Travel Reference
In-Depth Information
segmentation is exploratory by very nature, requiring the data analyst to make a number of
crucial decisions in the process (e.g. how many variables are in the segmentation base, which
algorithm is used, which distance measure is used, how many segments are created etc.),
which have a major impact on the resulting segments.
• 65 per cent of marketing managers surveyed believe that market segmentation only leads to
segments if they are actually present in the data. This is also incorrect. The job of a segmen-
tation algorithm is to split objects contained in a data set into groups. The segmentation
algorithm will obediently do so, whether or not segments actually exist in the data.
• 70 per cent of managers state that market segmentation reveals naturally existing group-
ings of consumers. Whilst it is correct that segmentation algorithms will fi nd naturally
occurring market segments in the data, such cases are extremely rare in consumer data.
More typically no segments exist. If no segments exist, the segmentation algorithm will
create some.
• 30 per cent of marketing managers either believe that the number of segments chosen when
the market segmentation task was performed does not affect the nature of the segments or
are unsure about this. Although this is a minority belief among marketing managers, it is
troubling: if the same data set is used to create either fi ve or ten segments, it will clearly have
a major impact on the results, making the decision on how many segments to create highly
critical to the outcomes of the process.
• 30 per cent of managers either believe that the age of the data does not affect the nature of
the segments or are unsure about this. In addition, 27 per cent believe that market segments
stay the same over time. Both age and quality of data are critical to a good market segmentation
solution. A market segmentation solution aims to provide a snapshot of the market at the
time of analysis. Typically, given the time it takes to collect and analyze data, segmentation
solutions are already outdated when fi rst presented. Optimally, new data should be collected
regularly to ensure no major structural changes have taken place with respect to the targeted
segment or segments.
• About one third of marketing managers believe that, if a segmentation analysis is repeated, it
leads to the same solution. This is incorrect. All segmentation algorithms contain random
components and will therefore very likely lead to different results if recalculated.
• Finally, 30 per cent of surveyed marketing managers state that segmentation is independent
of positioning and competition, the other two key considerations of the strategic marketing
planning phase. This is a matter for concern because optimal segments can only be selected
when the positioning or intended positioning of the organization or destination is known,
and when it is clear what the competitive pressures are for the segments under consideration
for targeting.
It is also interesting that 38 per cent of the surveyed marketing managers report that they use
external consultants to run their segmentation studies, and only 32 per cent run their own
segmentation analyses. This may well be due to the fact that 20 per cent admit that they do not
actually feel that they understand the technical details of the market segmentation solution
which was used to derive segments for targeting.
It has to be concluded that nothing much has changed since McDonald's case study of
organizations revealed 'inadequate understanding and inappropriate use of SWOT analyses, the
directional policy matrix, market segmentation, and objective and strategy setting' (1996: 22-3).
There is indeed a substantial theory-practice divide in market segmentation. While the above
fi ndings are a result of surveying marketing managers more broadly, not specifi cally tourism
marketing managers, there is no reason to believe that the situation would be any different
