The problem is that a lot of information is needed when determining how much fibre to run. What is the anticipated load? What are your plans for redundancy? What equipment are you using? What expansion plans do you have? What is the budget like? How many clients are there now and how many will there be next year?
For example, if you are trying to get away with old switches using a bunch of 1gig ports in a port group you will need more fibre than a newer switch with a single ten gig uplink. If redundancy is too expensive and not intended to be implemented in the future (bad idea in my mind) then you just cut the needed runs in half. If you do plan on having redundancy and can tolerate half the bandwidth during an outage and are OK with using both links when possible you can also reduce the number of runs. Otherwise more runs are needed. So many variables!
