This is a very short article. It describes a simple rule we use for projects for our clients as well as for our internal projects. We use it to determine the scope of work which will be included in any given "project".
We are all familiar with situations when the amount of work is such that it is hard to embrace all the details and create a more or less reliable estimate for various aspects of the project: technical specs, work resources, budget, etc.
Our approach is this: if at the start of the project, 80% of the specs may be conceptualized by everyone involved within the first 10 to 20 minutes and it is perceived as doable within 1 to 3 weeks - such scope is a good candidate for a "project". If a project takes too long to explain, or if it will take longer than three weeks to implement, then it should be split into smaller parts, with the same methodology applied again to each of those.
The mileage will vary in each specific case. In cases of larger undertakings the term "project" may be substituted with other terms, such as "work unit" or "task" or "project step", etc. to recognize the fact the the actual word "project" is already used for something bigger, which consists of these smaller units.