I've finally received buy-off to take a baby step in the direction of implementing some project management.
The people I'll be relying on for data have never had to estimate (read as: commit to) task durations. I get blank stares and shoulder shrugging when I try and walk them through a process.
Without any historical data to fall back on, and *knowing* that there will be buffer time up the wazoo added... are there any effective approaches to creating realistic estimates from which to baseline a project by?
Yes, I believe in Parkinson's Law... but right now I'm getting ranges, instead of estimates. For instance -- I ask a developer how long it'll take to code xyz and he gives me a range of "a couple of hours to a month."
Any tips on this would be great. I have a wee bit of authority to enforce some guidelines if I watch my p's and q's and can justify the added value.
I have done that before. Works pretty well for fine tuning the range. I once had 3 senior people do an estimate. There was a 100% difference between the high and low!
More sources can only help an estimate. Saving Changes...
In the absence of historical data, the best way to get a head thru is to use the Industry standard figures for productivity / language wise OR skill wise.
The first thing to derive effort estimates is to do a size estimation first. Then break that thru modules and apply the industry productivity factors with a variation of +or- 15%. I am sure it'll help.
The second way is to do a ACM (Activity complexity Model). But for this you need to have a rating of the exsiting skill levels based on experience. You can model this methodology to generate the estimated LOC too.
The third is a wide band delphi method. This could be used in the absence of size estimates and skill level ratings.
As Jim has put it you can also try out PERT and believe me the book "Total project control" by Stephen D really helps understand this topic.
Need help...contact me. Saving Changes...
Amy SnowHeartland Payment SystemsDenton, Tx, United States
Can anybody recommend a good software product that can assist in project estimation? Saving Changes...
Frank WintersPhotographer and ConservationistSandwich, Ma, United States
The project that is almost always under estimated is the esitmation of projects itself. Development of a solid estimate for a major undertaking is not a trivial task nor can it be done without adequate staffing. A key to successful estimation is cultural and managerial support. Whatever method is used it must be accepted as "correct" by all those involved. From a cultural viewpoint this includes all stakeholders -- particularly the end beneficiaries of the project. The education of the organization is key to this goal and while difficult the process starts with a single step, (and a plan) as always.
One method that works in some organizations involves triangulation -- the use of three separate methods that are then rationalized to create a very solid final result. The three methods some use are Top Down, Bottom Up and the use of a tool. Top down requires that a high-level understanding of the systtem's functions, interfaces, risks and so on be developed. Then an estimate is developed by an experienced team doing independant estimates (2 or 3. This estimate depends on rules of thumb and experience. The bottom up estimate is done by a separate team who independantly break the system down to semi-atomic parts such as function points or lines of code, then use experience or industry standards to estimate the effortfor each piece part. There should be multiple estimates at this level as well with a rationalization process to come up with the final result.
The tool I've used is Estimacs, developed by Dr. Howard Rubin in the eighties and now available from CA. I don't know how well CA has maintained this product but it was excellent when I used it a number of years ago. Estimacs is a great combination of function point anaylsis and practical considerations. It has a database of thousands of projects and the effort it took to complete them within it. Dr. Rubin is a world class expert on many things including software project estimation. There is an article on Darwin (http://www2.darwinmag.com/connect/ask/expertbio.cfm?ID=101) that you might find interesting. Using Estimacs and the other two methods, then rationalizing all three by give and take discussions between the three teams will result in an excellent estimate. This sounds like a project all by itself -- and it is. Just don't under estimate its complexity or the effort required to do it well! If you can help your organization understand what should be done to estimate properly you will have gone a long way towrd the cultural change required to get good estimating going in your company. Saving Changes...
Frank WintersPhotographer and ConservationistSandwich, Ma, United States
I found a great article on the Rational website. I'll summarize. Measure how far a estimate is in terms of the square root of the duration. (you need historical data). For example, I say I can do x in 9 days. If I deliver in 12 I'm off by one square root (3)not bad. If I deliver in 15 days well I really missed the boat didn't I? Try it. In the meantime I'll find the original article and post it here. Just a good rule of thumb. Not a silver bullet Saving Changes...
