Types of Segmentation Analysis

Basic media segmentation is the most elementary form of segmentation analysis. Direct marketers must evaluate individual lists, publications, or stations, but not entire campaigns. In any campaign, some media will do better than others, some will breakeven, some won't. Some will hit sales objectives, some won't. The analysis is often done using CPR (cost per response) or CPL (cost per lead) and then the programs are ranked in order of CPR. Those with the lowest CPR are listed at the top and those with the highest at the bottom.

Internal segmentation

involves analyzing a list or media by its inherent characteristics, such as geography in the case of publications or any of the characteristics that a mailing list may be segmented on, like sex, RFM, etc.

Direct mail and email are especially responsive to internal segmentation.  With both house lists and rented lists it is possible to break down results into a wide variety of segments.  Any available characteristics can be analyzed.

External segmentation

Refers to the analysis of external characteristics that are not intrinsic attributes of the media.  The simplest one is identification of geographic areas of a market area that produce significant variations in response.  The most scientific application of this approach is cluster analysis popularized by Claritas.  All the obvious demographic characteristics available from the US Census can be applied as well as other enhancements added by Claritas and others. 

Deciles analysis and segmentation

Is a refinement of segmentation analysis and it is the practice of statistically analyzing the responsiveness of each segment of a mailing program.  You divide the segments into 10 groups, each representing 10% of the total.  You then list them in descending order of profitability. 

Dimensional segmentation

Is the practice of taking the decile groups and breaking them down even further into a third dimension which might involve geography for example.  It is most often combined with geographic analysis.  It’s theoretically possible to refine mailing list results on a three dimensional base.  For example comparing and weighting results for the list based on demographics and other factors such as size of market or weather.  House files can be refined based on R, F, M (Recency, Frequency, Monetary).

Regression analysis

is a process that compares the known characteristics of your customers with the characteristics of other mailing lists.  Mathematically, the matches are noted, counted and compiled.  The result is a “penetration analysis”—a ranking of those characteristics that were matched most frequently and the other list which shared them.