ERP Testing in a Nutshell
Remarkably, there has been very little research into the testing of packaged ERP suites. Even now, there are few books on the subject and those that do exist are focused almost exclusively on using tools bundled with the packaged software. Given the scale of this aspect of the business, it is very poorly supported.
The software vendors have their own methodologies for testing, but these focus on process flows through configured components. In effect, such tests prove the software ‘does what it does’. The behaviour of ERP suites are typically controlled by configuration parameters and do not involve software maintenance at all.
However, where the ERP system must trigger an interface and import from or export to legacy systems, ‘user-exits’, custom interface programs, or amendments to the legacy systems themselves are required.
Much of the testing has to focus on the behaviour of the system as experienced by its users, the output generated, and the integration with legacy systems. Needless to say, a lot of the functionality in integrated systems is master-data driven. Data migration is a challenging area of ERP projects and data migration efforts need evaluation, reconciliation, and testing using the integrated applications.
The ERP Testing ‘Problem’
The problem to be addressed is, ‘how to approach testing of ERP systems to demonstrate its behaviour, to meet the needs of its users and address the risks of failure’.
We suggest that an integrated set of test methods and tools is required to support the following:
Results-based management (aka Benefits Realisation), modelling of results-chains to provide input to a benefits-based test approach.
Experience-based risk assessments that can provide input to a risk-based test approach.
Modelling of integrated user, automated and manual processes and interactions in a format useful to testers.
Derivation of objective, measurable and manageable test objectives from these three models.
To integrate these three test objectives into a non-proprietary life-cycle test methodology that can be managed systematically.
Some foundation/visionary work has already been performed by the author, but the approach needs more research, rigor and proving in a commercial environment.