Digital Marketplaces Unleashed

by Claudia Linnhoff-Popien

  At the end of the data analysis phase the results can be of different nature. This can be a prototype of a product for predicting long term staff planning or new insights about the analyzed business process. And sometimes it is even the realization that that the data quality or data coverage is insufficient to get reliable results.

  57.3.4 Presentation of the Results

  After the data analysis phase has ended, the results collected by then are firstly presented to the business department and afterwards to the executive board. The business departments have to assess the potential of the solution by calculating a business case and verifying the practicability of deploying the solution in daily business. According to the outcome of this evaluation the Smart Data Lab results are considered to be enhanced to an operationally mature product.

  57.3.5 Transition into Production

  The Smart Data Lab plays an important role in the procedure of problem solving steps, but it is only the first step. If a result is considered to be used in production the responsibility is handed over to the Smart Data Factory in order to create stable and reliable products running in productive environment. The Smart Data Factory has the responsibility to implement the product using standard components and procedures and to maintain and manage the life cycle of that analytical product. The most important tasks for the Smart Data Factory are the completion of model implementation, the management of the realization project and the change management initiated by the lab’s results. Also, the factory is supposed to maintain and support the product, to monitor and retrain the model, to further develop the product and, not to forget, to train and coach the users.

  The integration of analytical components into a productive application requires more than classical application maintenance and support like keeping up a system and running and guaranteeing a certain performance. The use of models means to ensure high quality of model generated information and to deliver the right information at the right time to the right people. Therefore, it is necessary to monitor and recalibrate models or to re‐engineer them immediately if basic assumptions have changed. Most of the products are used in high critical production environments like operational control centers or security offices. The requirements regarding quality and availability are accordingly high.

  57.4 Results from Smart Data Lab

  57.4.1 Retail‐Oriented Positioning of Flights

  A big share of the profit in airport business is realized in the segment of retail and properties. Airports mostly act as concessionaire and not as shop owner. In that way, they earn money based on revenue participation and store rent with high margin and low cost. At Frankfurt Airport, the shopping range is significantly different within the terminals. Product offering, premium brands availability and shop presentation can vary considerably. The offering ranges from fashion & accessory over cosmetic and electronic devices to luxury goods like valuable watches or exquisite liquor. As manifold as the offering is presented, so are the preferences and needs of the customers. At an airport like Frankfurt, up to 200,000 passengers are moving through the terminal infrastructure passing shops and marketplaces constantly. At the end the purchase trigger depends on various factors like personality traits, the route to the gate, shops passed on this way, waiting and walking time or travel occasion. These are just a few examples of potential influencing variables explaining shopping behavior at an airport, s. Fig. 57.2: Variables and their relationships in the positioning model.

  Fig. 57.2Variables and their relationships in the positioning model

  The positioning of a flight directly affects some of these variables. The amount and type of shops available close to the departure gate is one important factor directly affected. For transferring passengers, the positioning of arrival and departure flights results in routing and thereby in time needed for walking through the terminal and processing time for passing required check points. Frankfurt Airport guarantees a minimum connecting time of 45 min from any gate‐to‐gate location. During the flight planning phase this objective is currently one of the highest priorities for positioning.

  The task of the Smart Data Lab was to identify important variables and the strength of their influence on retail revenues and purchase probability by combining all available data and processing them with data analytics techniques. It is important to know which variables can be controlled by Fraport and how to adjust them to maximize retail revenues without running into other operational problems. For the first time the correlation of high waiting times, transfer distances or nationalities regarding retail revenues could be expressed in a multivariate model helping to understand their influences.

  An analytical challenge was the substitution of known effects which was necessary for different reasons. Some of the passenger characteristics cannot be tracked for reasons of data privacy, e. g. their nationality. We used destination instead but found that data about the nationality itself, theoretically available from existing data sources, would lead to more accurate predictions. So, there are ongoing discussions on how that information could be gathered without violating privacy. Other characteristics like spending capacity or the motivation for traveling simply are not available on a customer level. An organizational challenge was the overcoming of existing decision rules that turned out to be myths. It was hard work to explain the difference between correlation and causation, including the issue of spurious or partial correlation, and the difference between a multivariate prescriptive model and descriptive, bivariate analysis. That challenge required excellent communication skills including an idea of political structures among the organization.

  Two very important applications can be derived from the results generated out of the Smart Data Lab. First, the insights can be used for future negotiations with airlines about positioning scenarios or the exclusive and prioritized use of terminal areas. Second, current flight planning process in agreement with existing arrangement can be optimized to achieve some quick‐wins on revenue lift without impairing superior objectives like the guaranteed transfer time.

  57.4.2 Prediction of Arrival Time

  Each time an aircraft arrives at a gate, various activities are kicked off immediately. Passengers jump off their seats impatient to grab their belongings and leave the aircraft while ground handlers open the airlock to unload baggage and freight containers are just a few examples of what might happen directly after the aircraft has stopped at the gate. What seems to be an ideal situation where ground handlers arrive just in time ready to start aircraft operations, is quite challenging in reality. There are two extreme situations which describe how the problem the Smart Data Lab was confronted with effects daily business. On the one hand this would be a situation where the aircraft arrives at the gate without all required equipment and staff available yet. On the other hand, it would be a situation where ground handling resources are bounded on position waiting for the aircraft to come to rest. While the first problem results in an inconvenience for passengers with missed connection flights and malus payments due to failing service level agreements, the second one implies high cost because of resources occupied while waiting for operation to start. The task of the Smart Data Lab was to improve the prediction of an arrival flight currently approaching the Frankfurt Airport in order to optimize dispatching of resources and thus reduce waiting time for both the passengers and ground handlers.

  The final model combined different components: a model for the airspace movements, a ground movement model, a model for the usage of runways (especially dealing with swing‐over probabilities) and a model for wheeling time. Not only the construction and parametrization of the different models was a challenge, but so was their combination and the interaction of parameters, including restrictions. An organizational challenge occurred as well. Data that were so far no
t included in managing the business process turned out to be extremely valuable in predicting arrival times. It also turned out that innovative operating figures invented by the Smart Data Lab would significantly improve the prediction. That means, it could become necessary to change the whole process, involving more than one department.

  By combining data from radar, weather and flight operations, the Smart Data Lab has implemented a prototype of a prediction model which would reduce waiting times up to 60% and would contribute to improve the quality of service for passengers and airlines and to reduce the cost for ground handlers caused by idle capacity. Since the solution worked out by the Smart Data Lab affects a broad range of stakeholders including ground handlers, airlines, federal authorities and shippers, it is considered to be a mission with critical components essential for smooth operations in daily business.

  57.5 Summary and Outlook

  One of the key lessons learned was that problem solving with big data and analytics needs a lot more than expert knowledge in analytics and big data. The task of persuading other departments that analytics would be inevitable to make smarter decisions and would turn Frankfurt Airport into a digital‐driven marketplace, turned out to be more difficult than any technological issues that came up. The incorporation of interdisciplinary practical expertise into nearly every single phase of the analytics process was crucial, in order to correctly interpret the results and to identify pitfalls which would have led to severely misleading conclusions.

  The change process so far has only started. New business models discussed are cooperative Smart Data Labs with partners like Lufthansa, Deutsche Bahn etc. in order to create novel, integrative passenger services. Fraport also thinks of partner data models or even open data strategies, involving as many stakeholders as possible, not only as data providers, but also as users of the data and as developers of new applications. Finally, analytics‐as‐a‐service driven by customer needs might lead to the development of smart applications. We observe a huge cultural transformation of the whole organization, turning Fraport into a digital marketplace both regarding the inner core of the company and its interactions with other parties.

  References

  1.

  R. Kitchin, The Data Revolution : big data, open data, data infrastructures & their consequences, London: Sage, 2014.

  2.

  S. Goldsmith und S. Crawford, The Responsive City : engaging communities through data-smart governance, San Francisco: Wiley, 2014.

  3.

  R. Sharda, D. Delen und E. Turban, Business Intelligence and Analytics : Systems for Decision Support, Essex: Pearson Education Limited, 2014.

  4.

  G. S. Linoff und M. J. A. Berry, Data Mining Techniques for Marketing, Sales and Customer Relationship Management, Indianapolis: Wiley, 2011.

  © 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_58

  58. The Digitization Dilemma of Europe’s Non-Profit Organizations: Software as a Service to the Rescue!

  Florian Fuchs1 , Michael Liebmann1 and Frank Thelen2

  (1)doo GmbH, Munich, Germany

  (2)e42 GmbH, Bonn, Germany

  Florian Fuchs (Corresponding author)

  Email: florian.fuchs@doo.net

  Michael Liebmann

  Email: michael.liebmann@doo.net

  Frank Thelen

  Email: frank@e42.com

  58.1 Introduction

  Europe boasts more than 5.5 million non‐profit organizations (NPOs) with 250 million members and volunteers. Examples are local clubs (e. g. sports, social, environmental), governing bodies (national/regional associations) and professional associations as well as educational institutions. The European Union (EU) has for many years emphasized the “social and economic cohesion” promoted by non‐profit organizations (NPOs) and their “considerable economic contribution” [1]. It clearly recognizes both the challenges faced by NPOs and the potential of cloud computing for business model transformation.

  However, the EU’s recommendations for the “professionalization” and “modernization” of NPOs and educational enterprises are far from being met. According to the Commission “the main challenge for this sector are […] the changes in the way people volunteer”, and “growing tension as volunteers are confronted with increasingly demanding tasks that require specific competences and skills”. The Commission therefore recommends that the “professionalization” of non‐profits and “the specific need of various groups (elderly, young people) must be better taken into account” [2].

  On the other hand, the EU pushes the Digital Agenda for Europe and emphasizes the opportunities of cloud computing and the need to “enhance interoperability” and “strengthen online trust and security” in Software as a Service offerings in the cloud [3]. It works towards establishing one Digital Single Market in order to unlock the scale necessary for cloud computing to reach its full potential in Europe.

  In this chapter we describe the specific challenges, which NPOs face and share in terms of digitization. We explain why NPOs are particularly threatened to be left behind and argue that the Software as a Service (SaaS) model of software delivery represents the only path to survival. To underpin our conclusion we compare SaaS adoption in other industries and estimate the social impact of a SaaS offering for NPOs on Europe’s society.

  58.2 Europe’s Non‐Profit Organizations Struggle with Digitization

  Europe’s NPOs are being left‐behind by digitization: In spite of generating >160B € in revenues each year, the vast majority of Europe’s 5.5 million NPOs are still using manual, ‘archaic’ processes for communication, interaction, administration and payments [4]. Member interactions mostly happen offline and are un‐automated, unscalable and typically very local. Even very large associations on national level still work with fax and offline processes. Even they are missing out on being role models for their smaller and more local affiliates. In a time of rapid societal change, this implies specific drawbacks.

  58.2.1 Need to Digitize

  Firstly, they have limited access to young people. Today’s students face increasingly demanding curricula and employers expect higher flexibility and mobility from young professionals. This means that young people simply have less time for community and extracurricular activities. Furthermore, the “digitally native” generation tends to avoid traditional interaction (post, fax, phone, checks …).

  Secondly, 85% of NPOs face difficulties staffing leadership positions [5]. As such, many NPOs are characterized by aging leadership, typically with less IT knowhow.

  Lastly, increasing regulation and cost pressure increase workload and create frustration within NPOs – particularly as they heavily rely on manual processes.

  To overcome these challenges, NPOs – just like any other enterprise – have to start taking advantage of digital technologies and automate manual procedures. At the same time, they must also increase member engagement through greater online and mobile interaction. However, this digital transformation is significantly more complex for NPOs than for profit‐driven businesses, because of their inherent innovation dilemma and lack of scale.

  58.2.2 Digitization Dilemma

  The innovation dilemma in NPOs means that they face much greater difficulties adapting to the digital transformation than other business areas. As demonstrated in Fig. 58.1, lack of budget means that buying software from a typical systems provider far exceeds the budgets available to NPOs. Furthermore, with largely unpaid staff, lack of IT know‐how prevents NPOs from choosing the best IT partners and managing implementation projects. Finally, leadership characterized by outdated admin procedures and unawareness of cloud services means that, more generally, NPOs have low innovative capacity.

  Fig. 58.1Viciou
s cycle of innovation dilemma in NPOs

  This unfolds a vicious cycle: NPOs with lower innovative capacity drift further away from attracting new and young members. Yet, without youth engagement, there is limited chance to increase innovative capacity. As a result, the digital landscape within NPOs ranges from purely paper‐based processes to a patchwork of half‐integrated solutions. Indeed, for NPOs that are using some SaaS applications, a second dimension to the dilemma emerges – with NPOs struggling to manage multiple, stand‐alone applications that have to be used separately and have different policies for using (or misusing) their data.

  58.3 Characteristics of Software as a Service

  Fostered by the proliferation of cloud computing, Software as a Service (SaaS) has become a popular licensing and delivery model for software:[Software as a Service is] software that’s developed and hosted by the SaaS vendor and which the end user customer accesses over the Internet. Unlike traditional packaged applications that users install on their computers or servers, the SaaS vendor owns the software and runs it on computers in its data center. The customer does not own the software but effectively rents it, usually for a monthly fee [6].