Process Model for Technology-Push utilizing the Task ...

15 downloads 105551 Views 746KB Size Report
application area, we selected biomedical laboratory automation on the one ... shows that an extraordinary technical idea or solution will not coercively lead to ...
Process Model for Technology-Push utilizing the Task-Technology-Fit Approach Raimund Hartelt1, Florian Wohlfeil2, and Orestis Terzidis3 [email protected] [email protected] [email protected]

All at the Institute for Entrepreneurship, Technology Management & Innovation, Karlsruhe Institute of Technology, Kaiserstraße 12, 76131 Karlsruhe, Germany

The development of new technologies mostly requires an extensive amount of time and capital investment. To recover these efforts, an adequate commercialization is essential. However, even auspicious technologies do not usually sell themselves. For utilizing a new technology an adequate application field has to be identified. Therefore, the results of technology development have to be matched to a market need. The objective of this paper is to present a systematic process model for technology-push to assess if a chosen technology fits to the tasks of a potential application. We refer to it as the Technology-Utilization-Model (TUM). Taking the Task-Technology-Fit approach of Goodhue and Thompson, we derived six concrete steps to evaluate how good a technology fits to the tasks of a potential application compared to existing technology alternatives. We furthermore provide a handson guideline to calculate an overall Task-Technology-Fit value, which includes the distinct evaluation criteria, their weighting, and the specific technology evaluation characteristics. On this basis, an overall ranking of the potential technology alternatives for a concrete use case could be generated, which corresponds to the likelihood of technology utilization by the customer. In order to empirically test the TUM, we studied an innovative software technology to realize intuitive robot handling and programming in the field of adaptive robotics. Concerning the application area, we selected biomedical laboratory automation on the one hand, and automated cable manipulation in the automotive industry on the other. Within both applications, the intuitive robotic technology has distinct advantages, but also faces some limitations if it is compared with manual operation.

1. Introduction It is generally recognized that technological innovation has a positive impact on international trade, industry structure, growth, and development of new and existing firms and industries (Utterback, 1971, p. 76). Technological innovations are based on the employment and commercialization of new technology. Trott emphasized that technology is not an accident of nature, but is the product of deliberate action by human beings (Trott, 2005, p. 18). According to Bullinger, technology is the knowledge about potential approaches to solve technical problems (Bullinger, 1994, p. 34). Commercialization on the other hand, is the conversion of an idea from research into a product or service for sale in the marketplace (Rogers, 2003, p. 152). As the development of new technologies usually requires an extensive amount of time and capital investment, a broad commercialization is essential to recover these efforts (Ullman, 2003, p. 69). However, good technologies do not usually sell themselves (Gibson and Smilor, 1991, p. 291). Traditional engineer thinking sometimes leads to the fallacy that every good technical idea could be transformed to a functioning technical artifact or process. The experience of technology-driven companies

2

shows that an extraordinary technical idea or solution will not coercively lead to commercial success (Bullinger, 1994, p. 85). For utilizing a new technology an adequate market has to be identified. Therefore, the results of technology development have to be matched to a market need or a new market has to be developed for them (Ullman, 2003, p. 69). However, searching for market opportunities for a technology is a demanding challenge (Keinz and Prügl, 2010, p. 269), as technology development, on the one hand, and successful market introduction, on the other hand, have independent key success factors. Moreover, starting product development without a present market demand is a specific challenge that is called technology-push. In contrast to market-pull innovations, potential market opportunities and application fields are initially unknown in the case of technology-push. A major driver of market success is the ability to identify and exploit these opportunities (Henkel and Jung, 2009, pp. 1–2). Usually, the range of market opportunities emerging from a new technology and thus the generality of a technological competence are rarely clear from the outset (Gruber et al., 2008, p. 1653). Additionally, systematic and consistent approaches for identifying market opportunities for new technologies are scarce in literature and a lack of adequate tools therefor is existent (Henkel and Jung, 2009, p. 1). The aim of this paper is to present a systematic process model for technology-push utilizing the task-technology-fit approach of Goodhue and Thompson (1995). We refer to it as the Technology-Utilization-Model (TUM). It addresses six steps to assess if a technology fits to the tasks of a potential application and is intended to serve as a roadmap for opportunity evaluation with regard to the respective technology. We conducted a multiple-case study to verify the applicability of the TUM in real life. The remainder of the paper is organized as follows: in the next section we present the task-technology-fit-model of Goodhue and Thompson. We then describe the research approach in section 3. Section 4 contains the TUM and the findings of the multiple-case study. We end up in section 5 with the discussion and conclusions.

2. Task-Technology-Fit Approach Goodhue and Thompson (1995) proposed a Task-Technology-Fit approach for information systems research (cf. Figure 1). At the heart of their model is the assertion that for an information technology to have a positive impact on performance, the technology naturally has to be utilized and therefor has to have a good fit with the tasks it supports. They empirically tested their approach by utilizing data from over 600 individuals using 25 different information technologies and working in 26 different departments in two companies (Goodhue and Thompson, 1995, pp. 213, 214).

Figure 1: Task-Technology-Fit Approach (Goodhue and Thompson, 1995, p. 220)

In their study, Goodhue and Thompson found evidence that supported the applicability of the Task-Technology-Fit approach for the information system discipline. Their study highlights the importance of the fit between technologies and users' tasks in achieving performance impacts from information technology. It also suggests that task-technology-fit, when decomposed into its more detailed components, could be the basis for a strong diagnostic tool to evaluate whether information systems and services in a given application area are meeting user needs (Goodhue and Thompson, 1995, pp. 213, 228). On this basis, we took the Task-Technology-Fit approach and derived a systematic process model with concrete steps for the challenge of technology-push.

3

3. Methodology The central research goal for the study at hand was to develop a straightforward process model for evaluating the technology fit regarding the task of a concrete application area. This process model should be empirically tested. Therefore, we followed a qualitative research approach. In particular, we chose the case-study methodology as case studies are multi-perspective analyses that serve as an instrument to analyze complex issues in real life (Tellis, 1997). Compared to single-case studies, multiple-case studies are advantageous when it comes to generalization logic. Thus, we followed a multiple-case study design (Yin, 2014, pp. 63-64). Yin proposed a procedure (cf. Figure 2) for conducting multiple-case-studies. This procedural model is subdivided in three general phases: define and design, prepare collect and analyze, and analyze and conclude. In the first phase the theory which should be analyzed has to be developed (Yin, 2014, p. 84). Based on the Task-Technology-Fit approach (cf. Figure 1), we derived a systematic process model for technology-push, the TUM (cf. Figure 3). Subsequently, the TUM formed the underlying theory and guided the case study. Case selection for this study has been carried out on the basis of content related and research pragmatic factors. We deliberately searched for a technology-push situation with an auspicious technology that exhibits a high level of newness and disruptiveness, but similarly an adequate maturity level for being introduced to the market. Furthermore, we looked for adequate application areas for this technology to evaluate the task-technology-fit. Therefore, the general market openness to adopt technological innovations, the potential usability, and the existence of a concrete application area were considered. Moreover, the access to key informants was decisive. As fundamental technology, we finally chose a software concept that is based on an innovative algorithm to realize intuitive robot handling and programming in the advanced field of adaptive robotics. This software technology was developed by a university spin-off and was already brought to application maturity. Concerning the application area, we selected biomedical laboratory automation on the one hand, and automated cable manipulation in the automotive industry on the other. Both application fields are generally open-minded regarding technological innovations and offer concrete but basically diverse use cases for the central software technology. This constellation represents an opportunity for the current study, as the task-technology-fit of the underlying technology could be analyzed and subsequently compared for two different application areas.

Figure 2: Multiple-Case-Study Procedure (Yin, 2014, p. 60)

Each of the two use cases for the fundamental robotic software technology consisted of a whole study. For reasons of triangulation, we collected data from multiple sources. We utilized primary data like expert interviews, direct observations, and technical discussions. Therefore, we visited an industrial fair and an application site to get in contact with technology users. Additionally, we constantly communicated with the technology provider. Furthermore, secondary sources like technical documents, websites, press releases, and articles have been considered during data collection. For each individual case study, a data collection protocol was developed, which guided the respective case study. Afterwards a separate case study report was prepared for each case. Within the conducted case study, we followed an iterative approach by integrating learnings and emerging questions in the research process. In phase three, we compared the results of the two application fields for the basic technology. We drew cross-case conclusions and documented the results in a cross-case report (Yin, 2014, p. 84).

4

4. Findings A. The Technology-Utilization-Model Figure 3 shows the Technology-Utilization-Model (TUM), which represents a systematic process model for technology-push utilizing the Task-Technology-Fit approach. It contains six steps to assess how a concrete technology fits to the tasks of a potential use case. It is intended to serve as a roadmap for opportunity evaluation with regard to the respective technology. The core of the model is a comparison with potential technology alternatives evaluating each alternative by an extensive, preassigned set of evaluation criteria. By weighting the evaluation criteria from the application perspective the technology performance impact can be estimated, which also helps to derive the likelihood of utilization by the customer. Step 1: Technology Characterization

Step 2: Task Characterization

Step 3: Derivation of Evaluation Criteria

Step 4: Technology Assessment comparison with potential alternatives

Step 5: Performance Impact Evaluation weighting the evaluation criteria from customer perspective

Step 6: Task-Technology-Fit Conclusion and Customer Utilization calculating an aggregated value and generating an overall ranking Figure 3: Technology-Utilization-Model (TUM)

Step 1: Technology Characterization To be able to evaluate the task-technology-fit, at first, it is necessary to characterize the fundamental technology. Therefore, its basic structure, its physical operating principle, its general functionality, and its workflow principles have to be illustrated. Moreover, the feasibility and maturity of the respective technology have to be analyzed. Based on the underlying technological problem, the technology is addressing, the main purpose of the technology, its most promising utilization possibilities, and its potential application fields should be described. In this context, the corresponding technology characteristics, features, effects, general advantages, and disadvantages should be elucidated. If available, technical data would complement the technology characterization. For this initial step, consulting experts, reviewing the technical press, visiting trade fairs, and analyzing technological alternatives are suitable data sources. Step 2: Task Characterization Beside the technology, the addressed task has to be characterized in detail. In this context, a task can be seen as a technological problem which has to be solved. This could be a process in which a technology is deployed or a product in which a technology fulfills a specific function. In both cases, it is important to understand the task’s purpose, its significance, and how the task was previously solved. Data for task characterization can be collected by interviewing application experts, involving lead users, conducting internet inquiries, or by observing and analyzing concrete use cases.

5

Step 3: Derivation of Evaluation Criteria Within step 3, the criteria by which the task-technology-fit will be evaluated have to be derived. Based on the concrete application task, these criteria should be figured out that are mostly relevant for solving the respective task. Therefore, the technology users have to be questioned as their perspective is decisive for the technology success in the end. In addition, attributes that are not obvious to current customers but could be relevant for them in the future should be considered. This could be realized by integrating lead users or utilizing scenario techniques. Step 4: Technology Assessment (comparison with potential alternatives) The assessment of a concrete technology can only be done by comparing it with the existing alternatives. Therefore, the most relevant technological alternatives have to be considered. Based on the set of evaluation criteria derived in step 3, the potential alternative technologies have to be assessed relative to each other. To support this process, we have developed a ten point rating scale, ranging from one (very negative) to ten (very positive). For each technology, the individual parameter value for each particular evaluation criteria has to be determined relative to the existing alternatives. This evaluation should be conducted by consulting technology experts, who are experienced in the field. Obtaining more than one expert opinion, critical scrutiny, and cross-checking broadens the result validity. Step 5: Performance Impact Evaluation Since some characteristics are usually more important to customers than others and thus more relevant for the tasktechnology-fit, the evaluation criteria have to be weighted from the customer perspective. The more important an evaluation criteria is perceived, the bigger is its impact on the task performance. We suggest utilizing a three point scale to weight the importance of an evaluation criterion (weight of one means less important, two means important, and three means very important). The evaluation should be done by interviewing technology users, lead users, or application experts. Step 6: Task-Technology-Fit Conclusion and Customer Utilization In order to judge the resulting task-technology-fit for a concrete technology, the assessment with regard to each evaluation criteria and the corresponding weights of these criteria have to be considered. Therefore, we suggest calculating an overall Task-Technology-Fit value (TTF) that includes the weighting (W1-n = 1 to 3) of each evaluation criteria and the specific technology evaluation characteristics (TE1-n = 1 to 10) for each technology (X): 𝑛

𝑇𝑇𝐹𝑋 = ∑(𝑇𝐸𝑖 ∙ 𝑊𝑖 ) 𝑖=1

On the basis of the TTFX-value, the analyzed technology alternatives can be compared with each other and finally an overall ranking could be generated. This value exhibits the task-technology-fit of a technology with respect to a concrete application task relative to its alternatives. With this value it is possible to predict the likelihood of technology utilization by the customer.

B. Empirical Evidence Step 1: Technology Characterization - Intuitive Robotic Software Technology The core function of robotics is the development and control of robots. In the recent past, one of the main challenges in robotics is to program fast and intuitively complex robot tasks. Especially tasks that need to be controlled sensoradaptively via force-torque sensors or via computer vision are hard to realize with conventional automation software. The intuitive robotic software technology analyzed in the current case study addresses this matter. This software technology allows to program complex force-torque adaptive robot motions via wizards in relatively short time while supporting several robot types (Robotic Startup, 2014a, b; BCG, 2015). The fundamental core of the software technology is an internal representation of an adaptive robot task including all relevant constraints with regard to geometric, temporal, force, and tactile factors. In contrast to classical robotic software, in which robot paths are usually clearly defined by simple vectors, parameter variances and strategies to handle these variances can be represented within the internal model. The software is able to translate this data model automatically into a coded robot program, which, ultimately, can be deployed directly within a stand-alone robot application. Furthermore, the intuitive robotic technology includes new methods to parametrize the model depending just on a few, intuitive key inputs. By providing a library of elementary motion types, the software allows

6

the user to compose complex program sequences and to parametrize the elements for instance by a small number of kinesthetically taught robot poses (Robotic Startup, 2014a, b). According to the technology developers, the new internal task and motion model involves economic and technological advantages. First, it enables robot systems to cope with significant process and component variances. Hence, less technical requirements have to be met, which simplifies application engineering. In addition, being able to cope with variances the technology allows deploying robotic systems in areas of application, in which robotic automation was just not possible before. Second, the technology allows generating not only classical but also complex robot programs faster and more intuitively. This enables a more efficient initial commissioning phase, but even qualifies local staff to adjust and change robot tasks easily without consulting an expert. Yet, it is a powerful tool to handle adaptive motion tasks at the same time (Robotic Startup, 2014a, b).

Use Case – Laboratory Automation Step 2a: Task Characterization – Laboratory Automation A potential application field for the intuitive robotic technology is the task of laboratory automation within bioanalytic life science research and personalized diagnosis therapy. This task contains basic laboratory activities such as pipetting and sample handling, but also extensive activities like sample preparation and processing by following laboratory protocols and methods. For instance, the process of DNA or RNA extraction from different sources is complex, needs multiple steps to be accurately executed, and is often the basis for further experiments and analysis (NCT, 2014; Müller and Röder, 2004 p. 161). For this kind of laboratory work, special laboratory equipment and devices are necessary like for thermal cycling, testing, or other specific purposes (Mülhardt, 2009, pp. 7, 8). In general, demand for sample analyzes is increasing due to new medical opportunities, like personalized diagnosis and therapy, and due to our increasing aging society. Both trends lead to higher sample numbers (NCT, 2014, 2013, pp. 14-17). Another aspect is, that processed liquid volumes are becoming smaller per sample by using microtiter plates in order to save chemicals, to reduce costs and to increase throughput. Hence, laboratory automation solutions have to be able to handle these little volumes precisely. Several laboratory automation solutions exist in terms of special purpose automation systems, such as liquid handling work stations. These automation solutions are mainly used for bigger settings when large sample numbers are similarly processed (NCT, 2014; Röder, 2004, pp. 157-163). Furthermore, several classical robotic solutions were adapted for laboratory application. Their hardware differs from standard industrial robots through specific materials and designs, which allow implementing robot systems also in cleanroom conditions (Denso Robotics Europe, 2014, p. 2; Hartelt, 2015). These systems can be used for instance within isolated work cells in order to handle very dangerous or sensitive substances (Hartelt, 2015; AST, 2010, pp. 4-5). Traditionally in the case of bio-analytic research institutes and smaller laboratories, many laboratory activities are carried out by manual labor. There are some sensitive tasks which technically can’t be automated yet like sample cutting, separating cell layers, or quality control activities for example. Moreover, the numbers of processed samples can diverge significantly among individual research projects and are sometimes on a low level. Therefore, even if they technically could be automated, the pipetting work and other activities are primarily performed manually because it takes too much time and effort to parametrize and adjust the existing automation systems for these less extensive experiments. Since these activities are still repetitive, there can be seen a potential for automation. This would help to relieve researchers and well trained medical-laboratory assistants so that they can concentrate on more demanding tasks (NCT, 2014; Poremba, 2014, p. 2). Step 3a: Derivation of Evaluation Criteria Within the use case of laboratory automation, we have derived nine evaluation criteria that are decisive for assessing a technological solution. First, throughput implies how many samples can be processed by the technology within a defined time frame. Second, process robustness signifies how autonomous a task can be executed and how errorprone a technology alternative is with regard to significant process variances. Third, quality in terms of process results covers precision and reproducibility. Precision means how accurately the automation system can execute laboratory processes for example to handle small liquid volumes. Reproducibility expresses how identically a laboratory task can be repeated resulting in the same analysis outcomes. Fourth, acquisition costs refer to initial investment costs which contains also hardware but also set-up costs. Fifth, application range indicates how flexible a technology alternative can be applied to different kinds of laboratory tasks. Sixth, adjustment flexibility addresses technology flexibility concerning hardware and software changes due to new laboratory processes or protocol changes, when the system already has been set up. Seventh, the handling of a laboratory automation technology implies how intuitive and facile the operation, programming, and maintenance can be executed and how much training effort is necessary in order to be able to parametrize laboratory tasks. Eighth, collaboration implies

7

not only the technology ability to collaborate with people or to be safely implemented among people, but also implies the general technology acceptance among laboratory personnel. Ninth, cleanliness denotes how well a technology is able to process sensitive samples without contaminating them. The evaluation criteria were derived primary based on the input of expert interviews and observations (NCT, 2014; Robotic Startup, 2014 a, b, c). Step 4a: Technology Assessment The intuitive robotic technology has been evaluated with respect to the laboratory automation tasks in comparison to three potential technology alternatives: manual labor, special purpose automation solutions, and classical robotics. The assessment has been conducted primary based on observations and interviews with technology and application experts (Robotic Startup, 2014a, b; NCT, 2014; Hartelt, 2015). The results are presented in Figure 4 and will be briefly explained in the following.

Figure 4: Technology-Assessment Laboratory Automation

By nature, humans are predestined to adapt their actions to variances and changing situations. Hence, they are able to conduct laboratory tasks in a robust manner and are able to adjust the latter if this is necessary for example due to protocol changes. A wide range of laboratory tasks can be well applied by manual operation requiring only basic capital equipment. Since laboratory personnel are usually highly skilled, little training is needed in order to start manual operation. This also leads to the lowest level of acquisition costs compared with the alternative automation technologies. Manual operation is well accepted among laboratory personnel and usually can be easily applied in laboratory settings. However, especially during processing large sample numbers, manual operation lacks a high level of quality, since mistakes or slight unnoticed process changes occur. Furthermore, manual operation is accompanied by a higher risk of sample contamination. In contrast, the considered automation technology alternatives excel in high quality process execution. However, since robots have more degrees of freedom and may apply adaptive sensors, they aren’t able to measure up to special purpose automation solutions regarding quality. Automation solutions also excel in a lower risk of sample contamination and can be designed to process much larger sample numbers than can be achieved by manual labor. Although robotic systems currently don’t reach the same level of throughput as they are usually not faster than manual operation, they are still able to execute large numbers by continuously processing for a long time. In general, laboratory automation technology consists of hardware and software, which both initially have to be set up according to a specific laboratory task that induces relatively high acquisition costs. In addition, since adaptation is limited and if realized, often difficultly to withdraw, automation technology features minor application range and minor adjustment flexibility. As robotic technology is usually more versatile due to the typical robot design, it allows to cover more application tasks with less required hardware engineering compared to special-purpose automation technology. As noted above, utilizing intuitive robotic software technology further helps to implement tasks even quicker and easier than before and enables to cope with significant variances which improves robustness and enlarges application range once again.

8

Despite classical robotic technology similarly allows using adaptive sensors, applying them during process execution is more difficult due to their complex software implementation process. Facile handling of intuitive robotic solutions simplifies operation and collaboration in daily use, whereas the handling of previous automation solutions requires specific training. Step 5a: Performance Impact Evaluation The weighting of each evaluation criteria have been listed in Figure 4. Since the amount of sample analyses is expected to increase, throughput is a relevant factor regarding laboratory automation solutions. However, also high sample numbers have to be processed accurately and identically in order to achieve a reliable data basis for further research, medical diagnoses, and therapy. In addition, very small liquid volumes have to be processed which contributes to quality and is thus very important. Process robustness is crucial, as samples should be able to be processed autonomously without having to worry about execution problems. Since laboratory protocols and research projects may change on occasion, adjustment flexibility is essential in order to be able to implement these changes easily into the present process execution. Furthermore, the handling of laboratory automation is highly relevant not only because users alternate and have often rare experience with automation solutions, but also because it affects daily operation. Since laboratory automation has to cover several kinds of activities while steadily sample contamination has to be avoided, application range as well as cleanliness are important. Equally, the degree of collaboration is relevant, as automation technology has to be integrated within a complex working environment where users have to interact with each other and similarly with devices. Acquisition costs are perceived to be less important, but still should be considered as laboratories have a limited budget. The performance impact evaluation was primary based on expert interviews (NCT, 2014; Robotic Startup, 2014a, b). Step 6a: Task-Technology-Fit Conclusion and Customer Utilization The calculation of the overall Task-Technology-Fit value (TTFX) for each technology alternative leads to the following ranking (cf. Table 1): Table 1: Task-Technology-Fit Assessment Laboratory Automation

Laboratory Automation Technology Alternatives (X): Manual Operation Intuitive Robotic Technology Classical Robotic Technology Special-Purpose Automation Technology

TTFX 159 150 113 105

In relation to Intuitive Robotic Technology 106 % 100 % 75 % 70 %

Apart from manual operation, it seems that the new intuitive robotic technology fits best to the use case of laboratory automation according to the calculated values. Hence, the likelihood of utilization is supposed to be high in relation to the other automation technology alternatives. Nevertheless, manual operation exhibits currently major advantages, which lead to the highest TTFX-value. Comparing the characteristic importance weighting (cf. step 5) with the technology assessment results (cf. step 4) of the intuitive robotic technology (cf. Figure 4), it could be realized that intuitive robotics has relative advantages regarding those characteristics, which are highly relevant for laboratory activities at bio-analytic research institutes. Thus, the advantages of intuitive robotic technology achieve a correspondingly high performance impact which enhances the likelihood of technology utilization compared to the alternative automation technologies. In contrast, the special-purpose automation technology and the classical robotic technology have distinct disadvantages regarding the highly relevant criteria of process robustness, adjustment flexibility, and handling. Correspondingly, the Task-Technology-Fit values of both technologies clearly turn out to be on a lower level compared to intuitive robotic technology.

Use Case – Cable Manipulation Step 2b: Task Characterization – Cable Manipulation At the second use case for the intuitive robotic technology, the primary focus lies on cable manipulation regarding medium sized components for the automotive industry like cable harnesses for air conditions or motors. Although production in automotive industry is usually highly automated, cable manipulation is still executed by manual operation. Since cables are limp components, which are characterized by large variances regarding their shape and behavior, it is very difficult to automate this challenging task. For human beings, it is relatively easy to identify the

9

exact position of limp objects and to react on their deflection in a robust manner. In contrast for automated solutions, it is difficult to grasp, flex, and mount cables due to the significant variances that occur (Robotic Startup, 2014b; System integrator, 2014; Leoni, 2011). The automotive industry is highly competitive and cost-driven. Hence, there is a great interest in automated cable manipulation. One approach is using classical robotic technology. A large robot manufacturer has already demonstrated handling non-rigid hoses by utilizing torque-force sensors. However, setting up these tasks on robot systems takes a lot of time up to several months (KUKA, 2014a, b, p. 10; Robotic Startup, 2014b). Step 3b: Derivation of Evaluation Criteria For the use case of cable manipulation, ten evaluation criteria have been derived, which form the basis for assessing the technology alternatives. First, the working speed for cable manipulation describes the time it takes to grasp an object like a cable, move it to the correct position, and assemble it as desired. Second, process robustness expresses how capable an automation system is dealing with significant variances of limp objects’ shapes, positions, and behavior. Third, quality covers precision and reproducibility of the assessed technology. Precision implies the capability to handle small-diameter cables with high accuracy while positioning and assembly. Reproducibility expresses how accurately a cable assembly task can be reproduced achieving uniform results in terms of connection quality and cable position. Fourth, return on investment describes the time it takes for cost amortization. Fifth, running costs corresponds to the amount of operational expenditure. Sixth, integration effort into production expresses how much effort (e.g. hardware engineering and software programming) it takes to initially integrate a new automation technology into an existing production system. Seventh, adjustment flexibility indicates how well an already installed technology can be adapted with regard to hardware and software changes due to new products. Eighth, the handling of a cable manipulation technology characterizes how intuitive and facile a system can be operated and programmed. Ninth, collaboration represents how well a technology is able to be implemented within a working environment, which excels in active interaction of personnel and collaboration. This also implies the general technology acceptance among users. Tenth, the scalability denotes how well a technology system can be adapted to higher capacity requirements. The evaluation criteria have been derived primary based on the input of expert interviews (System Integrator, 2014; Robotic Startup, 2014a, b). Step 4b: Technology Assessment The assessment has been conducted primary based on the input of expert interviews and observations (Robotic Startup, 2014a, b; System Integrator, 2014; Hartelt, 2015). The comparison of potential alternatives has been executed by considering manual operation, special-purpose automation technology, classical sensor adaptive robotic technology, and the new intuitive robotic technology (cf. Figure 5).

Figure 5: Technology Assessment Cable Manipulation

10

As human beings are very sensitive regarding visual and tactile stimuli, manual operation has several advantages for cable manipulation compared to general automation technologies. Human beings can align oneselves to potential variances of limp objects and can handle them intuitively, which leads to a high process robustness. With manual operation, it is possible to grasp, feel, flex, and push these objects simultaneously within one single motion, which is accompanied by very fast working speeds. Robotic solutions utilizing sensors for adaptive task execution can’t reach this speed level. Despite there may emerge careless mistakes during manual operation also quality is still acceptable. In addition, manual operation is quite flexible regarding task adjustment or process changes and allows intense interaction and collaboration. However, this positive picture gets tainted by relatively high running costs, in particular in countries with high labor costs. In contrast, automation technologies excel in relatively low running costs. However, they require longer break-even periods than manual operation solutions as initial set-up costs are quite high. This is due to the fact that it takes much effort to integrate automation technologies into production settings since their hardware and software need elaborate customization. If variances of limp objects occur, automation technologies face difficulties with cable manipulation, which cause problems to maintain robust process results. Since intuitive robotic technology is more suitable to cope with variances than conventional automation technologies, process robustness is perceived to be higher. As automation technologies usually achieve rather homogeneous process results with few mistakes, quality is high. Since special-purpose technology usually requires the most elaborate hardware and software customization, this technology is accompanied by the highest effort to be integrated into production and lacks most adjustment flexibility. Thus, special-purpose automation solutions are relatively difficult to be scaled up. Due to their structure they seem to be less appropriate for interaction with people and also less intuitive to handle, but can potentially reach high working speeds. Robotic solutions can be easily scaled up, as they consist of standardized basic hardware and require relatively minor hardware engineering. Compared to classical robotics, intuitive robotics excels in a broader range of application without being based on additional hardware engineering. As complex tasks can easily be represented via the software program, robot capability could be better exploited in case of intuitive robotic technology. This is the key for higher adjustment flexibility. Similarly, this reduces running costs. Furthermore, intuitive robotic solutions have advantages regarding collaborative environments since handling is more facile. Step 5b: Performance Impact Evaluation Working speed is very important in the automotive industry, as this branch is highly cost-driven. Moreover, working speed impacts directly the workload capacity of the overall system. As highly automated and linked production lines may not stop due to unreliable automation solutions, the robustness of process execution is very crucial for automotive suppliers and manufacturers. In contrast, quality doesn’t play a major role in this context. Of course, cables and connectors have to be mounted correctly, but the uniformity and the bending lines of the cables are not particularly important. Due to the cost-driven environment within the automotive industry, profitability is one of the most important factors. Correspondingly, running costs and especially return on investment were considered to be very crucial. Furthermore, integration effort is relevant not only since an automation system usually has to be integrated into existing production lines and has to meet cycling times and process interfaces to adjoining workstations. Adjustment flexibility is important in order to be adaptable to product changes or to be able to produce several product variants. Since end customers want to operate an automation system as easily as possible, intuitive handling and operation is perceived to be important. Collaboration on the other hand, was seen to be less important as automation systems may run separately and are in general fairly accepted. However, new automation systems have to be easily scaled up in order to meet potential capacity increases. The performance impact evaluation was primary based on the input of application experts (System Integrator, 2014; Hartelt, 2015; Robotic Startup, 2014a, b).

11

Step 6b: Task-Technology-Fit Conclusion and Customer Utilization The calculation of the overall Task-Technology-Fit value (TTFX) for each technology alternative leads to the following ranking (cf. Table 2): Table 2: Task-Technology-Fit Assessment Cable Manipulation

Cable Manipulation Technology Alternatives (X): Manual Operation Intuitive Robotic Technology Classical Robotic Technology Special-Purpose Automation Technology

TTFX 173 142 112 88

In relation to Intuitive Robotic Technology 122 % 100 % 79 % 62 %

Manual operation achieves the highest TTFX-value and therefore seems to be most suitable for the application cable manipulation. Consequently, the likelihood to utilize manual operation is relatively high within the automotive sector. Considering the automation technology alternatives, robotics seems to have clear advantages compared to special-purpose automation technology. If cable manipulation tasks should be automated, intuitive robotic technology correspondingly tends to be most likely to be utilized. Compared to conventional automation technologies, manual operation and similarly intuitive robotic technology exhibit several advantages that are often perceived to be highly relevant. Therefore, these advantages (e.g. process robustness or adjustment flexibility) lead to a relatively high performance impact and enhance the likelihood of utilization accordingly. Nevertheless, several disadvantages of manual operation (e.g. high running costs) and intuitive robotics (e.g. limited working speed reduces profitability) are perceived to be important just the same. Therefore, these characteristics impact the utilization probability of the corresponding technology negatively.

C. Cross-Case Conclusions In general, manual operation exhibits the highest TTFX-values of all technology options and has clear advantages regarding the evaluated criteria. However in case of laboratory automation, the manual operation has decisive disadvantages when it comes to quality and cleanliness issues. In case of cable manipulation, the emerging running costs and scalability requirements are disadvantageous for manual operation. These reasons may lead to the demand for automating these activities. This situation represents the main opportunity for the intuitive robotic technology. This is due to the fact that in both use cases the intuitive robotic technology has shown higher TTF X-values than the alternative automation technologies. Particularly regarding the overlapping evaluation criteria adjustment flexibility, process robustness, handling, and collaboration, the intuitive robotic technology excels. This is mainly caused by the fact that the effort for software programming can be significantly reduced during task implementation by utilizing the intuitive software technology. However, these advantages are limited if compared with manual operation. Although robotics seems to have reached the next stage by utilizing intuitive software technology, it still couldn’t close the gap to humans’ capabilities. Nevertheless, there is a chance for automation within these application tasks. Especially for cost-driven industries if the factors running costs and profitability excels. The application of the TUM helps in both cases to identify, structure, and summarize relevant key information. Since it is a qualitative analysis, the exact assessment values should not be interpreted to be highly precise. The TUM is meant to reduce complexity and point out the main issues. Therefore, the visualization diagram pointed out the main advantages and disadvantages of each technology alternative in the two application fields. Attention must be paid to situations when one characteristic outshines all remaining factors (e.g. running costs in case of cable manipulation). Its assessment is merged to the overall TTFX-value and its outstanding impact risks to fade.

5. Conclusion & Discussion The paper has both scientific as well as highly practical relevance. On the one hand, it fosters the understanding of technology management and contributes to the scientific community by generating new knowledge in this field. The TUM has been empirically validated by conducting an in-depth multiple-case study. On the other hand, the paper will be of great benefit for companies that face the challenge of commercializing a new technology as it presents a

12

hands-on guideline for evaluating the task-technology-fit. By applying the TUM, these companies can assess if the tasks of a potential application field fit to the respective technology and should be addressed. Nevertheless, there are several concerns and limitations with respect to the conducted research. First, a common concern about case studies in general is that they provide little basis for scientific generalization as they just analyze case specific phenomena. Hence, the base for broad generalization would be quite thin. According to Yin, the short answer to that issue is that case studies are generalizable to theoretical propositions like the TUM and not to populations (Yin, 2014, pp. 15-20). Second, the sample of the current multiple-case study design contains just one technology and two correlated use cases. Thus, the transferability of the results may be questioned. Correspondingly, a natural suggestion for future research is to increase the number of cases. Third, within the current study, the TUM has been tested just for the robotic branch. We assume that the TUM could be transferred to other branches and industries as well. However, this has to be empirically tested by conducting primary case studies. Fourth, the TUM does not provide a method to detect potential use cases. A precondition for applying the TUM is that potential applications for the respective technologies have been preselected. The TUM provides a guideline to assess if the task of the selected application fits to the concrete technology.

References Automated Systems of Tacoma (AST) (2010), “Restricted Access Barrier Systems (RABS) & Isolators: The Perfect Combination of Robot System Safety and Aseptic Drug Manufacturing”, whitepaper. The Boston Consulting Group (BCG), (2015), “Raffaello D’Andrea on the Future of Robotics”, https://www.bcgperspectives.com/content/videos/technology_strategy_technology_business_transformation_dandre a_raffaello_future_robotics/. Bullinger, H.-J. (1994), “Einführung in das Technologiemanagement: Modelle, Methoden, Praxisbeispiele“, Teubner, Stuttgart. DENSO Robotics Europe (2014), Leaflet: “New VS Series H 2O2/UV Resistant”, https://www.densoroboticseurope.com/sites/default/files/file/Downloads/DENSOnewVSH2O2122014V3EN.pdf. Gibson, D.V. and Smilor, R.W. (1991), “Key variables in technology transfer: A field-study based empirical analysis”, Journal of Engineering and Technology Management, Vol. 8 3–4, pp. 287–312. Goodhue, D.L. and Thompson, R.L. (1995), “Task-Technology Fit and Individual Performance”, MIS Quarterly, Vol. 19, No. 2, pp. 213-236. Gruber, M., MacMillan, I.C. and Thompson, J.D. (2008), “Look before You Leap: Market Opportunity Identification in Emerging Technology Firms”, Management Science, Vol. 54 No. 9, pp. 1652–1665. Hartelt, R. (2015), Several expert interviews at the industrial fair “Hannover Messe 2015”, conducted by R. Hartelt. Henkel, J. and Jung, S. (2009), “The Technology-Push Lead User Concept: A New Tool for Application Identification”, http://www.econbiz.de/archiv1/2010/106892_technology_lead_user.pdf. Keinz, P. and Prügl, R. (2010), “A User Community-Based Approach to Leveraging Technological Competences: An Exploratory Case Study of a Technology Start-Up from MIT”, Creativity and Innovation Management, Vol. 19 No. 3, pp. 269–289. KUKA Robotics GmbH (KUKA) (2014a), “Adaptive content/themes/kuka_microsite/modal.php?ID=714.

Assembly”,

http://www.kuka-lbr-iiwa.com/wp-

KUKA Robotics GmbH (KUKA) (2014b), “Specification LBR iiwa”, KUKA Laboratories GmbH, http://www.kuka-labs.com/NR/rdonlyres/5F941729-D36E-4012-B1E4C58E93D45C4F/0/Spez_LBR_iiwa_7_en.pdf. Leoni 2011, Interview with Probst, K., CEO of LEONI http://www.elektroniknet.de/automotive/sonstiges/artikel/82538/.

AG

(conducted

by

S.

Janouch),

Mülhardt, C. (2009), Der Experimentator: Molekularbiologie/Genomics, Spektrum Akademischer Verlag, 6. Auflage, Heidelberg. Müller, H.-J. and Röder, T. (2004), Der Experimentator: Microarrays, Elsevier, Spektrum Akademischer Verlag, 1. Auflage, München.

13 National Center for Tumor Diseases Heidelberg (NCT) (2013), “Connect. Das NCT Magazin”. National Center for Tumor Diseases Heidelberg (NCT) (2014), Several expert interviews with bioanalytic scientists, laboratory visits, website, observations, conducted by R. Hartelt from October 2014 to March 2015. Poremba, C. (2014), „Automation in der Pathologie: Keine Angst vor dem ‚Roboter‘“, in: „Wie viel Automation braucht die Pathologie?“, Trillium diagnostik, Zeitschrift für interdisziplinäre Medizin, Sonderdruck aus Heft 1 – 2014, p. 2. Robotic Startup (2014a), Internal documents, website, press information and observations (company name is disguised), protocoled by R. Hartelt from October 2014 to May 2015. Robotic Startup (2014b), Several expert interviews and technical discussions with CEO of Robotic Startup (concrete company name could not be disclosed), conducted by R. Hartelt from October 2014 to May 2015. Robotic Startup (2014c), Several expert interviews and technical discussions with Sales and Marketing representative of Robotic Startup (company name is disguised), conducted by R. Hartelt from October 2014 to April 2015. Rogers, E.M. (2003), “Diffusion of innovations”, 5th ed, Free Press, New York. Rothwell, R. (1992), “Successful industrial innovation: critical factors for the 1990s”, R&D Management, Vol. 22 No. 3, pp. 221–240. Souder, W.E. (1989), “Improving productivity through technology push”, Research-Technology Management, Vol. 32 No. 2, pp. 19–24. System Integrator (2014), Several expert interviews with System Integrator (company name is disguised), conducted by R. Hartelt from December 2014 to March 2015. Tellis, W. (1997) Introduction to Case Study. The Qualitative Report. http://www.nova.edu/ssss/QR/QR32/tellis1.html (accessed on January 18, 2013). Trott, P. (2005), “Innovation management and new product development, Pearson education”, 3. ed, Financial Times Prentice Hall, Harlow [u.a.]. Ullman, D.G. (2003), “The mechanical design process”, 3rd ed., McGraw-Hill, Boston, Mass. Utterback, J.M. (1971), “The Process of Technological Innovation Within the Firm”, Academy of Management Journal, Vol. 14 No. 1, pp. 75–88. Walker, A. and Ellis, H. (2000), “Technology Transfer: Strategy Management, Process and Inhibiting Factors. A Study Relating to the Technology Transfer of Intelligent Systems”, International Journal of Innovation Management, Vol. 04 No. 01, pp. 97–122. Yin, R.K. (2014), “Case study research: Design and methods”, 5th edn., Sage Publications, Los Angeles.

Suggest Documents