1. Many empirical studies have validated the original model and its interrelationships,
whereas other studies have recommended enhancements to the original
model. Based on these contributions, we propose an updated D&M IS Success
Model to serve as a foundation for the positioning and comparing of IS empirical
research. The model should continue to be tested and challenged. The
changes introduced in this paper are examples of this continued growth and
refinement. We encourage others to join in this effort.
2. The updated D&M IS Success Model is a useful model for developing comprehensive
e-commerce success measures as demonstrated in Table 1.
3. We recommend that “service quality” be added as an important dimension of
IS success given the importance of IS support, especially in the e-commerce
environment where customer service is crucial.
4. The complex, multidimensional, and interdependent nature of IS success requires
careful attention to the definition and measurement of each dimension
of this dependent variable. It is important to measure the possible interactions
among these success dimensions in order to isolate the effect of various independent
variables with one or more of these dependent success dimensions.
The updated D&M IS Success Model in Figure 3 presents the interdependent
relationships that should continue to be considered and tested.
5. For each research endeavor, the selection of IS success dimensions and measures
should be contingent on the objectives and context of the empirical investigation,
but, where possible, tested and proven measures should be used.
The Seddon et al. [42] context matrix is a valuable reference for selection of
success measures based on context.
6. Despite the multidimensional and contingent nature of IS success, an attempt
should be made to reduce significantly the number of measures used to measure
IS success so that research results can be compared and findings validated.
Some good progress has been made in this area as noted in the
Measurement Enhancements section of this paper. Where possible, we advocate
the application of existing, validated measures rather than the development
of new measures.
7. With the growth of management support systems and the advent and development
of e-commerce systems, voluntary systems use is more common today
than it was a decade ago. We, therefore continue to advocate the inclusion
of “System Use” as a critical dimension of IS success measurement. Actual
use measures should be preferred to self-reported use measures. Also, usagemeasures should capture the richness of use as a system phenomenon including
the nature, level, and appropriateness of use, and should not simply measure
the frequency of use.
8. Finally, more field-study research should investigate and incorporate “Net
Benefits” measures. Yuthas and Young support this conclusion: “[E]xamining
satisfaction and usage measures is not an acceptable alternative to measuring
performance [i.e., Net Benefits] directly. Although the three variables are correlated,
the relationships between them are not sufficiently strong to warrant
their use as substitutes for one another” [60, p. 121]. Good progress has been
made in the development and testing of “Net Benefits” measures on the individual,
group, firm, industry, and national levels