Introduction
Innovation is a central issue in contemporary organizations (Dodgson, Gann, and Phillips 2013b; Fagerberg and Verspagen 2009; Salter and Alexi 2013). When managed effectively, innovation can improve both the top and bottom lines of businesses worldwide (Dodgson, Gann, and Salter 2008; Tzeng 2009). However, ineffective management of innovation may lead to an organization becoming uncompetitive or reliant on technologies that, once successful, are now obsolete (Dodgson, Gann, and Phillips 2013a; Christensen 2013; Freeman and Soete 1997). Technology can facilitate effective innovation management; frameworks are useful in guiding interaction with, and use of, technologies. One such technology, which is changing the way businesses are conducted, is artificial intelligence (AI), in particular machine learning (Domingos 2015; Michalski, Carbonell, and Mitchell 2013; Kolbjornsrud, Amico, and Thomas 2016).
This chapter explores how machine learning technologies can better enable the management of innovation. In particular, it outlines how activities related to data mining, predictive analytics, statistical modeling, and administrative tasks are better performed by machines (Bose and Mahapatra 2001; Kolbjornsrud, Amico, and Thomas 2016). It also discusses how making sense of the output of those activities, in the context of an ambiguous and nuanced business environment, is better achieved by people. It is argued that innovation managers who harness the power of machine learning technologies will be able to streamline, accelerate, and bring more effective innovations into the markets in which they operate.
Preliminary Definitions
First, it is important to define innovation management and machine learning in the context of this chapter. The definition of innovation management used in this chapter is simple: it is the processes and activities that support innovation (Dodgson, Gann, and Phillips 2013b). The definition of machine learning is less straightforward, however. Machine learning is a subset of AI (Domingos 2015), which is a broader concept, encompassing the study, design, building, and analysis of computational models of action, perception, and cognition (Honavar 2006, p. 3; Langton 1997; Minsky 1974). Machine learning then is the processes that computational systems manifest to accurately recognize patterns, discover knowledge, manage and mine data, and perform predictive analytics (Domingos 2015; Michalski, Carbonell, and Mitchell 2013).
Managing and Balancing Innovation
The nature of innovation is evolving, presenting a range of challenges in its management (Salter and Alexi 2013). The challenges of innovation management are not one-dimensional: from a managerial perspective, they include the effective organization of diverse resources, actions, choices, procedures, and practices. The management of innovation literature explores how an organization can integrate its multiple innovation processes into one comprehensive innovative approach to deploy new products and services (Sundbo 1997; Trott 2008; Dodgson, Gann, and Phillips 2013b, p. 20). Other studies in the literature discuss how organizations manage their intellectual property (Malerba and Adams 2013), how to manage the process of mergers and acquisitions that may increase a company’s innovation capabilities (Ahuja and Novelli 2013), and external collaboration between companies and external entities to bring innovations to the market (Chesbrough, Vanhaverbeke, and West 2006; McKelvey 2013). This includes understanding the role of users (Franke 2013) and consumers (Osaki and Dodgson 2013) so that companies know how an innovation is used and consumed.
In the midst of these challenges, particularly relevant is the question: how should a company balance its innovation portfolio, that is, how can it better manage the radical versus incremental innovation output of a company (Dodgson, Gann, and Phillips 2013b)?
Radical innovation refers to the level of market disruption from an innovation, which is evaluated through its capacity to create new value in a given market (Christensen, Raynor, and McDonald 2015) and even through its capacity to create new categories of products and services. Incremental innovation focuses on improving internal processes, and combining and recombining internal assets and capabilities to drive overall innovation output (Tzeng 2009). This is considered a more subtle and also more common approach to innovation.
These two approaches contrast fundamentally in their impact. In attempting to find a balance between radical and incremental innovations, internal innovation capabilities, insights from market knowledge, and the evaluation of risks associated with each type of innovation are used and applied differently (Dodgson, Gann, and Phillips 2013b; Stringer 2000).
How to balance and manage an innovation portfolio is relatively unknown. This is partly because there is limited information about what a company believes a balanced innovation portfolio should be and look like and also about the ways in which individuals try to manage and achieve balance. While people are trained in structured and rational thinking in which facts drive most of the decision-making process within an organization (Locke 2005; Simon 1979), in practice decisions are often made without a full understanding of all the factors and/or full access to all information available (Gilovich, Griffin, and Kahneman 2002; Jensen et al. 2007).
Successful innovation companies clearly articulate their innovation ambition through managing their total innovation portfolio rather than specific projects (Nagji and Tuff 2012). However, many organizations lack effective management tools for managing their innovation portfolio. In the most successful organizations, internal frameworks and capabilities are implemented to provide fast, clear, comparative, and automated analysis of the success of past innovations. In doing so, these organizations aim to provide an accurate forecast of the performance of future endeavors. For example, successful innovation companies know how to diffuse their innovations and understand when to undertake radical innovation, depending on the level of encroachment into the market (Schmidt and Druehl 2008). However, there is little evidence as to how these companies document, track, and create knowledge from their innovation diffusion. Further, successful innovation may also be underpinned by advances in technology and a careful evolution in which the level of disruption the innovation creates in a given industry is more gradual or nuanced (Norman and Verganti 2014). However, these companies may not be able to explain how this innovation evolution works in practice, in terms of the connections these different levels of disruption have with other industries, or how consumer preferences have evolved as a result.
Against this background of uncertainty experienced by companies when innovating, and to effectively manage their innovation portfolio, the provision of fast, accurate, and predictive analytics is paramount to assist in tracking past innovation efforts and forecasting the success of new ones. Such analytical power, when harnessed appropriately, can provide a comprehensive view of the impact an innovation brings into the market. Hence, knowing whether, when, what, and how an innovation is incremental or radical is useful for effective decision making.
Better Solutions—Improving an Innovation Portfolio by Learning From It
The automation of administrative tasks behind innovation, the consolidation of diverse and vast inputs from the market, and the contextualization of these with the financial and sales performance of new products and services are critical tasks for companies seeking to manage innovation effectively. Through automation and systems that learn from large datasets, innovation managers will be better equipped with more complete information about the performance, profile, and diffusion of their different innovation projects.
In this regard, there is much to embrace, explore, and exploit in the capabilities of machine learning. Data mining is a powerful process that can transform large datasets into usable knowledge (Bose and Mahapatra 2001). The financial services industry is exploring the prospect of using machine learning to profile credit scorings (Kruppa et al. 2013) and smart web-browsers exploit machine learning techniques to better enable the suggestion of targeted and retargeted advertisements based on an internaut’s search behavior. Large online retailers have turned to predictive analytics to facilitate marketing specific, tailored products to a returning customer, with the aim of increasing their conversion rate while visiting an online store.
The benefits of machine learning for innovative businesses are manifold. Machine learning can streamline and provide informed solutions to managers. Innovation managers who embrace, explore, and exploit machine learning techniques will be better placed to map market expectations of their different innovation projects by gaining knowledge of how these projects operate, that is, broader understanding of potential and past customer behavior according to the characteristics of a product or a service (Tkáč and Verner 2016, p. 792). Effective innovation managers understand that machine learning algorithms can transform large amounts of data into valuable information and relevant knowledge; correctly harnessing this process provides a competitive advantage.
Effectively managing an innovation portfolio requires full understanding of which machine learning techniques can provide a company with better data management and mining capabilities. This also means understanding that a company that has robust data management and data mining frameworks can also have a robust predictive analytics output (Ali and Arıtürk 2014).
In the context of large datasets, the more data companies have at their disposal for analysis, the more they can learn. However, they also face the challenge of how to automate the analytics behind the data. Those companies that excel at transforming their current frameworks and capabilities to accomplish this will have a technological advantage over their competitors (Domingos 2015). For example, within the financial services industry, companies can use these techniques to better equip their managers in making more accurate and fact-based decisions related to future investments according to market trends (Kolbjornsrud, Amico, and Thomas 2016).
Businesses that understand how to integrate the power of machine learning are better able to exploit their benefits. For instance, online retailers are better placed to understand how browsing patterns provide rich datasets to personalize consumers’ shopping experience and segment their catalogs accordingly (Baluja 2006; Mobasher 2007).
Innovation Manager—Machine Interaction for Improved Management
Companies that have successfully provided a framework and the internal capabilities for computational systems to learn from the dynamics of an innovation portfolio will realize that innovation managers are no longer users of information but rather interactors with the machines providing it (Suchman 1990). This means that the management of innovation should clearly state the role that humans and machines have in the interaction.
In the context of human-machine interaction, having a clear intention, allowing the machine to make an accurate selection of the action resulting from such intention, and executing the action make human evaluation of the results of such execution possible (Norman 1984). The successful interaction of innovation managers with machines implies a clear characterization of what the manager wants the system to provide, in other words, the innovation manager can identify the future action intended, and the machine recognizes the intention and complies with an action accordingly (Tahboub 2006, p. 36). For example, an innovation manager wants to understand the financial impact of past innovation considering a specific time frame and the resultant increase in net revenue. The machine in turn will comply with this request, the intention of which is to understand the performance and financial impact of the company’s innovation portfolio.
Certainly, this type of request can be done—and still is mostly done—by people. However, in the face of larger and larger datasets becoming available, having machines perform these tasks means companies will have faster and more accurate answers (Kolbjornsrud, Amico, and Thomas 2016; Domingos 2015; Bose and Mahapatra 2001). The sheer speed in the provision of information and knowledge means that the decision-making process itself is faster, with the potential to increase a company’s market competitiveness.
Machines are better at data management, mining, and predictive analytics if the right technological capabilities are put in place. In terms of managing an innovation portfolio, the machines will own the calculation of probabilities of success of a company’s innovative endeavor but the possibilities will still belong to humans.
Conclusion
This chapter discusses how effective management of innovation can improve the overall performance of organization. To be effective, however, companies must embrace technological advances, such as machine learning, to improve the way in which innovation is managed. In particular, this chapter focused on the potential of machine learning techniques to improve the overall management of a company’s innovation portfolio by providing faster, better, and more accurate information obtained from large datasets.
Exploring the possibilities of machine learning in the context of innovation management is an ongoing and evolving endeavor, in part because machine learning techniques are in a nascent stage. Despite this, some industries are already exploiting the capabilities that these techniques provide.
At the moment, these technologies allow us to command the lights to be turned on, to find out the weather forecast for today, or to dial a colleague with a simple voice command. Our command can be acted upon almost instantaneously. Soon, these same technologies will allow innovation managers to ask for—and almost instantaneously receive—the financial contribution made by a set of new products or services in the company’s profit and loss statement, how the consumer received an innovation initiative, and even what is the probability of success of an innovation project based on previous innovation efforts and market input. There is much potential to be exploited in machine learning capabilities for innovation managers.
But always, in the human–machine interaction, there is a role for humans. In the age of larger and larger datasets, innovation managers embracing machine learning techniques must also understand that making sense of complex information is still a human endeavor.
References
Ahuja, G., and E. Novelli. 2013. “Mergers and Acquisitions and Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 430–45. Oxford, UK: Oxford University Press.
Ali, Ö.G., and U. Arıtürk. 2014. “Dynamic Churn Prediction Framework with More Effective Use of Rare Event Data: The Case of Private Banking.” Expert Systems with Applications 41, no. 17, pp. 7889–903.
Baluja, S. 2006. “Browsing on Small Screens: Recasting Web-Page Segmentation into an Efficient Machine Learning Framework.” In Proceedings of the 15th International Conference on World Wide Web, 33–42. ACM. http://dl.acm.org/citation.cfm?id=1135788
Bose, I., and K.R. Mahapatra. 2001. “Business Data Mining—a Machine Learning Perspective.” Information & Management 39, no. 3, pp. 211–25.
Chesbrough, H., W. Vanhaverbeke, and J. West. 2006. Open Innovation: Researching a New Paradigm. Oxford: Oxford University Press on Demand.
Christensen, C. 2013. The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Boston, USA: Harvard Business Review Press.
Christensen, C., M. Raynor, and R. McDonald. 2015. “Disruptive Innovation.” Harvard Business Review 93, no. 12, pp. 44–53.
Dodgson, M., M.D. Gann, and N. Phillips. 2013a. “Organizational Learning and the Technology of Foolishness: The Case of Virtual Worlds at IBM.” Organization Science 24, no. 5, pp. 1358–76.
Dodgson, M., M.D. Gann, and N. Phillips. 2013b. “Perspectives on Innovation Management.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 3–26. Oxford, UK: Oxford University Press.
Dodgson, M., M.D. Gann, and A. Salter. 2008. The Management of Technological Innovation: Strategy and Practice. Oxford: Oxford University Press on Demand.
Domingos, P. 2015. The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World. New York, NY: Basic Books.
Fagerberg, J., and B. Verspagen. 2009. “Innovation Studies—The Emerging Structure of a New Scientific Field.” Research Policy 38, no. 2, 218–33. doi:10.1016/j.respol.2008.12.006
Franke, N. 2013. “User-Driven Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 69–83. Oxford, UK: Oxford University Press.
Freeman, C., and L. Soete. 1997. The Economics of Industrial Innovation. Abingdon, UK: Routledge.
Gilovich, T., D. Griffin, and D. Kahneman. 2002. Heuristics and Biases: The Psychology of Intuitive Judgment. Cambridge, UK: Cambridge University Press.
Honavar, V. 2006. “Artificial Intelligence: An Overview.” http://web.cs.iastate.edu/~cs572/handout1.pdf
Jensen, M.B., B. Johnson, E. Lorenz, and B.Å. Lundvall. 2007. “Forms of Knowledge and Modes of Innovation.” Research Policy 36, no. 5, 680–93. doi:10.1016/j.respol.2007.01.006
Kolbjornsrud, V., R. Amico, and J.R. Thomas. 2016. “How Artificial Intelligence Will Redefine Management.” Harvard Business Review.
Kruppa, J., A. Schwarz, G. Arminger, and A. Ziegler. 2013. “Consumer Credit Risk: Individual Probability Estimates Using Machine Learning.” Expert Systems with Applications 40, no. 13, pp. 5125–31.
Langton, C.G. 1997. Artificial Life: An Overview. Cambridge, USA: Mit Press.
Locke, E.A. 2005. “Why Emotional Intelligence Is an Invalid Concept.” Journal of Organizational Behavior 26, no. 4, pp. 425–31.
Malerba, F., and P. Adams. 2013. “Sectorial Systems of Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 163–83. Oxford, UK: Oxford University Press.
McKelvey, M. 2013. “Science, Technology and Business Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 69–82. Oxford, UK: Oxford University Press.
Michalski, R.S., J.G. Carbonell, and T.M. Mitchell. 2013. Machine Learning: An Artificial Intelligence Approach. California, U.S.: Springer Science & Business Media.
Minsky, M. 1974. A Framework for Representing Knowledge. Memo No. 306. Artificial Intelligence. Boston, Mass: Massachusetts Institute of Technology. http://18.7.29.232/handle/1721.1/6089
Mobasher, B. 2007. “Data Mining for Web Personalization.” In The Adaptive
Web, 90–135. Springer. http://link.springer.com/chapter/10.1007/978-3-540-72079-9_3
Nagji, B., and G. Tuff. 2012. “Managing Your Innovation Portfolio.” Harvard Business Review, Spotlight, 1 (May), pp. 1–9.
Norman, D.A. 1984. “Stages and Levels in Human-Machine Interaction.” International Journal of Man-Machine Studies 21, no. 4, pp. 365–75.
Norman, D.A., and R. Verganti. 2014. “Incremental and Radical Innovation: Design Research vs. Technology and Meaning Change.” Design Issues 30, no. 1, pp. 78–96.
Osaki, R., and M. Dodgson. 2013. “Consumption of Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 248–71. Oxford, UK: Oxford University Press.
Salter, A., and O. Alexi. 2013. “The Nature of Innovation.” In The Oxford Handbook of Innovation Management, eds. M. Dodgson, D.M. Gann, and N. Phillips, 26–50. Oxford, UK: Oxford University Press.
Schmidt, G.M., and T.C. Druehl. 2008. “When Is a Disruptive Innovation Disruptive?” Journal of Product Innovation Management 25, no. 4, pp. 347–69.
Simon, H.A. 1979. “Rational Decision Making in Business Organizations.” The American Economic Review 69, no. 4, pp. 493–513.
Stringer, R. 2000. “How to Manage Radical Innovation.” California Management Review 42, no. 4, pp. 70–88.
Suchman, L.A. 1990. “What Is Human-Machine Interaction.” Cognition, Computing, and Cooperation, pp. 25–55.
Sundbo, J. 1997. “Management of Innovation in Services.” Service Industries Journal 17, no. 3, pp. 432–55.
Tahboub, K.A. 2006. “Intelligent Human-Machine Interaction Based on Dynamic Bayesian Networks Probabilistic Intention Recognition.” Journal of Intelligent & Robotic Systems 45, no. 1, pp. 31–52.
Tkáč, M., and R. Verner. 2016. “Artificial Neural Networks in Business: Two Decades of Research.” Applied Soft Computing 38, pp. 788–804.
Trott, P. 2008. Innovation Management and New Product Development [4th Ed.]/Chapter 12: New Product Development. Essex: England: Prentice Hall.
Tzeng, C.-H. 2009. “A Review of Contemporary Innovation Literature: A Schumpeterian Perspective.” Innovation 11, no. 3, 373–94. doi:10.5172/impp.11.3.373