multivariate regression ?

[lambrechtPJ]

Hi guys,

I would like to look whether gross margin % (GM/net sales) is dependent on several parameters for example volume, weight, product, geographical location of customer ... The problem is that there are several categorical variables (product, geography, ..). I was thinking of doing a multivariate regression, as in gross margin % = a1 + a2* volume + a3* weight + a4 * product1 + a5 * product2 .... + ax * geography 1 ....
with each of the different categorical values (fe products) having a separate column with binairy values.
Is this a correct approach or should I just test each variable separately to see whether there is a correlation?
The thing is I would like to know f.e. if product A is sold in country A and in country B whether the gross margin is significantly higher or not.

Thanks in advance