Consider a rather large project that like so many projects had some difficulties. The project team had a major component (subsystem) delivered from a supplier. The supplier has one set of processes and the customer organization another.
This supplier delivers multiple versions of this major subsystem. The customer integrates this subsystem into the larger system and performs integration testing on the product. In the course of this integration work, the customer finds numerous problems the supplier did not find before sending the product to the customer. These faults are then reported into the company sanctioned reporting system and assigned to the party responsible for the corrections. After a review we find that the subsystem supplier did not find these failures prior to sending to the customer for a few reasons:
- The test group at the subsystem supplier did not have available the top-level product into which this sub-assembly went (lack of test objects)
- The supplier did not have system test cases that mapped to the functions of the entire system
- The supplier did not have test personnel or automation in place to the level the customer organization
Throughout the project we have seen large numbers of faults reported. We are late in the project and we continue to see a large volume of fault reports with a significant number of those failures having a high level of severity. Given the fast rate of defect arrival and the slow rate of defect closure, it is very difficult to predict success for the project by the required launch date. A launch date, by the way, that is very rigid due to a regulatory requirement.
The supplier sends over a vice president to discuss the defect reporting and closure method with the customer staff. Specifically, the vice president wants to be able to close the failures reported quickly. The customer organization process requires that the person that found and reported the problem perform the test case again on the material that is similar, or better still, the same system in which the failure was originally reported. These stipulations attempt to ensure the variables that influence the failure are all accounted. The vice president wants to be able to close the fault reports presented by the customer, he does not want to perform the specific test case nor be required to test on a similar system as that of the originating test failure. Moreover, he wants to be able to close the fault reports when the developers believe they have addressed the problem (no test evidence required). To that end, the developers make the changes to address the system then artificially manipulate the system as verification of the correction.
The vice president secures a conference room and invites the responsible test personnel and managers from the customer organization. He then proceeds to tell them he wants to close the fault reports based upon:
- Engineering revision level of software (not a released software)
- Not conducting the specific test case (tested by those who did not find the problem the first time)
- Not using similar test subject
There are twenty or more test people from the customer organization in this conference room. He lays out his case for how he wants to handle the fault closures. The test experts explain to him the reason the fault should be closed via the customer process. It is also mentioned that the supplier did not find the problems in the first place. They do not have documented test cases, and neither do they have the test subjects. In the discussion, the vice president continues his demand that the supplier should be allowed to close the fault reports as supplier wishes, including confirming corrective action on development software in spite of the limitations. The latter amounts to closing the fault report based upon the hope we really corrected the problem.
Ultimately the meeting terminates and not one customer test person agrees to his way of handling the testing results. The vice president then goes to the management hierarchy where the decision is made to handle the fault reports as the supplier wishes, all while using the word “cooperation.” In the end, there are many fault reports that are opened, closed, open, closed, and opened again. The test group begins to be analogous to Sisyphus, pushing a boulder up the steep hill to have it fall back at the end of the day only to start the process all over again the next day.
The customer’s verification staff had valid and sound reasons for not wanting to deviate from their processes. The evidence to support this assertion is the fact that the customer’s processes found the problems while the supplier’s did not. What message do we send to our employees about the processes when we coerce them to deviate from what appears to be a valid way of working? After all, the verification group did find many major issues and save the company from certain disaster. We harp on the fact we want our teams to follow some rationale process then when they do, we coerce them to deviate from what they believe is simultaneously company process and demonstrably better way than the supplier’s method.
We have seen once-proud firms descend into this pit—leading to declines in market penetration and substantial cost in returned material and customer dissatisfaction. We know of a major cell phone vendor that has gone through this nightmare. In the case we are discussing, the result was quality cost in the millions of dollars.
We provide the definition of the words from dictionary.com below for your consideration:
Cooperation an act or instance of working together for a common purpose or benefit.
Coercion force or the power to use force for gaining compliance
Hiding the problem is never the way to resolution. The processes are in place to help the organization especially during times of stress. If the processes are not able to meet that demand; then those processes should be reworked so they will deliver the results we desire. We must ask the questions:
1. Was the vice president’s domain knowledge more significant than the mass of domain specific employees?
2. What message do we send to our employees when we demand they dump the organizations process?
3. What impact did this “cooperation” have on the morale of the employees doing the work?