In order to build statistically

'accurate' forecasts of which products would work ideally for a person, the questions would need weighted answers.

Some products are "dark" and some are "light", and dark products tend to benefit those with light characteristics,

and light products tend to benefit those with dark characteristics. Therefore if you were to first poll each current

member and ask them to rate each product on a "darkness" scale (1-10), you could take the average of the ratings for

each product and use the data the assign a numerical weight to each product

Next build a set of questions that

indicate how dark or light one's characteristics are (1-10 scale again). This will help define a person's inherent

characteristics, and will be the basis of predicting which products to use.

Lastly, use a group of questions to

help determine what kind goals and reactions a person is hoping for (also based on a 1-10 scale)



I do have

my reservations about this though. I think that anything that deals with human attraction is way to unpredictable to

apply statistical concepts to. I mean, no matter how good the data we collect is, whether a product works or not is

more a matter of a person's ACTIONS. and thats something that we really can't account for statistically.