On this particular project, the performance of the testing team is being tracked against the number of scenarios tested compared to the plan. Using Earned Schedule analysis, the resulting values are calculated:
- Earned Schedule (ES) = August 15
- Actual Time (AT) = September 10
- Independent EAC (IEACt) = November 6
- Planned Time at Completion (TAC) = October 2
- Schedule Performance Indicator (SPIt) = 0.84
- To Complete SPI (using planned EAC) = 2.28
Here's the graph (series 4 is the actual performance):
A TCSPI, like TCPI using Earned Value, greater than 1.10 is considered to be not credible, i.e., impossible. Since our TCSPI is 2.28, recovery should be considered not possible and other alternatives should be considered.
This was presented to the team (not in a manner using a ton of ES technical babble but in a way for most others to understand), suggesting that recovery was unlikely (did not use the word impossible) and that other alternatives needed to be analyzed. The testing team, developers, and other support personnel rejected this notion wholesale, indicating that recovery is not only possible but that they will recover, and dismissing the presented analysis as "textbook." The alternative, more optimistic points of view had no basis behind it; it was just they thought was more true. Others who were listening to this debate tended to lean towards the more optimistic opinions and the general consensus was to continue on and no other mitigating or contingency alternatives were explored.
My question is: why?
Why do we spend time learning the mechanics of various way to analyze data but then reject it in practice? I have observed this chronically across many different types of projects and customers and teams.
There is not a "right" answer to my question so I know this violates the rules a bit, but I think a right answer sort of lives in the collective responses from all of you, those who live this every day. I would like to hear your critical thoughts and opinions on this subject. Thanks.