 |
The Mythical 4-Month Project
June 2007
When evaluating the quality of your software development process, the predictability of project planning and estimating deserves a careful look. When many projects are over time and budget, it's time to look at the estimating processes. So let's understand where estimating typically goes awry.
The Mythical 4-Month Project. Let's consider a typical four-month software development project. At the outset, the activities and budget would be as follows:
| Planned |
| FTE |
Phase |
% of Project |
Budget |
Duration (weeks) |
| 4 |
Requirements |
35% |
$44,800 |
5.6 |
| 4 |
Coding |
35% |
$89,600 |
5.6 |
| 2 |
Testing |
20% |
$25,600 |
3.2 |
| 1 |
Project Mgmt. |
10% |
$6,400 |
1.6 |
| |
Totals |
100% |
$166,400 |
16 |
Note: Average rate $100/hr
Yet every study done on software development shows that the final cost, whether measured in hours or dollars, is expected to be 40-60% higher than initial estimates. To the project manager and the stakeholders, most of this additional time appears to be spent during the testing phase of the effort. However, the actual work effort is expended not in testing, but in additional analysis and design to clear up ambiguities and mistakes in requirements, business rules, and design. So the 4-month project now becomes a 5-6 month endeavor. More importantly, the distribution of work effort looks like this:
Table 1
| Planned |
Actual |
| FTE |
Phase |
% of Project |
Budget |
Duration (weeks) |
Rework
Effort |
Final
Cost |
Final
Duration |
| 4 |
Requirements |
35% |
$44,800 |
5.6 |
$22,400 |
$67,200 |
8.4 |
| 4 |
Coding |
35% |
$89,600 |
5.6 |
$44,800 |
$134,400 |
8.4 |
| 2 |
Testing |
20% |
$25,600 |
3.2 |
$12,800 |
$38,400 |
4.8 |
| 1 |
Project Mgmt. |
10% |
$6,400 |
1.6 |
$3,200 |
$9,600 |
2.4 |
| |
Totals |
100% |
$166,400 |
16 |
$83,200 |
$249,600 |
24.0 |
Note: Average rate $100/hr
As this table clearly shows, our mythical project just got 50% bigger as a result of the need for rework. In general, requirements rework drives the costs in all phases of the work effort. Getting the requirements details correct ahead of time is a major leverage point for controlling project cost and duration. As we have suggested in previous Critical Knowledge installments, tools and processes which lower requirements-driven rework are the best investment an IT shop can make. Our experience is that with such solutions, requirements rework can be reduced to 10% of initial requirements cost, and the overall rework costs should not exceed 20-25% of the "mythical" budget for any project.
Experience the difference of getting it right from the start. Visit critical-logic.com or email moreinfo@critical-logic.com for more information.
|
 |
|
Check Out Our Past Issues
April 2007
AGILE, INCREMENTAL, WATERFALL: Is There A Difference?
March 2007
Automating Test Execution: What You Need To Know
February 2007
REWORK: Stopping Software Development Overruns
January 2007
Creating An Effective Software Testing Process
Upcoming Events
We're coming to a race near you! Come check us out at one of our upcoming racing events.
July 6-8
Critical Logic Racing "Regates de Valleyfield"
(Valleyfield, Quebec)
July 13-15
Critical Logic Racing "Silver Cup"
(Detroit, MI)
July 27-29
Critical Logic Racing
"Tri Cities Water Follies"
(Tri-Cities, WA)
August 3-5
Critical Logic Racing "Unlimited Lights Championship presented by Graham Trucking"
(Seattle, WA)
|
|
|
