You're talking about two things: probability and "chance". Probabilities are fixed - probability of throwing a 1 when you toss a one sided dice. Probability of winning the lottery. Probability of having a boy or a girl (well this might be affected by a ton of other things - I dunno). In general though, probabilities are something one could calculate without ever having to perform the experiment.

Chance though refers to some observed and studied result. 43% of Americans are against government healthcare. 32% of all white female drives 18-24 will have a car accident this year. Does that mean if I gather three white females 18-24 into a room, one of them is guaranteed to have a car accident? No. That is the difference between "chance" and probabilities: in probabilities an outcome MUST occur. With chance, you could gather 100 females, and not a one will have had an accident. You would EXPECT 32 of them to have had accidents, but it is not a guarantee.

What does that have to do with your question: well what you are describing is more chance than probability. I mean there is a probability a single product will have Ingredient X - 50%. It either does or it doesn't. However the chance someone buys an Ingredient X item is determined by so many things, not the least of which is the volume of product they purchase. How would you represent that mathematically - who knows. Perhaps a linear regression line: as volume of food increased, chance of purchasing Ingredient X product also increases:

Where "y" would be chance of buying Ingredient X, "x" would be volume of food purchased, and beta 0 and 1 would be slope and y intercept (referrred to as different things when talking about regression, but that is what they are) of this line.

Hopefully that answers (part) of question, or gives some food for thought.