1
$\begingroup$

I have an (offline) advertising campaign that I'm running in one city. I'm trying to figure out how to answer the following questions:

  1. What are the chances that the advertising campaign has no positive effect?

  2. What are the chances that the advertising campaign increased sales by at least X%?

I have historical monthly data for the sales going back 6 months. Given that these are small samples, what are the best statistics to use here? I expect the advertising campaign to have a significant effect, so ideally I'd like to be able to say something even with one new data point (one month of data.)

$\endgroup$
  • $\begingroup$ Do you got sales development from other cities to compare with? $\endgroup$ – Michael M Jun 30 '14 at 7:04
  • $\begingroup$ Yes, I can compare to other cities or to historical performance for that city (or both.) $\endgroup$ – Satvik Beri Jun 30 '14 at 10:32
1
$\begingroup$

Because your data is temporal in nature, there really isn't anything significant that can be concluded. You definitely need some kind of a control group (ie. a group of users who were not subjected to the advertising campaign in any way), with both groups being studied in the exact same context (same period, and absolutely no bias other than said advertising campaign).

If you have no access to such data, there really isn't a lot you can say out of 6 points of time-series data.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.