When performing the task of estimating you usually have three primary dimensions that you look at occurrence, timing, and magnitude of uncertain future events. A quick way to remember this is to look at estimating like this: if (occurrence), when (timing), and how much (magnitude).
The occurrence is the highest level in the hierarchy chart for estimating:
Occurrence "Will uncertain future events take place?" If the answer you come up with is no, you do not need to continue estimating. If my answer is yes, you will need to continue to the next level.
So you get your proposal approved so you will need to consider the timing of all project deliverables. So you ask "When will my client request implementation of the application?"
Magnitude=volume. Ask yourself "How much implementation will be required?"
There are also three additional estimating questions you should ask:
1.Cause and effect relationship does it exist?
Cause and effect is not always an assumption, it deserves analysis.
Since there is seldom perfect information at the start of a project, it is rare to have a perfect estimate. Try (this is hard) to identify how imperfect the information while you are making your estimate. If you feel you have imperfect data make as many allowances for error and adjustments to the project that you need.
3.Accuracy of an estimate decreases as time periods increase.
Estimating months in advance is always a tricky subject, and the bigger and longer the project is will create estimates that will become increasingly inaccurate. This is where you MUST update your estimates on a timely basis. BE PROACTIVE if things change drastically, change your estimates be detailed, precise and document them in a timely manner.
Risk management is I believe then the major project management area of concern, I have seen too many projects run longer then they needed to because the risk is not identified and contingency plans are not prepared. A good rule of thumb is that while you are creating your estimate and your estimate has a variation of +50% that is huge and important to plan for if it has a plus or minus of 2% then the estimate will need very little risk management applied to it.