Log-in | Contact Jeff | Email Updates

Question 631:

My company has been arguing over this topic for quite a while, and I'm looking to put an end to this mathematical question.

We are an ad agency who analyzes marketing events (promotions) for our retail clients. We collect all of the metrics (response rates) in a database to eventually get to a point where we could realistically predict what the response rates will be by category (ie Vertical = Automotive, home furnishings, etc) and marketing vehicle (Newspaper Inserts, postcards, etc).

Needless to say we have some categories in which we have a lot of data, others, not so much. So what I'm asking is, how many marketing events should we have before we can safely rely on the response rate averages (i.e., when is it statistically signficant to rely on these averages)?

Also what type of test could we perform to ensure the averages we are sharing with clients are indeed significant?

Thank you

Not what you were looking for or need help?

Ask a new Question

Browse All 872 Questions

Search All Questions: