In my own experience, building a spreadsheet has been more of an application of mathematical concepts, or reinforcement of those concepts, rather than the learning of new mathematics. (Although I have certainly seen the reverse--e.g., graphing trig functions using the spreadsheet's graphing utility).
I spent a great deal of time as a lawyer practicing in the field of trusts and estates. Much of what we do in that area consists of very sophisticated tax planning, involving frequent computations of estate and gift taxes, present values of streams of future payments, etc. Once we all got computers at our desks and spreadsheet software (by the early to mid 1980's), it became obvious that the spreadsheet was an answer to our prayers for doing the kinds of computations that we do, especially with its ability to do iterative computations at enormous speeds (for our so-called "inter-related computations").
I became the unofficial Lotus "maven" in our group, probably because I very quickly learned that every hour I spent learning the techniques, formulas, and built-in functions of the software would be repaid a hundred-fold, if not more. I can still remember a very simple spreadsheet I did for one of the senior tax partners, which really did nothing more than compute simple and compound interest, maybe with a few bells and whistles. One would input the starting and closing dates for the running of interest, the annual rate, the principal amount, compounding periods/year, etc., and the spreadsheet instantly produced all the answers.
My boss and I were showing this to the tax partner, who was a big wheel in this firm who billed out his time (way back then) at maybe $500/hr. He was just amazed. "Mr. Silverman," he said, "have you had graduate work in mathematics?" "No, Mr. Wolf," I replied. "In fact, the mathematics involved in this spreadsheet is nothing more advanced than I learned in first-year algebra with Mr. Monteverde in 8th grade!" He wanted to see the "guts" of the spreadsheet, so we looked at the formulas, not much more than i=prt and FV=PV*(1+r)^n. And, of course, a simple subtraction to get the number of days for which interest runs. I reminded this guy (who was very nice) of those problems he most likely did not much enjoy, in which Susie's age was 1/3 mother's age, but in 10 years . . . (fill in the rest). How old are Susie and mother now? I pointed out to him that by doing those problems, you develop that skill for converting facts stated in English into algebraic expressions, so they really did have a purpose (as Mr. Monteverde used to assure us). One uses that skill constantly in creating non-routine spreadsheets.