location – the address in Paris, the price, facilities, etc.

You, in Paris, also need this platform to reach a large number of interested customers. This platform allows you to specify the characteristics of your apartment or property (address, price, etc.). By communicating accurate information through this platform, users will be able to find your property and communicate easily with you. In return, this property will earn you money.

We are clearly seeing a connection between supply (offering a room in Paris) and demand (looking for a room in Paris), by providing participants with the necessary information. This connection is therefore based on profit, but also on mutual trust, since the transaction and the consumption rely on the relationship between these people via digital platforms, where information is transparent and symmetrical.

In this context, the idea of sharing requires three elements that foster a sense of social cooperation and integrity:

• – mutual understanding;
• – goodwill;
• – honesty.

Several companies have developed their own platforms based on these three elements to allow owners of goods and services and users of these goods and services to connect for the benefit of both. Uber, Airbnb, and BlaBlaCar are obvious examples, and they represent platforms among many other platforms, too numerous to mention.

Thousands of homeowners put the characteristics of their apartments and properties online to attract users’ attention. But what do these characteristics really represent? In other words, when you create a full online reputation profile, you’re not just enabling communications and safer transactions on these platforms; you’re sharing something else. What are you sharing?

The answer to this question can only be massive amounts of data. Particularly of unstructured data generated by thousands of users on the sharing economy’s many platforms. Data is therefore at the center of this movement and of all these different categories (P2P, B2B, B2C, C2C, etc.). And companies that do business in this economy should not only think about ways of doing business or creating new services, but also about effectively using large amounts of data to deliver goods and services to people who want it.

It’s important to see that data – like the platforms that are at the heart of the sharing economy, and that, unlike companies based on an economy without sharing, such as search engines – is very valuable to the platforms’ fundamental activities.

7.3.2. The data of sharing economy companies

You can see from the previous sections that the emergence of the sharing economy and its different categories influences not only demand but also the entire value chain. Its impact on different parts of the value chain, including production, logistics, product design, and supply management, becomes clear. This is enhanced by increased computer technology, the Internet of Things (IoT) and blockchain.

For sharing economy companies, the data that is captured, stored, and analyzed along this chain is indicative of value creation. This means that, in order to meet the needs of owners and users of different platforms, companies like Uber, Airbnb, BlaBlaCar, etc., need to manage all of their data in real-time.

The question now is: what types of data must these firms manage and analyze in order to generate value? And what are the different challenges that these companies face?

To answer these questions, we need to dive back into Chapter 5 of this book, where we showed that, just as these data can have different units of measurement (they are now being measured in petabytes, exabytes, and zettabytes), they also occur in a variety of forms and can be classified into three categories (Table 7.2).

In the context of the sharing economy, we can cross-reference these three categories of data (structured, semi-structured, and unstructured) by analyzing the large quantities that are produced. But the most relevant data for companies in the sharing economy is relational data (Smichowski 2016) that reflects the way users interact on a platform.

In this framework, we distinguish three types of data, without neglecting the type already illustrated in Table 7.2:

• – data provided directly by a user, such as profile information, photos, contact lists, etc.;
• – data on user behavior from a platform or browser;
• – data generated by the analysis of the previous two data types.

Managing these different types of data emanating from the sharing economy’s platforms favors cooperativism, which can maximize the economic and societal effects, etc.

But in the sharing economy, the first type of data (provided by a user) is highly relational. That is to say that if, for example, a person X (from London) rents the house of another person Y (in Paris) using a platform, this data refers to these two persons. This may also be the case with some data provided, such as photos where several people appear or a contact list, etc.

This poses a problem because almost all individual users must agree that the data can be used. But the context of big data involves more than these two persons (X and Y), as in our example. If fact, the opposite is true, because Big Data analytics is only useful when large amounts of data are analyzed. What counts is the use of data, which involves millions or even billions of units of different types of data.

So, different people use sharing platforms that each gather large amounts of personal data about consumers, potentially creating challenges for protecting this data.

In the case of data-driven services, as provided by most sharing economy platforms, the problem lies not only in the sharing, but also in the data, or in a sufficient amount of data, and consists of how to exploit them.

7.3.3. Privacy and data security in a sharing economy

We can agree that when we use platforms such as Airbnb, Uber, Lyft, BlaBlaCar, and others, we are not only offering or sharing our properties, or looking for specific goods or services; we are also sharing our personal data.

Yes! Our apartments’ addresses, the facilities available, our credit card information when we make transactions, our destinations, availability, etc. We are communicating this data on sharing economy services platforms.

Are you aware that we are providing this data to other users via some companies’ platforms – which have no vehicles and have not recruited a single driver (Uber, BlaBlaCar), which have no houses, apartments, or hotel rooms (Airbnb) – which do not have the goods and services that we are sharing through their platforms? These platforms are designed for a specific purpose – driving, accommodation, self-employment – from X to Y.

Via these platforms, X and Y will:

• – create profiles;
• – describe their property (goods or services);
• – search in lists based on predefined criteria;
• – book rooms or apartments in advance;
• – make payments safely.

These platforms are therefore digital spaces where we can find or develop several functions at once, all while providing the necessary data.

From an ethical point of view, each of these functions poses one or more problems related to the protection of participants’ privacy.

Although these platforms allow users easy access to the sharing of goods and services with a single click on their connected devices and objects, it is important to examine some questions and requirements that seem to be neglected and that are related to data security and confidentiality.

Different types of data described previously must be protected in the context of the sharing. To explain this, let’s refer to our example. The example of X, who has a room in Paris, and Y, who is looking for one in the neighborhood.

In this case, X will, at first, create a profile, on the Airbnb platform, for example, which incorporates some basic details such as name, username, etc., which must be visible to all. This type of data can be stored in a distributed storage system in unencrypted format.

To describe his or her property, X must add additional information to the profile that will be publicly visible, such as the description of his or her apartment, photos, etc.

This description contains data related to the apartment’s address: the neighborhood where the apartment is located, its proximity to public transportation, supermarkets, tourist attractions, parking, etc. This information is very helpful to potential client Y, who can easily access these details.

X can also set the apartment’s availability. But Y has no way to access this information (the apartment’s availability), which can be only be managed by X. X must update this information in order to approve new requests from other clients.

Notice also that X and Y can communicate securely in private. Once Y confirms the reservation, this decision is only shared with X. The payment details between the two parties remains anonymous and must be fully secured. Obviously, this security also concerns Y’s trip, because he or she doesn’t want this information to be visible to other people.

However, some metadata related to booking and payment can be published to show that X actually has the number of people it claims to have. On the other hand, Y will have temporary access to data about the apartment (the exact address, door codes, etc.). After his or her stay, they receive a notification to provide his opinion, which future clients will be able to see.

The example shows that there is a range of requirements for data privacy via shared platforms. This concerns not only the Airbnb platform, but all sharing economy platforms.

The previous example allows us to see that the following two statements highlight a dual aspect (Ranzini et al. 2017):

• – The first concerns the security of the data exchanged between users and the companies that created these platforms. In return for the security of their data, these users are able to participate on these platforms. But it should be noted here that these platforms are reluctant to share data from their users, which is important for better estimating the impact of the sharing economy (Frenken and Schor 2017). This can harm the platforms themselves, by limiting the potential size of the sharing economy.
• – The second is related to the security of the data that users exchange with each other in order to access goods or services. The creation of sharing platforms thus raises privacy protection issues to the extent that they involve not only the sharing of goods and services but also the sharing of data by simply using the platforms.

The emergence of this economy has raised challenges concerning the private lives of participants that involve the sharing of their data. It would take a whole book, or even a series of books, to analyze these points in detail. But we wanted, through this short discussion, to draw your attention to confidentiality and data security, which are two elements to be considered in the context of the sharing economy.

7.3.4. Open Data and platform data sharing

In the context of Big Data, the challenges faced by businesses in the sharing economy extend beyond privacy and security. Other challenges, such as problems related to the complexity, scalability, heterogeneity, quality, and timeliness of data, are also a major concern (Table 7.3).

These problems should be taken seriously by companies in the sharing economy during the analysis of massive data-sets and the development of their analytical process. This must be done in advance, even before building a data-driven approach.

But beyond all of that, another very interesting point we wanted to look at here is sharing and ease of access to the large data-sets that are generated. We believe that, contrary to what its name suggests, the sharing economy, or, more precisely, this economy’s platforms, do not consider the value of sharing their users’ data.

Collaborating and making data open will result in greater value for all stakeholders. But, unfortunately, this is not the case.

We need data to build trust with other users and to improve our reputations. Enhancing this reputation can be presented as a factor motivating users to interact in a community (Wasko and Faraj 2005) and as a commitment based on trust between users (Utz et al. 2012).

In other words, we need to communicate with trustworthy people. Because, even if the platform allows us to connect with other users, no one is willing to connect with someone who cannot be trusted. Since data quality depends on the amount of available data, the user needs to access data from other users of the platform.

By connecting to these virtual spaces where billions of people cooperate and work independently (Rifkin 2014), we transmit large amounts of data. Harnessing this data can increase the benefits of the sharing economy, and users and participants can improve their reputations by sharing information.

It is clear that the economic benefit of the sharing economy lies in the efficient use of resources, resulting in reduced transaction costs (Lobel 2018). But sharing data would provide access to large data-sets, which would open up many opportunities to improve and operationalize various practices in the sharing economy and to create new business models.

To exploit the potential of large data-sets, data must be shared, and its reuse must be allowed. Building a data-driven culture depends on easy access to this data as a value-generating resource. However, as mentioned previously, companies always seem reluctant to share their data.

One of the main responses raised by private platforms is that their competitors could access their data, because there is no single platform supporting transparent data sharing. Therefore, the key question is how this sharing can be encouraged.

Maybe you’ll think of something that would boost the culture of data sharing. For example, helping businesses understand the value of their data, or creating a platform for sharing data-sets in a transparent manner, etc.

But the more reasonable approach is to focus on the platforms. The different sharing economy platforms are far from being considered as simple mechanisms for data access. They represent a third party that enables a systematic exchange of data flowing between users.

As we noted at the beginning of this chapter, these platforms exist, as does data, at the heart of the sharing economy’s various practices. They represent the foundation of users’ activities by giving them access in exchange for providing data (Srnicek 2017).

It is time therefore to rethink data sharing and openness for effective reuse for the economy in general. This means that the techniques and algorithms for data analysis are essential for better leveraging the flow of data.

In this context, companies like Uber, Airbnb, BlaBlaCar, Lyft, and others must follow an approach based not only on data but also on making it available, or Open Data. This is very important for drawing more benefit from the third-party services (platforms) and data reuse.