What does uniformity look like?QA is about more than testing to catch bugs before release. It also involves putting practices in place that prevent trouble from disrupting development in the first place. Uniformity plays a part by setting and communicating standards, so everyone knows what is expected and what success looks like. Although there’s a broad spectrum of practices that contribute to uniformity, they can be grouped under three major components: process, communication, and testing.
ProcessAgile development is a responsive, flexible practice, but that doesn’t mean there’s no place for uniformity. Think of it as a similarity of process, not specific actions that must be taken every time. Some examples:
- Following consistent standards during discovery ensures the requirements gathered are as complete as possible before development starts.
- When sprints have established guidelines to include contingency plans, the team can respond to change orders quickly and efficiently.
- Uniform processes make it easier to manage staff fluctuations. If a team member leaves the project, their replacement can get up to speed with a minimum of disruption.
CommunicationClear, regular lines of communication keep projects on track and within budget. Without uniform communication, small problems can brew into huge misunderstandings and new requirements can be lost. At a minimum the communication plan should include:
- Who has the authority to approve change orders
- How problems within the team are resolved
- What the process is for handling issues that affect the sprint schedule
TestingTests are most relevant when their results are measured against those from identical tests. This might not seem like a huge issue since the advent of automated testing, and it’s true that automation has made QA a lot easier. However, automation only covers about 80% of the testing process. Aspects like usability still need to be tested manually. It also matters which tests are run and when. To build a uniform testing process:
- Be sure engineers know what tests are expected and what the software is expected to do.
- Have a minimum testing schedule so tests are run at the same points in the process.
- Make sure manual testers follow a consistent set of procedures so feedback can be compared “apples to apples” instead of “apples to oranges”.
What happens when uniformity fails?The damage from irregular processes varies. In the majority of cases it simply makes the development process more expensive and complicated than it needs to be. Bugs are found later than with uniform testing procedures, resulting in costly fixes when a smaller one would have worked earlier. Change orders are missed and must be awkwardly accommodated where they don’t fit. Stakeholders who were overlooked during discovery add vital features towards the end of the project. These kinds of changes won’t derail the project entirely, but they do threaten the budget and timeline. Sometimes, though, neglecting uniformity causes serious quality issues. The software makes it to release with major flaws that damage the company’s reputation (or worse, cause a data breach). The weak documentation makes it hard to fix mistakes. Added features eat into the project’s profitability. Worst of all, missing requirements can lead to a final product that doesn’t meet the client’s needs. The software is what they asked for, but not what they actually needed. Features were suggested that clash with the customer’s workflows. The scope is misunderstood entirely and the software can’t handle the necessary workload. These aren’t easy problems to fix at the end of development, if it can be done at all, and they could have been avoided by emphasizing uniformity during discovery.
Does uniformity fit into Agile processes?On the surface uniformity seems out of place with Agile methodologies, but in practice it supports them. For example, in Agile development everyone is responsible for quality. There’s a need for clear standards everyone can reference and understand. Plus, iterative development geared towards continuous delivery naturally requires uniform testing to catch errors before they snowball. This includes both testing new features and checking for cumulative functionality. Uniformity doesn’t mean things can’t change during development. Instead, it creates a process for how things change and lays out contingency plans so that when change comes everyone still knows what’s going on. In a very real sense, uniformity isn’t a cage- it’s a foundation.
Did your last software project run over budget? Schedule a complimentary meeting with one of our developers to explore how uniformity helps keep development on track!