Spend against views/day. Include days of no spend?

Hi everyone,

just looking at doing some simple linear regression and I've got views on an article. Occasionally we choose to spend a bit of money to promote the article and boost views. I want to see how spending affects our views.

It's pretty simple, but since spending is so on and off i'm not sure. I get a good R-squared value when I include those days but a terrible one when I don't. How I see it is including days of no spend gives a good R-squared because spending definitely boosts our views. Not including those days has a terrible R-squared because the amount we're spending isn't hugely affecting the views we're getting.

If I'm looking to do simple linear regression, should I include those days where we haven't spend any money? If I do include those days, how do I choose which days to include? Do I pick days in between spending? Would I pick days before/after spending has started/ended?

Cheers guys!