Distribution to model time between web page accesses

[oneamp]

I have a formula to get an average time that an average user spends reading a web page before going to the next one, that takes as input the number of words on the page. I want to use this in conjunction with a probability distribution to try to accurately model an 'average user' that is not exactly average; ie. a real user. Is a random exponential distribution appropriate to use in this context? If not, what are the alternatives and their benefits?

Thank you

 

[NascentOxygen]

oneamp, it sounds an interesting approach. You'd be likely to get more specific suggestions if you asked on one of the maths sub-forums. This is sure to be something that others have spent time on.

Good luck.