What method would you use?

Hi all,

I have a bunch of product sales data, and the marketing methods (say, group 1 and group 2) used to sell those products. I'm trying to show that products marketed using group 1 methods sell better and faster than those marketed with group 2 methods. So group 1 might be "TV, radio, web, leaflet" and group 2 might be "TV, radio". In other words, I guess I'm looking at relative effectiveness.

Now, my puny brain suggests a Pearson correlation or maybe a t-test but there must be something a little more advanced than that...maybe SEM?

Oops, I should have mentioned, the sales figures are obviously ratio (dollar values or days), but the group methods are dichotomous (1 if TV is used, 0 if not).

Predictably, the 'sales' and 'time to sell' variables are not normal... :(