Core Challenges and Conclusion

This is part 7 (and the final part) of my master thesis on Monetization Strategies and Business Models Behind Consumer Data. Find other chapters here:

4.6 Challenges for Data-Based Businesses

As soon as companies start implementing models that are based on consumer data as a core value stream, they will be faced by a multitude of non-technical challenges, no matter how attractive the value proposition towards the customer is. The following subchapters will briefly highlight a selection of potential hurdles for data-based businesses, namely, the cost of data, talent, culture and leadership, legal barriers, data security, and privacy.

4.6.1 Cost of Data

As this thesis has mainly dealt with the revenue generating parts of the business, a short insight into the costs of operating a data-based business is needed. In this case, focus will be laid on costs directly related to data, omitting costs for administration, customer service, HR and any overhead. It has to be assumed that these businesses do not build infrastructure and connectivity at a significant cost ($30,000 to $80,000 per mile), but operate on existing networks (Finley, 2015; Lupinacci, 2015).

Companies are shy about disclosing the cost of their data storage and analysis, making S-1 IPO paperwork a good source for some insight. Snapchat made public that is was spending $2 billion on Google Cloud services for computing, storage, bandwidth and further services over five years (Spiegel, 2017, p. 15) and an additional $1 billion on AWS (Amazon Web Services) for redundancy (Spiegel, 2017, p. 69). Bandwidth, computing and storage are listed as the primary factors of the cost of revenue - almost fourfold the amount spent on sales and marketing (Spiegel, 2017, pp. 71–73). A similar picture is painted at Dropbox, where the cost of revenue is 33% ($368.9 million - 2017) of overall revenue, with expenses “associated with the storage, delivery, and distribution” (Houston, 2018, p. 68) of the service as the primary driver.

Snapchat made public that is was spending $2 billion on Google Cloud services for computing, storage, bandwidth and further services over five years and an additional $1 billion on AWS

Although it is difficult to find the exact cost of data beyond the prices listed by Google (Google Cloud, 2018), Amazon (AWS, 2018) or Microsoft (Microsoft Azure, 2018), it is clear that bandwidth, storage and analysis are a significant part of the operating cost of data-based businesses. The cost of data for data-based businesses is also reflected in growth predictions of the leading cloud service providers (Microsoft, Amazon, IBM, Oracle, SAP and Google) all of which expect more than a 30% year-on-year revenue growth (Evans, 2018).

4.6.2 Talent, Culture and Leadership

In light of the substantial cost of storage and computing, companies should try to optimize their systems to the largest possible extent. In order to do so, data-based companies have to hire the best talent they can find, which has to have equally good data and business skills (Davenport & Patil, 2012; Katal, Wazid & Goudar, 2013, p. 407). Here, data-based businesses face similar challenges to every company with the goal of harnessing the power of data: identifying capable leaders that spot opportunities and set clear goals, effective talent management and hiring (McAfee & Brynjolfsson, 2012, p. 8).

As data-based companies will not only be using consumer data as part of their product offer, but will be data-focused in general, leaders have to be able to make decisions based on real-time flow-based data (rather than post-ex one-time analyses) (Davenport, Barth & Bean, 2012). In order to successfully create products that are based on data, leaders and employees alike have to understand and identify opportunities of data enhancement for consumer-value-creation. The cultural changes that arise as data science becomes ingrained in every part of the business, especially product development, might be the most significant challenge for current businesses aiming to transform themselves into data-based businesses (LaValle et al., 2011).

It is highly challenging to find people that have the ability to work with data in every part of an organization and every company will see itself in fierce competition with companies in related and unrelated industries (Meer, 2015). This challenge is also reflected in the low employee tenure of just under two years at top technology companies (Paysa, 2017) - which also reduces the chances of acquiring expertise and human capital through M&A, where employee retention is a crucial success factor (Ranft & Lord, 2000, pp. 312–313).

For companies to succeed in this environment and to build data-based products, utmost priority has to be given to talent acquisition and development, establishment of a flexible organizational culture based on meritocracy and enablement of employees to take decisions on all levels (Barton, Carey & Charan, 2018; Zuckerberg, 2018).

For companies to succeed in this environment and to build data-based products, utmost priority has to be given to talent acquisition and development, establishment of a flexible organizational culture based on meritocracy and enablement of employees to take decisions on all levels

4.6.3 Legal Barriers in Using Data

As data-based companies are continuously working with consumer data on an international scale, they have to comply with a multitude of laws and regulations in all local markets they operate in - most notably data- and consumer protection law, as outlined by Svantesson (2017). In addition to the international complexity of these laws, legislation is changing on an ongoing basis, as seen in the General Data Protection Regulation (GDPR) coming into force in the European Union (EU) on May 25, 2018 (Tankard, 2016) For the purpose of this chapter only EU legislation shall be considered.

It is apparent that companies are not allowed to collect, transfer and process consumer data without the user’s consent. German law even stipulates that a company enables a consumer to wholly understand the implication of consent in order to make an informed decision (Landgericht Berlin, 2018, pp. 32–33). Furthermore, considering that most users do not read privacy policies, the GDPR requires access to granular opt-in control of different data handling purposes. It further stipulates that companies give commonly understandable reasons why and how the data is used. (Information Commissioner's Office, 2018)

Should a company have the permission to collect and store data of an EU-citizen, the individual has the right to have this data deleted (“right to be forgotten” Mantelero, 2013) and to have it made fully transferable to a different service (“right of data portability” Hert & Papakonstantinou, 2016, p. 190). A detailed evaluation of the GDPR can be found at Hert and Papakonstantinou (2016), with a brief overview and comparison to US legislation at Butterworth (2018).

This chapter is not able to cover all the details of legislative rules and challenges that data-based companies will face, but it highlights that compliance with the law of the respective location is an essential part of the design of a data-based business model and all products reliant on consumer data.

4.6.4 Data Security

If a consumer entrusts a company with his or her data, one of the core challenges is to guarantee the security of this data - mainly protecting it against theft and breaches (Lafuente, 2015). Thus, data security is based on four pillars according to Denning (1982): encryption, access management, communication control and inference control.

Even the slightest glitch or inaccuracy in one of the four pillars can have substantial consequences as the nature of current data systems can lead to access to millions of accounts. The biggest data breach so far demonstrates the scale of the threat: Yahoo reported that up to 3 billion accounts were compromised in a hack that was based on a single Yahoo employee clicking a faulty link in 2014 (Armerding, 2018; Gallangher & Kravets, 2017).
Therefore, Heiser and Nicolett (2008, pp. 31–33) argue that modern data security is based upon seven new criteria: Privileged user access, regulatory compliance, data location, data segregation, recovery, investigative support and long-term viability. In order to successfully reach a satisfactory level in all these aspects, Chen and Zhao (2012, p. 651) analyzed the data-life-cycle and concluded that data security has to be embedded within the design of the data system rather than implemented as an afterthought.

Data security, as one of the core parts of the trust-based interaction with the consumer, is a crucial component for the success of a data-based business.

Data security, as one of the core parts of the trust-based interaction with the consumer, is a crucial component for the success of a data-based business. On the one hand, public attention to a security problem have a negative impact on the business (Facebook’s stock declined by 13% after an alleged data breach in 2018 Lim, 2018), but one the other hand, access to unique data can be one of the most valuable assets and a core part of the defendability of a product.

4.6.5 Privacy Concerns

Protection of privacy and a breach of trust are the pivotal issues here. The concept of privacy is complex, indistinct and ambiguous (Solove, 2009, pp. 1–2), but of increasing interest in public discussion (Cichy & Salge, 2015). Generally, informational privacy is described as the ability of an individual to control information about themselves (Smith, Dinev & Xu, 2011, p. 995).

Accommodating for privacy wishes of individuals is of high importance for data-based businesses, as access to consumer data is essential to better understand the individual person, which, in turn, is a core part of these models (see chapter 3.2 on p.22). This requires the highest possible amount of accurate externalized personal information - the most extreme form of disclosure, and the least privacy (Zwick & Dholakia, 2016, p. 35).

Companies can try to alleviate privacy concerns by providing accurate and understandable information about the usage of data, giving the user full control over data collection and storage, collecting the smallest amount of data necessary, and by delivering substantial value in return for the loss of privacy (Chen, 2017; Phelps, Nowak & Ferrell, 2000, p. 34; Prince, 2018). Fear about loss of privacy often stems from the difference between consumer’s expectation of the degree of privacy of their actions and the information that actually can be inferred from the data that is collected about them (McKenna, Richardson & Thomson, 2012, pp. 808–809).

But even if companies chose to be as open as possible about data collection and privacy, an individual might not always make a rational choice. Social, cultural and situational influences on self-disclosure and privacy preferences are a major part of privacy-related decision-making (Acquisti, Brandimarte & Loewenstein, 2015, pp. 509–511). As a consequence, companies should respect their responsibility for privacy protection and take into serious consideration that faulty assumptions about consumer preferences can have a long-term impact on business success.

companies should respect their responsibility for privacy protection and take into serious consideration that faulty assumptions about consumer preferences can have a long-term impact on business success.

The above-mentioned Facebook privacy discussion in 2018 is based on the decision of the company to make private information increasingly public, and only rolling back on that decision after a backlash from users (Mckeon, 2010). Although Facebook’s policy had minimal impact at the time, the alleged exploitation of this policy during national elections has caused a major public and political debate about Facebook’s business practices in general.

4.7 Outlook and Implications for Business

Despite a multitude of challenges, the outlook for companies pursuing data-based business models is favorable. The volume and variety of data produced is steadily rising, leading to an increase of information and knowledge that companies are able to extract from this source. As a consequence, the data models of consumers will become more accurate, leading to an even better understanding of customer needs and the possibility of creating more personalized products.

As the trend of a customer’s decreasing willingness to pay is seen as one of the core challenges, using consumer data as a basis for enabling a financial value creation presents a viable monetization strategy. In addition, cases of established companies (with revenue generating products) demonstrate that embedding data as a core value stream within the business model can increase the product-market fit. Thus, higher competitiveness can be achieved.

embedding data as a core value stream within the business model can increase the product-market fit

Data-based products enable a business to monetize their value delivery in a wide variety of ways, as there will not be a one-size-fits-all solution. Each company has to decide which of the three presented business models will allow for the most fitting value delivery to the consumer - and which monetization option is the most realistic. For those companies that are competing in markets where the penny gap is present, two-sided business models are the more attractive choice - keeping privacy in mind, a company should not see an individual merely as a data resource. Even though each individual product will have its own monetization- and business model, companies can combine the three conceptual models that are presented in this thesis.

The most advanced data-based businesses will develop models that are able to predict consumer behavior and deliver value based on prescriptive actions. This requires continuous improvement of the data model of a consumer through strategic data acquisition and consolidation. In addition, sophisticated real-time analytics systems based on basic statistical methods as well as ML and advanced AI will be necessary. This increasing knowledge about consumer actions and future behavior, however, also implies the responsibility of the company to make far-reaching decisions surrounding social reinforcement, sampling bias, political selectivity, time management in the attention economy, content exposure of individuals and many more ethical problems. Data-based businesses are likely to find themselves in an ethical dilemma where they have to choose between their social responsibility as a part of society and their short-term business interests - on a global scale.

The most advanced data-based businesses will develop models that are able to predict consumer behavior and deliver value based on prescriptive actions. This requires continuous improvement of the data model of a consumer through strategic data acquisition and consolidation.

Nonetheless, the major challenges that encompass data-based business models are related to the acquisition of knowledge (talent and culture) and questions surrounding data security and privacy. It is highly recommended that these aspects are not handled in a reactive manner, but are addressed proactively and solved within the design of the data-based business model. Overall, products based on consumer knowledge that is gained through the analysis of data will be a core factor for the success of all B2C companies.

5 Limitations and Research Opportunities

The limitations of this thesis present a wide variety of research opportunities. While this thesis used a pure case-based conceptual approach, the most salient extension is the introduction of a quantitative study on the topic of data-based business models. An approach of a quantitative study design can be found at Hartmann et al. (2016) where data-driven business models are analyzed from a dataset of startups.

In addition, the cases chosen for this thesis have a strong US-focus and the inclusion of international examples might lead to different findings. This is especially true following the idea that a consumer’s approach towards data disclosure has a significant cultural component (Acquisti et al., 2015). Thus, selecting examples from South- and East Asia with their influential and transformational online businesses can be of great interest.

With regard to the methodology used, the focus was deliberately laid on the transactional, flow-based approach of analyzing business models. Although this very abstract approach has been used before, success and failure of a business model often depend on details of the configuration within this model. Breaking down models on the basis of the business model canvas by Osterwalder and Pigneur (2010) can be a more accurate approach which would have exceeded the scale and scope of this thesis.

From a technical perspective, the topic of the data value chain and different methods of data analysis have to be discussed in more detail for the successful implementation of data-based business models. A research focus on ML and AI seems to be the most relevant approach, where companies will have to collect data in order to enhance these analytics systems. In light of the current discussions surrounding AI (Chollet, 2018; Harris, 2018; Metz et al., 2018; Vincent, 2018a), ethical research covering the impact of different technologies, data-based products as well as privacy-related topics are worth further research as they have not been covered in detail in this thesis.

6 Conclusion

This thesis has shown that data will be an essential part of growth and change in the coming years. Three different business models have been identified to help businesses make strategic choices and find fitting monetization options. From a simple “data or insight sale model” over the “transactional model” to the “improve and build model”, a business can offer data-based products with varying complexity.

With the “data or insight sale model”, a business will simply build upon the large availability of data without delivering explicit value to the individual. Companies following the idea of the “transactional model” see two sides to their business: consumers of value as a data source and customers. In this constellation, consumers can expect to benefit from more product value with an increase in the amount of data collected. And, finally, in the “improve and build model”, a company designs a product for the individual customer and uses data for continuous improvement. All these models exceed the scope and scale of data-driven business models as defined by Hartmann et al. (2016) by being more than mere components of the data value chain. As a basis for the successful implementation of data-based models, a new five phase approach of defining the data value chain was proposed.

there is a high probability that companies will be able to deliver an unprecedented amount of personalized value to the individual consumer at a lower cost

The models that have been described have two sides to them from a consumer point of view. On the one hand, consumers will have to put up with the fact that they will lose more of their privacy as they are becoming more transparent, while on the other hand, there is a high probability that companies will be able to deliver an unprecedented amount of personalized value to the individual consumer at a lower cost.

Overall, it can be said that building a business on the idea of consumers as data models presents a major opportunity for companies to create better and more profitable products for the consumers they are serving. Now, it is at the discretion of both the businesses and the customers to define and shape the road ahead for data-based business models.

Now, it is at the discretion of both the businesses and the customers to define and shape the road ahead for data-based business models.