We're trying to estimate our likely testing hours for a project based on a known number of requirements hours. Has anyone seen "industry-standard" ratios that would helpful for that type of estimating?
For the specific question I need to answer right away, I'd like to utilize the requirements hours consumed by *business* resources to estimate *business* resource time in testing. Saving Changes...
Estimating testing time as far as accuracy goes can be something of an art or even just magic! There are so many variables that can derail the estimates, but the key one of course is quality. The poorer the quality the greater the number of testing cycles required and this is not something that is known when you are estimating. I try and use at least 3 full testing cycles and where the system is complex sometimes even more.
Sometimes there is not a direct correlation between requirements and testing effort, or not a 1 to 1 correlation. What I mean here is that one test can cover multiple requirements and visa versa. Saving Changes...
Jayson ReadProject ManagerEden Prairie, Mn, United States
What we've typically done (not to say it's industry standard) which has helped us from an initial estimation perspective is to estimate the construction and then use a factor of the construction effort to estimate the testing effort.
We've used between 25% and 35% depending on the expected complexity of the modification. The greater the complexity, the greater the amount of testing required. Saving Changes...
