Descripción de los componentes arqueológicos

By comparing the initial questionnaire-based pre-study in ADR Cycle I, with the results from the field workshops and interviews, a high degree of consistency between the study participants’ perception of problems and challenges in the field and the envisioned merits of the artifacts were found. Both, the pre-study and the actual on-site evaluation, highlighted similar issues in supply networks and thus indicate the validity of the two defined design principles in terms of potential performance benefits. DP1 even received greater support in comparison with the pre-study. Prevention of document exchange between users and systems, and collaboration on networked business objects were mentioned as ways of avoiding system discontinuities, avoiding information redundancy and complicated document management, reducing the number of necessary interfaces between systems, and ensuring a harmonized information base. It was also perceived that DP1 leads to avoidance of unsynchronized document versions, standardized supplier collaboration and a more transparent supply management process. All the dependent aspects were named as key factors for performance increases.

The question of compliance with current financial and tax laws was raised however by no longer creating several documents for an ordering process and storing them in each involved on-premise supplier relationship or customer relationship management system. The question is raised for example of legality if the order as a legally binding document is only kept once in a cloud environment rather than twice as a purchase or sales order in the on-premise SRM or CRM. One reason why this might be legally possible is that the financially relevant document, the account payable and receivable bookings in the finance management applications of both buyers and sellers companies normally need at least one unique reference to a supply management document (in case the business transaction implies a stock or supply management related activity). With the proposed design, this would be provided by the unique identifier of the n-BO of type ORDER for both buy and sell side financial documents. Another reason is that similar concepts are already deployed, for example in popular forward auction platforms like eBay. The related activities there, such as offers, bids, payments, and the financial bookings, are related to a single, unique article number on the auction platform. This issue with regard to storing transactional data in cloud-based or on-demand system landscapes does need further clarification, compliance investigation, and research however.

Further challenges were mentioned by the supply network experts regarding the implementation of DP1 in productive system environments. Despite interviewees

7 Discussion 116

agreeing in general with the benefits of the design principles DP1, some expressed a need for extensive change management with this new approach. To mitigate the risk of implementation failures because of the fundamental change in executing business transactions, how to establish business relationships and conduct collaborations, the proposal is to start with smaller sets of business transactions, for particular categories and organizational units. Based on the learnings and refinements of the business processes, broader intra- and inter-organizational rollouts can be started. In any case, introducing the proposed design to a complete organization all at once (a ‘big bang’ scenario) would not be recommended.

The most significant indications from the research were found regarding DP2. The results revealed that the design decision for social network elements were very well perceived as providing strong potential to address key challenges in supply networks, especially in sourcing and the need for full supply network process coverage. From the user-centered perspective in particular, the interviewees anticipated significant performance improvements by seamless navigation from unstructured activities, like contact initialization, communication, supplier qualification and on-boarding to activities around structured business objects, like quotes and orders. They also confirmed the potential to comprehensively support their various activities and increasing responsibilities without being hindered by process breaks and defragmented system landscapes.

In this respect, privacy and data protection played an important part in the interviews, especially challenges in terms of who is allowed to access which personal data, including relationship data (i.e., not everyone should know with whom a person is in contact). The suggested three level privacy concept in B-Zone of general public data, private data for groups/teams and private data specific to business transactions (e.g., one-on-one relations between buyer and supplier in concrete business transactions) was stated as sufficient in the first place. Another privacy challenge which came up was regarding attrition: who owns the relationship information when an employee leaves the company? If employee leaves the company, is s/he allowed to retain possession of the business related contacts, or do these contacts have to become the full property of the former employer company. Of course, the answer to this question depends on the jurisdiction of the country in question, but there seems to be a common consensus that the contact relation is normally owned by the employee, even after leaving the company they established the contacts for. The reasoning is that interactions similar to the ones in public social media are primarily related and driven by natural and not by juristic persons. According to the German interpretation of law for example, ownership depends on certain parameters, such as who originally created the user account, the type of social network, who carries potential costs related to the account, whether the name of the company is contained in the name of the account, whether the email address associated with the account is owned by the company, whether the account primarily uses privately or for business. It is possible however and in fact recommended for employers to codify in the work contracts or amendments that contacts and content ownership remain

organization (Ulbricht 2012).

7.1.2 Discussion of ADR Cycle II Results

Summarizing the results of ADR Cycle II, hypothesis 1 that the artifact results in higher task efficiency is supported by the experiments, thereby fulfilling the fundamental objective of this research, namely that the artifact with its underlying design principles improves individual efficiency. This result can be generalized using causal explanations and surface similarities (Shadish et al. 2002). First, there are no differences between subjects in the experiments and the desired target group. Supply management experts participated in the experiment, which can be regarded as a representative sampling of one of the two target groups of the supply network project (besides suppliers), as they came from different companies, operated in different departments and purchased different goods. Strictly speaking, the snowball and judgment sampling method does not constitute a truly probabilistic sample (Bhattacherjee 2012). Compared to a simple convenience sample however, this expert selection has a higher external validity and provides relevant insight into the solution to the problem.

Secondly, individual performance, which is measured in this experiment, is a part of group performance. Unless the performance of suppliers at the other side of a supply relationship would decrease by using a network like the artifact, it is reasonable to assume that the overall network performance would also improve when using the artifact as a supply network environment.

Finally, if design principles, n-BOs and social augmentation reduce the time needed to perform a supplier qualification use case, they could also help to reduce time in other use cases where the combination of different types of data and exchange of documents play an important role.

Hypothesis 2 that the artifact results in higher task effectiveness could not be supported despite there being a higher quality indication using the artifact, but no significant difference in the mean values of quality compared to the comparison tool. The average quality score achieved was relatively high, with a score of 3.54 (comparison tool) and 3.58 (B-Zone artifact) out of 5, but did not differ much between the comparison tool and the artifact. There are a number of explanations for this. One reason for the means could be that the use case designed for the experiment was finally not complex enough to yield substantially different results in terms of quality. The assumption that the supplier qualification task in general imposes high intrinsic cognitive load thus continues to hold true, but the intrinsic load in this designed experimental task might not have been high enough to influence quality. It is highly likely that individuals tend to adjust the time effort rather than to compromise on a high quality outcome. Nevertheless, complicating the use case by increasing the amount of suppliers to be qualified and adding more decision rounds to it would have easily exceeded the original time plan of maximum 90

28

7 Discussion 118

minutes per experiment. Some individuals could also still have compensated the higher complexity with a higher time investment. Related to this is the fact that quality is an aspect of performance that is very difficult to control in an experiment, on account of its high level of subjectivity. From the questionnaires at the decision points, it appeared that while the question of where the supplier was located (Singapore as opposed to Kuala Lumpur for example) was more important for some participants in the experiment, others considered it more important for suppliers to be rated well or recommended by others, or for some suppliers to be more experienced in their job (senior trade agents as opposed to junior trade agents for example). Some even stated nationality, native language or gender as decisive factors. Also the medium- and long- term performance of suppliers as important factor of supplier selection quality, for example regarding medium- and long-term responsiveness, quality of supplied services and goods, the actual execution quality of supply processes (e.g., in time delivery), medium- and long-term collaboration behavior and others play an important role in the ongoing supplier evaluation. It is obviously impossible to include these factors in an experimental setting focusing on the initial supplier qualification.

Hypothesis 3, the artifact results in lower mental effort was supported by the results of the experiment. This means that B-Zone and the design principles therein significantly reduce the mental effort and consequently cognitive load that has to be invested for a given task compared to a comparison tool. By ensuring that different types of structured and unstructured data are tightly connected and displayed in one platform, and are not distributed in different applications, such as email, enterprise resource planning and supplier-relationship management systems, the design principle social augmentation reduces the information split that is responsible for cognitive (over)load. Consequently inappropriate information presentation and information disperse are reduced. N-BOs, as the first design principle supports users in automating administrative tasks and providing a single version of truth of data documents (such as qualification criteria sheets). Accordingly n-BOs avoid redundant and disperse information sources, help to reduce split attention effects, thereby freeing cognitive space for learning and important decision tasks.