Like the fax and e-mail, MDS are a network-enabled innovation. And, in the sense that they have the potential to alter industry structure, MDS are disruptive (Christensen, 1997). From the demand perspective, MDS have the power to "restructure the patterns of everyday life", a class of innovation first identified by the business historian Braudel (1985). By creating new freedoms rather than mere conveniences , MDS have the potential to create entirely new markets (Keen & Mackenzie, 2001). Yet operators are understandably reluctant to commit resources until they can clarify the nature of the demand for MDS.

We can apply motivation theory to interpret our findings. Non-gamers noted a lack of (1) time, (2) interest, and (3) attraction towards gaming, which would be the source of any intrinsic reward. However, many appear to be drawn to functional uses of MDS, an extrinsic motivation. Casual or social gamers may also see mobile entertainment as a way to be "part of the crowd ", and service providers can market entertainment as a "collective activity" to motivate these subscribers. For example, packaging and marketing of entertainment with competitive pricing and group connectivity, to emulate LAN gaming centers. Although we have no firm data as yet, this seems a plausible direction to seek empirical evidence to identify the characteristics of this unique segment.

By mapping the segments revealed in these studies to the motivational and social needs of subscribers, we can synthesize a two-dimensional needs-based marketing model. One dimension captures the expectation of an extrinsic versus intrinsic reward, while the other captures the individual versus collective nature of this reward.

The diagram in Figure 2 arrays revealed segments according to the needs subscribers seek to meet through MDS use. The numbers in each quadrant refer to the sequence in which segments are likely to adopt, while the segments for mobile gaming are in italics. Note that Lifestylers (and casual gamers) do not exhibit the same clear needs patterns as other segments. They are essentially followers, or, as described by Rogers (1995), majority adopters.

Figure 2: A needs-based marketing model

The approach to data gathering we have used in this work, derived from the "information acceleration" methods advanced by Urban et al. (1996), identify specific target groups more purposively than simple demographic models. Segmentation by linking conventional demographic variables such as age, income, and occupation to service preferences is only temporarily useful, as the underlying structure of the segment is unstable. The structure shifts as subscribers learn because service preferences shift rapidly as new MDS devices, infrastructure, and services appear and as the adoption scenario matures.

Summing up, the limitations of qualitative research are offset by the advantages of survey research, and the two techniques are complementary. Whereas the focus group sessions lack empirical validity, the survey findings largely support the findings from our focus groups.

Segmentation based on psychographic and behavioral factors appears more durable, provided the research model correctly reflects the underlying structure linking these factors. In two of the three survey-based research projects, the survey instrument was grounded in theory, then validated through exposure to focus groups. The data collected through these instruments yielded segments that appeared in subsequent research.