Know-How Exchange : Other

It depends on the size of your ultimate target audience and the degree to which your sample households are representative of the target audience.

If the target audience is retired people, and your sample includes a large representation of young families, then the number is likely to be very high. If the sample is entirely within the group you're trying to reach, then it can be much smaller.

And if you're sampling nationally but expect to need regional breakouts, then you'll need a large enough sample to be able to look at the sub-segments and still have a reliable base size.

Net, you'll probably need a sample of at least 300 who are representative of the target population. The more sub-segmentation you want to do, the more you'll need in the sample in order to ensure large enough sample sizes for meaningful statistical comparisons.

Hope this helps. I haven't done any specific calculations to come up with the 300 number. It's just been my experience that 250-300 is usually a large enough sample for very large populations. The unknown variable is what sub-segments you'll want to use.

As for what marketers would want/require, it depends on the marketers and what their goals are, and how they're going to use the information. If they're local the needs will be different than if they're national or international. If they're B2B it will be different than if they're B2C. If they're using a shotgun approach, the needs will be different than if they're using a highly targeted rifle shot. If they're online it will be different than if they're bricks-and-mortar. Etc. Etc.

Hope this helps.