Approaches to Effective Estimate in Project Management

Effective estimating is the key to a plan in which one can have a degree of confidence. It is also one of the hardest parts as it involves making judgements based on knowledge, understanding and experience. As such, different – and sometimes conflicting – interests come to the fore.

Estimates will always be subject to human interpretation of these factors, and the Project Manager should ask himself /herself the following questions:

If estimates appear too long

  • Is the estimator trying to give himself/herself more time than he/she really needs to make life easier?
  • Is he/she unnessarily concerned (e.g. through lack of experience) about the complexity of the task?
  • Is he/she planning to deliver what you would consider to be an over-engineered solution (e.g. a 200-page report when a one-page memorandum would suffice)?

If the estimates appear too short

  • - Is the estimator over-confident?
  • - Does he/she really appreciate the complexity?
  • - Does he/she fully appreciate what must be delivered?
  • - Has he/she made unreasonable assumptions?

To take account of the ‘soft’ human factors mentioned above, we should try to apply a degree of objectivity and empiricism to the process. No method of predicting the future is fail-safe, but by ensuring that we are taking informed judgements on the basis of sound data and reasonable assumptions we can reduce the risk. There are many different techniques which can be used for estimating, each more or less appropriate to different circumstances. However, you should note that these can never be a substitute for judgement and experience. Some of the main principles to be observed when estimating are described opposite:

Principles of estimating

  • - Estimates should be in terms of days effort. The elapsed time taken to complete a task is a product of the resources which can be applied to it and other constraints
  • - Estimates should not include any allowance for contingency; that can be applied later at a global level
  • - Estimates should be ‘honest’
  • - Individual commitment should be sought
  • - The skill and experience levels of the available staff should be allowed for
  • - The procedure used and any assumptions made should be documented
  • - The process should be revisited throughout the project to ensure that assumptions and factors used in the derivation of estimates still hold true
  • - Estimates should not be ‘massaged’
  • - Always apply a reasonableness check

Estimating using historical data

Historical data about previous projects can be an invaluable source of base data for new projects. Generally, this data will only be available if your organization has well-established planning and control procedures, which have faithfully recorded the time and effort spent on each task during previous projects.

Estimating using relative time

There is frequently a direct relationship between the time taken to do one task and the time taken to do a subsequent one. If it takes two months to design a particular type of product, then it might generally take four months to build it and a further two to test it. Rules which specify these relationships may be well-known within your own industry.

Estimating using parameters

It is possible to derive formulae which can take into account the size of the deliverable and the complexity of the task. Some of these formulae are well-established in particular industries. In Information Technology there are well-known formulae based on the number of files a program will access, the number of decisions to be made within the program, and the language being used. These formulae are often derived and refined using many years of historical data.

Leave a Comment

About | Contact | Sitemap | Privacy Policy | Log in