If you wish to group people together without using something as complex as cluster analysis you might consider 'Count Commands' which enable you to create targets of people who (for instance) agree with between 3 and 6 attitude statements from a list of 8. Or, maybe, those who use at least 4 brands from a list of 10. For example, women who consume at least 10 Boots toiletries from a list of 20.

If you wish to target key product users - as determined by weight of usage - try using 'Volumetrics' which enable you to create your own segmentation of consumers according to their relative importance by weight of consumption. Volumetrics also allow you to estimate the quantity and average consumption levels within your target audience.

If any TGI database does not give you exactly the answers that you need, think about conducting TGI re-contacts in order to provide the information that is valuable to you. Here Ask Afrika contacts current TGI respondents and asks them additional questions, specific to your relevant business objectives.

With the introduction of biannual questionnaire changes it is increasingly important to ensure that analysis of variables that have not been measured for a whole year is sensibly 'factored' or weighted so that the duration of measurement is taken into account.

Remember, when doing trend analysis within FMCG markets for example, that questions asked and definitions used may change over time. For instance, from October 2001 TGI replaced 'housewives' with 'shoppers' and there was also a large scale shift from 'household' to 'individuals' usage of products.

When your analysis requirements stem from a 'blue sky' brief, try getting a TGI Searchlight run done. This will find the key variables that affect your market by searching the entire TGI database without bias from user/input selection.