An excellent resource for me over the years has been the Excel Developers Listserv. Whenever I've been stuck on an issue, posting a question here will always help me get the answer I need.
Some time ago, one of the members, Craig Hatmaker, posted an interesting note. Craig stated that he had been considering why so often it took longer to complete a project than originally estimated.
This is a very common problem and my guess would be that most people would say scope creep is the reason why it happens. Just do a better job of defining the scope, deliverables, and time frame. Then the actual time to complete will be the same as the original estimate.
Craig came up with a better answer. Essentially, the problem is that when hours are estimated most consultants only consider the time to actually create the spreadsheet. But in fact, this is only one step in the process.
Craig identified seven phases in a project. Here is the list, along with the estimated percentage of time required to complete each phase.
Research - What is the problem? 5-15%
Analysis - What is the solution? 5-15%
Design - How will it work? What will the User Interface look like? 5-15%
Construction - Build the spreadsheet/system. 25-50%
Testing - Identify errors and fix them. 25-50%
Implementation - Training and documentation. 1-20%
Note how the Construction phase, at most, takes 50% of total project time. Realizing this makes it easy to understand why the time to completion for so many consulting projects is greatly underestimated. Also, a good reminder that, before you enter any data into any cell in a spreadsheet, take time to properly plan your project.