Every edge-case, every test? The automated UAT answer

Judgments about the reliability of ERP software in production are based on continuity of service. Bugs and issues manifest as problems that impact internal end-users, partners, or customers and vary from small irritations to full-scale systems being down for protracted periods. A great deal of resource is rightly spent, therefore, on testing, QA, UAT, and change management processes.

There is, however, friction between the need to maintain continuity and the demand for new features, improvements to user experiences, extended capability of applications, and so on. Cloud-based point products from third parties can and do iterate quickly, and this influences stakeholders’ attitudes. Questions around why platforms can’t be so agile as off-the-shelf cloud apps can’t be met with negativity, so it’s incumbent on IT to respond to the demands of the business.

It’s a dependency thing

The nature of a business’s growth is organic, so software and supporting infrastructure have grown and adapted to changing demands. Burning down and starting afresh simply isn’t an option, so it’s to better, iterative software development that IT decision-makers must look (where “better” probably means faster and more reliable). But the development of software – the physical forming of code – is not necessarily the bottleneck. More often, the biggest time and resource sink is the testing and retesting of even a simple change’s effects on the rest of the software portfolio.

Applications and services’ performance and reliability depend on multiple components and rely on other applications and back-end systems. A so-called non-breaking change considered relatively trivial can alter a user’s experience with effects ranging from going unnoticed (good) to rendering an application unusable (very bad). Any development team that’s worked with a cloud provider knows that a small point upgrade might be catastrophic when the update comes downstream. And, as many ERPs are applying pressure on their customers to move to cloud-based solutions, more upstream changes could easily spell more upset.

Source: Shutterstock

Testing cycles

User acceptance testing ramps up the cost and time required to get new systems or features out into production, but it’s a 100% necessary evil, even for upgrades thought to be insignificant. Removing busy professionals from their daily tasks to test scenarios is a resource killer for the business, so not every test can be comprehensive. As the software stack increases in size, not every regression test can happen with “live test subjects” – there has to be some prioritization.

In most organizations, therefore, there’s a degree of user acceptance testing that’s automated, so most eventualities and even some edge cases can be tested.

Automated application testing has become its own subtle art, with reams of code formed over years of change management tests. Those testing libraries tend to be highly tuned to specific problem areas and are clearly going to comprise bespoke routines. That makes these repositories expensive to keep up to date. And sometimes, significant developer resources are dedicated not to front-line production-focused CI/CD, but to developing testing automation.

Smart Automation

That was the situation, at least, until Original Software’s UAT platform reached the market. It’s a code-free testing tool that examines every change in the context of every affected application across the interconnected web of a company’s ERP-based applications.

Source: Shutterstock

As iterations on applications are produced, the Original Software platform’s testing library increases in scope and size, creating a baseline from which tests are conducted to shorten the UAT cycle. Most users put the time & resource saving around the 60% mark, and in an environment where pre-production QA and testing take the lion’s share of time, applications get to production faster.

The reason why the UAT automation capabilities of Original Software’s platform are so significant is that the majority of UAT is, in fact, regression testing. Automated systems don’t have to prioritize where developers think processes might be most affected. Instead, every eventuality is covered and checked at every release point.

In our next editorial, we’ll look at how collaborative technologies can bring in stakeholders from across the business to help with UAT – at a much lower resource cost to the business. There’s an effective way that developers and end-users can work together to bring about better results in addition to those huge savings.

Whether your company’s applications run on Oracle, Infor, SAP, Salesforce, IBMi, Workday, or any other platform, you can cut time-to-production significantly across multiple applications and improve the chances of breaking changes being fixed before applications see the light of day. Find out how to begin with Original Software’s UAT solution.