Marketing campaigns rely on market segmentation for optimization - to send the right promotional stuff at the right time to the right target. Identifying optimum segmentations is one of the main purposes of market research, and typically done by data scientists.
With direct mail being replaced by email marketing and the recent problems with Gmail, email delivery issues and spam filtering errors, has segmentation shifted from being user-based and content-based to ISP-based?
For us, it is clear that considering Gmail vs. Hotmail users as two separate segments has a far bigger, more positive impact than considering users interested in visualization vs. data integration as two segments, or considering West Coast vs. Midwest, or student vs. retired. Ideally, a double segmentation works best, if the segments are not too small. In our case, the top five segments would be (for US / UK / CA users):
- Gmail, decision maker
- Gmail, tech geek
- Non Gmail, decision maker
- Non Gmail, tech geek
Interestingly, the segment "Hotmail, tech geek" is almost empty. These four segments overlap: we are dealing with overlapping segments, but it's not a big problem. The big issue is preserving high deliverability on Gmail - thus the need to consider it as a separate segment.
Have you experimented with such hybrid segmentations? What are the results?
Originally published on DataScienceCentral.com.