Last time we talked about DevOps in contact centers, we discussed the benefits and challenges of implementing the methodology in corporate cultures and tips on first steps to take. Now let’s talk about test automation progression and best practices that will give you the most ROI.
Test automation progression
When moving through your test automation pipeline, remind yourself this: Stay focused on the goals through it all.
- Unit testing: Focused on features or individual components.
- Systems integration testing: How components within a system communicate with each other.
- Regression testing: Do improvements cause legacy features to break?
- Scalability/performance testing: Did you implement any code or patches that presents a nonlinear scalability problems?
- Monitoring: Also known as post-deployment and Day 2 operations. Make sure patches or configuration don’t do more harm than good.
Anything you use in phase 1 should be reusable throughout the entire pipeline and be understood by the operations team.
Testing best practices
Perform functional and regression testing
This is especially important for IVR testing.
Test in the lab, verify in production
In this instance, size does matter. Your typical lab isn’t sized to production scale and you probably don’t have extra carrier capacity at your disposal. Therefore you must carve out specific test instances for production environments.
Test the full customer experience journey
Test everything from the carrier all the way to the individual call center or cloud-based locations.
Test for success, prepare for failure
Understand how the environment is supposed to scale and plan testing accordingly. Acknowledge that issues will be uncovered so you can have enough time to address them.
Do you have disaster recovery and failover scenario test cases? These are critical in your business interruption assessment testing (BIA Testing).
Collaboration is key
Culture change is critical. Corporate silos and “not my problem” mentalities will make this an uphill battle.
Make sure to involve subject matter experts and the full technology staff. Even desktop support should be involved in testing.
Establish performance baselines with load testing; monitor them in production
This is the only way you’ll understand how seemingly small changes can cause big issues. Without performance baselines, you won’t know whether a series of minor changes add up to a customer-impacting problem.