What is the cost of not testing?

The costs of not testing or of testing inadequately may include the following:

• More system outages
• More personnel for production system maintenance
• Slower software migration
• End-user dissatisfaction and potential loss of business
• Lost data
• Less new development
• Higher data processing costs
• Missed business opportunities.

The main objective of testing online systems is to find problems before you put the system into production. It is much less expensive to fix a problem found during testing than it is to have a system fail during production and then have terminal operators sitting idle.

There is no shortcut to the comprehensive testing of an online system. If your organization requires a high degree of system reliability, high availability, and good system performance, then you must dedicate significant effort and resources to testing.

The resources you need to run a test depend on what type of test you plan to conduct and how complex your system is. In general, you should treat the setting up of a test system like an application development program. You need to dedicate people to designing and developing the test system, but after it is developed you can often coordinate testing with only one or two people. The amount of time you need to code the test cases depends on how many different transaction types you have and how complex each transaction is.

However, as with any application development program, you can reduce the cost of testing by carefully planning your test in advance and by providing the test developers with a comprehensive test plan. For more information about creating a test plan for tests, refer to Creating a test plan.