CHAPTER 5. Evaluation
5.3 Experiment Details for Proposition 2
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
89
provide critical components to its service value proposition to the customers, hence results in a better consequence.
If applying these explanations to our propositions, we can see through Figure 5.4 that the comparison of setting 3 with setting 1, 2 is showing that interaction-centric SMEs are doing much better in alliance for service innovation than resource-centric SMEs. Accordingly, we can say that our proposition – interaction serves an important role in service innovation and service value creation is supported, because it has the same even better performance- acquire more customers, than the well-recognized resourced base alliance selection approach in this simulation. Hence, we could argue that interaction is an important factor in SMEs alliance building if they pursue better outcomes from the alliance, and are able to go on with our other experiments given this premise being justified.
5.3 Experiment Details for Proposition 2
5.3.1 Experiment Challenges and Design Principles
By proving the importance of interaction in section 5.2 through simulation, we then wish to prove the effectiveness of our proposed interaction pattern adjusting model mentioned in CH4 in helping SMEs analysing and managing their interaction pattern to obtain higher valued service innovation in alliance. In this experiment, the simulation approach will be used again, in consideration of the same condition as the previous experiment – it is hard to find SMEs to observe, requires very long time to perceive the result, and is vulnerable to other factor influences. With simulation, we can observe the outcome of SMEs who apply and who do not apply our model in different situations, and evaluate the usefulness of our model.
5.3.2 Experiment Design Details
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
90
To demonstrate our model within the simulation, two more factors will be added into the simulation process, a) interaction quality degree and b) company resource.
Interaction Quality Degree:
The interaction quality degree refers to the intensity and quality of the interactions between SMEs within the region. With higher interaction quality degree, SMEs shall have better and denser interaction with each other, and shall result in higher possibility of gaining benefits or service innovation insights from the others. During the simulation process, we model this through pretending SMEs with higher interaction quality degree will have higher alliance successful rate, thus will have higher chance to benefit from the alliance.
There are two considerations for adding interaction quality degree factor. The first is because the interview target we’ve chosen is Pillow Mountain Leisure Agriculture Area, SMEs in areas like Pillow Mountain Leisure Agriculture Area were considered as having a more loosely inter-business relationships due to their far distance with others, and are harder to form higher quality and intensity interactions; hence we design the interaction quality degree to model this phenomenon, and intend to see the differences might occur within different given values of interaction quality degree.
The second reason for creating the interaction quality degree factor is that this factor can serve as the experiment method of interaction pattern level we’ve mentioned in CH2 and CH4, in which higher in interaction quality degree will refer to higher interaction pattern level, which can help us to demonstrate and verify our proposed model in the simulation process. In the simulation process, very high interaction quality degree will be considered as having a level 3 interaction pattern with another SME, a normal interaction quality degree will refer to a level 2 of interaction pattern with another SME, and level 1 interaction pattern will be modelled
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
91
by having an inferior interaction quality degree with another SME.
For different configurations in the experiments, a rate of happening higher interaction quality degree will be designed. SMEs under a higher high-interaction-quality degree setting will have higher possibility to have a higher interaction quality degree with other SMEs, which simulates areas where SMEs are very cooperative and more tightly bonded. On the other hand, if the given situation is SMEs are having low high-interaction-quality-degree, than SMEs under this setting will tend to have lower degree of interaction quality between them, which is simulating areas where SMEs are highly competitive, consider other SMEs as an opponent, and seldom interact with other. Given an example, if the rate of higher interaction quality degree is high, two SME: SME A and B will have a very high chance to have higher degree of interaction quality; if the rate is low, then SME A and B will possibly have poor interaction quality. However, SME A and B can still improve their interaction quality degree by using their company resource, which will be explained in the next section.
Company Resource:
The second factor to be added into the simulation is company resource. This factor represents the resource SMEs hold to improve their interaction quality degree with other SMEs; when an SME A wish to enhance its interaction quality with another SME B, it will cost both SME A and SME B’s company resource to fulfil the enhancement. In addition, the costs of improving different degree of interaction quality will be different; while SMEs can relatively have a less expensive cost to reach a level 2 interaction quality degree with another SME, it will be more difficult to improve this relationship to a level 3 interaction quality.
The purpose of this factor is to serve as a limitation, and see the effectiveness of
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
92
our model under this given limitation. Our proposed interaction pattern adjusting model provides an analysis method and guidance for SMEs to decide their strategy to adjust interaction patterns. While given a limited company resource, SME could not choose every other SME to improve interaction quality (interaction pattern) with, so the strategy to choose proper target will be important. We would like to see the different outcomes of SMEs customers obtained by following or not following our model’s suggestion.
Also, the target users (Pillow Mountain Leisure Agriculture Area) of our model are SMEs who are scarce in resource, and we intend to use the different given value of company resource to simulate SMEs with different amount of resource within our simulation process. Relation of alliance successful rate, interaction quality degree and company resource are given in the following table 5.3
Table 5.3 Factors relation
Interaction Quality Degree Alliance successful rate Company resource needs
I Low Don’t required
II Medium Required less
III High Required more
Responding to the proposition 2, 2.A, 2.B, and 2.C, we design 4 settings with different configurations of these 3 factors, listed in the following table 5.4. Setting 1 refers to Pillow Mountain Leisure Agriculture Area, where we defined SMEs there to be loosely related, lesser in resource, and lower in knowledge and ability to form good alliance and benefit from it. Settings 2 refers to places where SMEs are higher in business knowledge (knowing more about how to make good alliance), but are having a loosely related inter-business relationship and little of resource. Setting 3 is for places where SMEs are having abundant of resource, but do not have a good inter-SME relation and not sufficient knowledge of doing business alliance. The last
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
93
setting 4 is for where SMEs have good interaction with others, but no resource and proper business knowledge.
Table 5.4 Details of each setting
Settings Interaction quality degree Alliance successful rate Company Resource
Setting 1 Low Low Low
Setting 2 Low High Low
Setting 3 Low Low High
Setting 4 High Low Low
5.3.3 Experiment Results and Conclusions
The following Figure 5.5~5.8 are the results of our experiment to verify our proposed model under different settings, Figure 5.9 shows the comparison between them. As the previous experiment, the way we define how well our model is by comparing the number of customers the SME could acquire after adopting our model or not adopting.
Figure 5.5 Results of Setting 1
0 500 1000 1500 2000
Original Random Strategy
A ver ag e cus tomer amoun t
SMEs allinace choosing strategy
166% 202%
100%
‧
Figure 5.6 Results of Setting 2
Figure 5.7 Results of Setting 3
Figure 5.8 Results of Setting 4
0
‧
Figure 5.9 Comparison of improvement ratio of average customer acquired under all settings
From Figure 5.5, the results of SMEs who apply our model suggested strategy makes the total customer acquired grows 102%, and the results of not following grows 66%; so while under setting1, following the strategy is better for SMEs.
However, from Figure 5.6, 5.7, 5.8 we can see a totally different situation that SMEs who follows the strategy are not doing as well as those who do not. Figure 5.9 shows a very clear comparison that only under setting 1 can our model do better than not applying, accordingly, we might only able to apply our model within the context of setting1.
To discover the reason of why our model only able to apply to setting 1 ( for where SMEs are less in resource, less acquainted in businesses managing knowledge and more loosely related), we gathered other data from the simulation process to see more details. From Figure 5.10 which represents the average number of customers each SME could acquire under each setting, we can see that SMEs under setting 1 has the lowest average, and setting 4 has the highest. Comparing with the data of
‧
when the average customer is lower, the effectiveness of building alliance grows more important. This information reveals a situation that where interaction pattern adjusting will be useful (i.e., where the average performance of the region is lower), but does not explain why our strategy is not applicable under other contexts than setting 1.
Consequently, we try to find the reasons from another perspective, ability gaining perspective.
Figure 5.10 Average number of customer acquired by SMEs under each setting
During the simulation process, there were two factors affecting SMEs acquiring customers as mentioned before: ability and marketing. The abilities decide the maximum customers this SME could obtain, and the marketing ability influences the percentage of customer the SME could acquire from its able-to-acquire customers.
Hence, the amount of customer acquired should somehow able to be decided by the accumulation of these two factors growth rate.
Figure 5.11 shows the comparison of the accumulation of abilities and marketing.
From Figure 5.11, we can see that only in setting 1, those SMEs who follow our model, their accumulation of abilities and marketing growth rate is higher than those who do not follow; in other words, only under setting 1 the strategy proposed based
0 Average custeomr acquired by each SME
Without resource Random choose With strategy
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
97
on our model, is benefiting the SMEs, which is same as what we discovered in Figure 5.5~5.9.
Figure 5.11 Comparison of the accumulations of abilities and marketing under different settings
However, although the comparison of results under each setting in Figure 5.11 are showing the same trend as Figure 5.9, there are still different discovery found in Figure 5.11. If we cross analyse the improvement ratio of customer amount growth and the accumulation of ability and marketing growth rate (Figure 5.12, 5.13), we can see under setting 1 and 4, the ratio of the accumulation of the two factors growth rate is lower than customer amount growth rate; and in setting2 and 3, the growth rate of two factors accumulation is the same as or higher than customer amount growth rate.
Combining this discovery with Figure 5.9, we can see whenever the effectiveness of following our model is higher or almost same as randomly choosing, the accumulation growth rate of ability and marketing is lower or almost same as customer amount growth rate. In other words, when the elasticity of accumulation growth rate with customer growth rate is high, our strategy is more effective. Figure 5.14 shows the comparison.
0 0.5 1 1.5 2 2.5
Settting1 Settting2 Settting3 Setting4
Accumulation of abilities and marketing growth rate Witthout Resource
Random Choose With Strategy
‧
Figure 5.14 Comparison of relationship of improving accumulation growth rate (ability and marketing) and customer growth rate under randomly choosing and
following strategy Customer amount grwoth rate/ accumulation grwoth rate
Random Choose With Strategy
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
99
This discovery reveals when our proposed model is effective and when is ineffective. First, while the effectiveness of gaining benefit from other SMEs through high interaction quality degree is lower, which means the difference between high and low level of interaction degree is smaller, the effectiveness of our strategy is lower (setting 1 compared with setting 2). Secondly, when the cost to achieve higher interaction quality degree is relatively lower (setting1 compared with setting3), the effectiveness of our strategy is also lower. Lastly, when the scarcity of higher level interaction quality is lower; in other words, when it is relatively easy to have a high interaction quality with other SMEs, our strategy will also be less effective (setting 1 compared with setting4). In contrast, when the effectiveness, cost and scarcity of interaction quality are all relatively high, SMEs are suggested to apply our model and follow the suggestion it gives.
This discovery is aligning with the design of the proposed model. In our model, we already presume the scarcity of higher interaction quality (interaction pattern) is high by defining that a level 3 interaction pattern could only be achieved when the all questions within the questionnaire (Ch4.3) are responded “high” in Figure 4.4. Also, there are only very few steps or improvement required for an SME to improve to level 5 service innovation value from level 1, which somehow can show the concept that each improvement is highly difficult and costly, and is in accordance with our discovery. For the last discovery, interaction is a crucial element in business is the fundamental concept of our theory, so our discovery also adheres to the design principle of our model. In brief, the discovery of our model is benefiting SMEs while the effectiveness, cost and scarcity of interaction quality is high not only shows when and what situation is appropriate to use our strategy, but also verified it is well-aligned with the design idea of our model.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
100
Concluding, the usability of our model must be evaluated by comparing the customer growth rate and the accumulation of growth rate of ability and marketing, if possible. When customer growth rate is lower or equal to the accumulation of growth rate of ability and marketing, which indicates that one of the factor that influences interaction quality: effectiveness, cost and scarcity is low, our proposed model will become inappropriate for SMEs to comply with. However, if the effectiveness, cost and scarcity of interaction quality is high, which will makes customer growth rate higher than the accumulation of growth rate of ability and marketing, our proposed model will be very useful to SMEs to improve their service innovation value.