Following from People and Process Parts 1, 2 and 3, in this post we’ll focus on the of key issue of TESTING. Poor processes when planning, executing and reporting on testing, let alone when expediting fixes, will also hit your bottom line by hampering your ability to move faster and execute cheaper and better.
As with earlier related people and process issues the value of comprehensive and trusted functional unit, functional integration, functional business/user acceptance, regression and performance testing cannot be understated.
When under project pressure the temptation to undertake a risk-based or sample-based approach to functional and regression testing is ever present. That is, to sacrifice comprehensive testing in favour of time and cost and, almost certainly, at the expense of quality.
In the context of customer communications where templates can produce multiple scenario-based, multi-recipient, brand-based and multi-channel outputs, that also may carry persona-based promotional messages, why run the risk of shortcutting?
After all, your communications are your most critical means of contact with your customers. Any mistake in data, content or context could cost you your customers.
We observe that testing processes, tools, outputs and reports typically sit outside core CCM solution capability – meaning that they must often be built, integrated and utilised from scratch.
The result of this discontinuity, from specification to build to testing, is:
- Test plans, test cases and test data are manually created and executed based on business requirements and technical specifications. Manual creation is time consuming and introduces personal interpretation.
- Requirements often do not cover all variations, i.e. all output scenarios. In most situations test cases and testing overall is limited / bound to information contained in, and data inferred from, the requirements document.
- Limited regression testing often due to poor data or poor understanding of data.
- Non, or inconsistent, use of testing automation tools – most tests are manually executed, followed by manual results reporting, which is time consuming.
- If tools such as Jira are in the mix, defect resolution can be delayed or confused because of the lack of integration with testing processes and tools.
Compounding the above we observe impacts related to review and approval of test outputs and results:
- Samples presented for approval are insufficient, have unclear context or are not aligned to the business’ expectations about variant outputs, leading to re-work.
- Business/User Acceptance testing planners and test case writers
do not always know about and are able to test all scenarios, resulting in
potential errors after deployment for specific scenario unknown at the time.
- Comprehensive and trusted unit testing is key to informing B/UA testing planners
- Review and approval process for business users is manual and slow if via email, even if referencing PDF outputs generated and manually attached or linked to Jira issues or added to shared drives, especially if access is restricted.
- Poor change tracking and auditability – review feedback communicated via emails may or may not be transposed into Jira issues.
- Test summary reports, often required by deployment gatekeepers, are manually produced.
Mitigate the risks of poor testing processes: improve turnaround, avoid rework and reduce your testing effort with structured, comprehensive, fully integrated and trusted testing using virsaicTM ASSURE.