The next phase would be, in our view, a digital mobility service. This new product would cover any type of movement of a customer (e. g. with own car, bike, walking, taxi, shared car, bus, etc.), with insurance coverage applied seamlessly and transparently (e. g., automatically at any time, or for a desired period of time, at the click of a button in an app, etc.). The advantage for the customer is that a mobility insurance can be adapted to the usage of any transportation means where she would like to be protected. The new customer would be sharing critical information (such as spatial data) just for the time the insurance is required. Big Data analytics would create models of adaptive risk prediction to estimate the best price for the current trajectory of the customer at a limited time span.
More interestingly, peer‐to‐peer services between costumers can be added in order to create a network to complement any sort of digital mobility journey. This can be used to share information on predicted dangerous routes, traffic conditions, or free parking lots which might also be shared or handed over between participants in the network. Other data‐driven insurance models can evolve around crowd insurance, where a group of comparable customers are pooled so that they jointly cover their risks backed by pre‐paid premiums. If paid claims fall below of the pooled amount, then insurance becomes cheaper for the whole pool of customers. If claims exceed the pre‐paid premiums, coverage is limited to the available amount (back pay or re‐insurance could also be applied). This approach also leverages a fully digital business model, hence significantly reducing operational cost. The choice of pooled customers as well as their likelihood to be accepted for the crowd framework, would be controlled once more by advanced predictive models that can predict individual risk and the effect on the whole pool.
Ultimately, the move from selling insurance products to managing risk‐preventing and social enhancing customer journeys requires continuous testing and evaluation in a controlled environment. New business ideas can be now easily tested given the new data‐driven approach. Data helps understanding what works and what can be done better to continuously add value, improve, and refine journey innovation of customers.
Data‐driven businesses can only deliver value when data is shared and is made available. Hence, next generation Data Ecosystems support data sharing and data consumption/use without the need for data ownership. Data ecosystems refer to a highly heterogeneous data environment, where users generate data from any connected devices, from cars to smartphones and toasters. Usually, each device initially only sees its individual and isolated context, hence is relatively blind and dumb. Data and information from various devices and potentially across corporate platforms should be shared with the underlying ecosystem. The ecosystem can open up data for a certain purpose or service at a certain point in time or timeframe (based on security/privacy rules, see Privacy by Design). The data ecosystem becomes the engine and brain behind digital devices, providing them with contextual information and predictive services.
These possibilities in fact raise however very important issues of data protection, trust, and data ethics, such as e. g., who owns the data in a data ecosystem? How can data be shared ensuring data protection? How can businesses prove acceptable use of data to their customers, and how can customers be in control? In the next section we will approach this issue and argue for technology‐driven framework and Privacy by Design specifically.
59.4 Data Protection, Trust, and Data Ethics
As data and analytics become increasingly business relevant, governments, agencies, and public companies increasingly accumulate massive amounts of person‐related data and information. The nature of Big Data, however, is to store data beforehand for later use, not initially knowing what that would be. As such, from a consumer perspective, immense knowledge about people is aggregated, potentially with negative impact. The GDPR accounts for this by putting a bundle of regulations around the acquisition, storage, and use of personally relatable data—along with substantial fines. In turn, companies may pile up the next asbestos if end‐to‐end data processes are not fully compliant and controlled. Hence, data privacy and ethics beyond IT security will tremendously grow in importance as leveraging data on the one side will deliver significant competitive advantage and eventually better products and services, on the other hand puts an immediate risk on companies if they do not comply with future laws.
Data protection basically comprises security plus privacy. Security is a rather technical IT3‐related topic, focusing mainly on protecting physical, infrastructural, and software‐related topics, such as confidentiality (information is protected from unauthorized views), integrity (data is not changed or removed without rights), availability (services and data are available when needed), and non‐repudiation (proved traceability of data‐related transactions, ACID‐guaranteed) (see e. g., [10]). Security is a legal requirement and obligation. (Data) privacy, in contrast, is closely related to analytics and not IT. Focus areas are typically principles/guidelines: transparency, freedom of choice, consent, personal identifiable information (PII), data economy (data minimalism), prior stated purpose, necessity, direct inquiry, and appropriation. Privacy is a legal right, often a constitutional right.
Data ethics is far less tangible and considers even philosophical questions, such as “who is the owner of the data?” vs. “will there be data ownership at all4”?—the human being creating the data or the supplier of the technical device recording it? What are the ethical limits of analytics and potential actions triggered? How can data be shared and leveraged while respecting individual’s rights? Data ethics still is more in the grey area of feelings, opinions, and right treatment. That is also why people are willing to open up much more data for a company if they receive a useful service in return. Also, the potential (use, value, damage) of data given away by consumers is usually opaque and perceived very differently by individuals.
The European Union (EU) rightly treats Big Data and privacy as topic on their own and discusses a data protection regulation since the first legislative proposal of the Commission in 2012. Data anonymity (or the concept of personal identifiable information, PII) is no longer sufficient given the new Big Data possibilities. By enriching and blending previously anonymous data with external or historical data personal information can be eventually reconstructed easily [5]. Moreover, extremely rich information, such as personal profiles (socio‐economic, psychological, political etc.) can be created, leaving individuals unaware of this analytical process happening in the background—plus, given the issue of analytical veracity, results can easily be wrong or misleading. For example, credit risk ratings based on Facebook profiles and likes have been proven to be inaccurate [11].
Hence, data protection, privacy, and ethics are important for any organization that runs a business built on trust. Frameworks such as Privacy by Design [6] are a must. In our perspective the major principles of a Privacy by Design for data‐driven services should consider the following. Data sharing with the authorization of the owner of the data, for specific time frame and for specific purpose. In a world of data ecosystems, the creator of the data should be also considered to be the owner of the data. Only the owner decides with which business to share the data and purpose should be always clear. For example, in the insurance business, a customer could be sharing her movement data stored on her smartphone for the last few weeks in order to get an instantaneous quote to insure a trip in a shared car for the next few hours. Sharing of the data should be then consented and limited to a specific time frame.
Data owner consent and data limits. Businesses should have a clear and unambiguous consent from owners to data usage. Use of data should be limited to the purpose of the request; for example, for providing a quote on traveling insurance, data from health records would not be considered as necessary.
Not using data against customers in
hindsight. Data should be used for the only purpose established between the owner and the business at the moment of applying data services.
Confidentiality. Data shared by an owner to a business should remain confidential and not be shared with third parties, unless in agreement with the owner.
Necessity and proportionality. Business should only store and process data that is necessary to fulfil its services and deliver data‐driven products.
Implementing Privacy by Design principles requires the technical enablers to ensure the above principles. For example, sensitive data may be kept on a user device, in an encrypted form where the user holds the key; or data may be distributed across business entities where the user is the only one who can re‐compile the data through personal keys. The goal is to create the enabler to give control to the user, while still sensitive data resides somewhere in the business systems. In this respect, privacy engineering, e. g. [12], appears as an emerging discipline aiming at providing tools and techniques such that the engineered systems have good levels of privacy. Machine Learning research develops strongly on the area of “differential privacy” [13]. Intuitively, it requires that the mechanism outputting information about an underlying dataset is robust to any random noise change of one sample, thus protecting privacy.
Beyond the technical capabilities for protecting data privacy, the notion of data ethics should be grounded in business processes and organizations. Having clear privacy policies and enabling user consent via Privacy by Design/PET frameworks gives users control over their data and enable transparency that builds the needed bond of trust.
59.5 Summary and Conclusion
In this paper, we discuss the challenges of digitalization for traditional businesses. We argue that the ability to improve the customer experience and innovate by using the model of digital customer journeys is the most tangible immediate change to leverage the “big data” era. Big Data Analytics already leverages the tools to improve, fine‐tune, and automatize customer experience. Companies can therefore continue staying in the relevant set of consumers by delivering value, customer‐centric offerings, and creating positive impact on the life of people.
Going beyond that, new digital business models around Big Data ecosystems can excel with the help of advanced analytics. The ecosystem can technically open up data for a certain purposes or services and Machine Learning is there to understand the consumer needs based on data. The data ecosystem becomes the engine and brain behind digital devices, providing them with contextual information and predictive services. In this context, privacy and ethics is of outmost importance—for consumers but also for companies themselves. Privacy by Design and Privacy Enhancing Technologies require technical enablers to ensure basic privacy principles which should be implemented on a case‐by‐case basis. Privacy by Design principles should be implemented in any digital product/service and data ecosystem. This is an engineering topic to ensure that analytics can continue working on data for the best purpose of the consumer, but only under the consumer control.
References
1.
D. C. Edelman, “Branding in the digital age,” Harvard business review, pp. 62–69, 2010.
2.
D. Elzinga, S. Mulder and O. Vetvik, “The consumer decision journey,” McKinsey Quarterly, June 2009.
3.
D. Edelman and M. Singer, “The new consumer decision journey,” McKinsey Quarterly, September 2015.
4.
B. J. Pine II and J. H. Gilmore, “Welcome to the experience economy,” Harvard Business Review, July-August 1998.
5.
Y.-A. de Montjoye, C. Hidalgo, M. Verleysen and V. Blondel, “Unique in the Crowd: The privacy bounds of human mobility,” Scientific reports, 2013.
6.
P. Schaar, “Privacy by design,” Identity in the Information Society, pp. 267–274, 2010.
7.
A. Cavoukian and J. Jonas, “Privacy by design in the age of big data,” Information and Privacy Commissioner of Ontario, Ontario, Canada, 2012.
8.
E. K. Strong, The psychology of selling and advertising, McGraw-Hill book Company, 1925.
9.
F. F. Reichheld, “The one number you need to grow,” Harvard business review , pp. 46–55, 2003.
10.
D. Chen and H. Zhao, “Data Security and Privacy Protection Issues in Cloud Computing,” in Computer Science and Electronics Engineering (ICCSEE), 2012.Crossref
11.
Knowledge@Wharton, “The ‘Social’ Credit Score: Separating the Data from the Noise,” 5 June 2013. [Online].
12.
S. Shapiro, N. Washington, J. Miller, J. Snyder and J. McEwen, “Privacy Engineering Framework,” MITRE Privacy Community of Practice.
13.
C. Dwork, “Differential privacy,” in Encyclopedia of Cryptography and Security, Springer US, 2011, pp. 338–340.
Further Reading
14.
G. Coulouris, J. Dollimore and T. Kindberg, Distributed Systems – Concepts and Design, Amsterdam: Addison-Wesley Longman, 2005.MATH
15.
K. e. a. Beck, “Manifesto for agile software development,” http://www.agilemanifesto.org/.
Footnotes
1Extending on the “Experience Economy”, we believe that eventually the biggest impact of big data analytics lies beyond customer analytics. However, customer analytics and journeys are used here to explain to the current state and conflicts with data privacy topics.
2 Agile development is a paradigm in software engineering that embraces lightweight processes, quick iterations, prototyping/trial and error, and subsequent improvement iteration (Beck).
3We define IT as infrastructure, hardware, and basic software, such as operating systems; we exclude higher‐level layers, such as business applications, data, and analytics.
4Another way to see this is to remove the notion of ownership from data, focusing the discussion around data controllers and data processors.
© Springer-Verlag GmbH Germany 2018
Claudia Linnhoff-Popien, Ralf Schneider and Michael Zaddach (eds.)Digital Marketplaces Unleashedhttps://doi.org/10.1007/978-3-662-49275-8_60
60. On the Need of Opening the Big Data Landscape to Everyone: Challenges and New Trends
Rubén Salado-Cid1 , Aurora Ramírez1 and José Raúl Romero1
(1)University of Córdoba, Córdoba, Spain
Rubén Salado-Cid (Corresponding author)
Email: [email protected]
Aurora Ramírez
Email: [email protected]
José Raúl Romero
Email: [email protected]
60.1 Introduction
Many organizations around the world are massively generating and analyzing large amounts of data, but they still look for better approaches to get significant insight with the aim of achieving a leading position in the marketplace. According to a survey on information technologies (IT) and business leaders conducted by Gartner in 2015 [1], up to 75% of companies will be investing in Big Data over the next two years. Similarly, a recent report presented by Accenture [2] lays out that high‐performing companies are incorporating analytics to support decision‐making and decision processes. In this industrial scenario, important players in the global market agree that Big Data solutions are a competitive addition to companies as a key basis to increase productivity and innovation.
To meet the great demand for Big Data applications in the industry, a large number of technologies and techniques has emerged in the last few years, composing a wide and heterogeneous Big Data landscape. Annually, FirstMark Capital publishes an overview [3] of the most important Big Data technologies classified into different categories like infrastructure, analytics and applications. All these technologies make possible the development of Big Data solutions in a wide range of application domains, such as healthcare, man
ufacturing or marketing, where the analysis of large amounts of data is essential to discover relevant knowledge.
Nevertheless, technologies within the Big Data stack usually have a steep learning curve due to the required knowledge in diverse areas of the computing field like data mining, machine learning, software engineering or distributed computing [4], among others. Therefore, Big Data seems to be more commonly adopted in those domains whose companies can hire experts in knowledge discovery. Other times, developments need to be outsourced. Thus, companies in those sectors where data processing is highly demanded require suitable mechanisms and tools to facilitate, open and promote the adoption of Big Data analysis.
In this context, workflow technology brings a framework to conduct data analysis processes closer to the business expert, who has the domain‐specific knowledge but probably not the necessary computing skills. A workflow [5] is a high‐level mechanism to automate and describe processes as a set of activities that collaborate to produce a desired outcome. In general, workflows allow business experts to provide the definition of domain‐specific actions for their data analysis processes without specifying infrastructure requirements. The workflow automation is delegated to a workflow management system (WfMS) [6], which manages and efficiently executes the corresponding actions using all the available computational resources in the environment. Thus, business experts only need to focus on the representation of the domain‐specific specification, without requiring computational skills that are not related with their own application domains.
Digital Marketplaces Unleashed Page 91