Running UAT and system testing in parallel

Conducting user acceptance testing and system testing at the same time may or may not be a good choice. Expert Baher Malek discusses the factors that testers need to consider when attempting a parallel UAT process.

Are there any adverse effects to running system testing and user acceptance testing (UAT) phases virtually in parallel? And if so, what would you see as the major issues and/or long term effects this may cause?

This is largely a question of acceptance tester availability and expectations. Most business acceptance testers are only allocated part-time to testing due to other job responsibilities.

It's often in the interest of the project team to discover their acceptance criteria sooner rather than later.

However since they typically provide sign-off for the production release, it's often in the interest of the project team to discover their acceptance criteria sooner rather than later. Although this sign-off responsibility may seem to imply a high level of business knowledge and testing credibility, that is not always the case. This in turn adds the basic question of quality management. Let's consider each of these factors individually and how they relate to one another:

Availability -- If the business testers are performing testing on a part-time basis, they typically have a higher expectation upon upstream QA system testing activities. In this case they are essentially expecting a "golden" system release to UAT. If a decision is made to run UAT in parallel to QA system testing without also adjusting expectations (in a downward direction) things could get ugly very fast.

Expectations -- Even if the business testers do happen to be available to participate in QA system testing, there is still a need to actively set expectations that things may be a bit "bumpy." There needs to be a distinction drawn between this type of collaborative testing versus the formal user acceptance test process that would follow. Furthermore, business and IT management must recognize that this collaborative testing is not UAT and that defects are to be expected. This type of testing partnership arrangement can be very effective, but can be difficult to negotiate with all the related parties.

Signoff -- The responsibility for production signoff typically rests with a business sponsor who receives input from individual user acceptance testers. In the best case scenario the user acceptance testers are business experts who are have high testing credibility. In the worst case the user acceptance testers are a necessary step in getting to production. In either case you have to get their sign-off, which means early testing collaboration is bound to yield a smoother UAT and will likely result in some useful business-IT cross training.

The ideal situation is one of testing collaboration and partnership that serves to front-load defect discovery and resolution by embedding business knowledge within system testing process. With the right partnerships and teaming strategy, a parallel UAT process can be successfully managed.

More on this topic

Dig Deeper on Software testing tools and techniques

Cloud Computing
App Architecture