I’ll take a stab at it, though without a bit more info, I may be barking up the wrong tree.
If the MSS figure is the total documented savings for the customer, I would think rather than using the built in variance feature, you would want to do a calculation of (MSS / Sales) * 100.
So, for example, if I have a customer with $1,000 in sales, and $120 in savings (MSS), then my calculation is going to be (120 / 1000) * 100, which gives me 12%.
Using these same example figures, if I use the built in Variance calculation, I’m going to get -88%. The variance calculation is basically doing ((a-b)/b) * 100. So plugging in my sample numbers, would be ((120-1000)/1000) * 100 = -88.0
Variance is giving you how far one number is off from the other number. Where as it sounds like your simply wanting what percent of one number the other is. So the “MSS / Sales” calculation I would think would be the way to go.
But again, maybe I completely miss understood. If so, any clarification or example figures would be great!