Information Technology Reference
In-Depth Information
be reduced by more than 70%. Customer also believed that the automation
tool will generate testing data and so time required to generate test data either
manually or using any other tool will be totally eliminated. Customer also
had a wrong notion that the automation tool can run in any environment
irrespective of platform, hardware, or operating system. Customer also had
expectation that the automation tool will generate scripts that will detect all
defects in the software application.
Customer attitude—Customer was not ready to listen to suggestions.
◾
Test automation process—Customer was totally ignorant about the whole
◾
automation process including detailed procedures required.
Test script maintenance effort—Customer was totally ignorant about the
◾
maintenance required in scripts.
9. 3.1.4
Solution
Customer expectation—Educate the customer about practical constraints,
◾
detailed test automation process, the real benefits of automation, which auto-
mation tools provide what kind of benefits, which tools are the best fit for the
given scenario, what are limitations of the tools, provide detailed automation
framework, provide ROI, provide technical expertise and latest trends in test-
ing tools, explain what metrics will be delivered to customer based on type of
data collected, what measures will be controlled and monitored.
9. 3.1. 5
Lessons Learned
It was a tough call to convince the customer about realities. Through con-
tinuous customer education it was possible for the team to explain what the
real benefits are and exactly how these benefits can be realized. The team also
educated the customer on the whole process to be followed in automation. The
team kept the customer informed on all developments and updated him on
regular intervals.
9.3.2
Business Case 2
9. 3. 2.1
Automation Strategy
To be successful every project should have a well-thought-out strategy before the
project actually starts. In many test automation projects, ad hoc measures are taken
instead of a strategy from the start. For instance, it is important that you select
which test cases should be automated and which should not. Depending on com-
plexity, the number of times the test case should be executed during regression,
and maintenance needed over the number of expected regression cycles, it is very
