0
$\begingroup$

Hi I have a table of data with 3 columns: Sales,Cancels and Net revenue. I am triying to perform a Linear regression to check how much the sales and cancels affect the Net Rev but the first thing i noticed is that there is a negative correlation between sales and Net revenue (So the less I sell the more I gain?) Corr(Sales,Net Rev)~-.68

These are monthly time series, where sales and cancels are dollars per month and the net revenue is '000 dollars per month. The revenue is the sum of different incomes but not directly the sum of the sales.

enter image description here

Then I run a linear regression net revenue v.s sales and Cancels (Using Excel) and I see that sales is significant at 99% but again it has a negative coefficient.

So my questions are:

1)Why do I get this negative correlation between sales and revenue? I supposed is an Spurious Correlation. 2)Is there some transformation I can apply to data to get better results? 3)How do I intepret a negative coefficient between sales and revenue?

Thanks in advance!

$\endgroup$
  • 1
    $\begingroup$ Why do you think there is necessarily something wrong with what you have? You should not interpret this as fewer sales causes less net revenue, but rather as they happen (in your dataset) to occur together. That may need interpretation in the context on exactly what your data is. Also note that nobody sold 0$ and all your data is a (relatively) narrow range of sales values, clearly this does not extrapolate arbitrarily far... $\endgroup$ – Björn Jan 9 '17 at 18:45
  • 1
    $\begingroup$ If it's longitudinal record it can also mean the company is getting high proportion of rev per sale dollar, not exactly a bad thing. I agree with the other comment that more context is needed to better interpret the data. $\endgroup$ – Penguin_Knight Jan 9 '17 at 19:07

Your Answer

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

Browse other questions tagged or ask your own question.