Information Technology
Article
News
Case studies
Trainer profiles

Better measurement for better quality management

Michel Benoit
Better measurement for better quality management

The implementation of quality measures is the most often forgotten or neglected element in a quality program, yet it is the most important element to "keep your program alive". How often are we aware that there is a problem in our projects, but are unable to demonstrate it clearly and accurately, using objective data? This is precisely the purpose of the metrics associated with quality measurement. And beyond simply pointing out what's wrong, quality assurance fosters a climate conducive to continuous improvement.

Michel Benoit, Technologia expert and trainer, revisits the importance of QA with an update of an article published in 2019, but still very relevant.

How can the right metrics help teams get management support?
It's a well-known fact that the earlier a non-conformance is discovered in a development process the less it costs the organization to correct it.

Therefore, it is essential to measure

  • The quality of the products;
  • The quality of the processes in place.

These measures must be explicit and easily understandable. First of all, they must arouse interest, but above all, they must encourage people to adhere to the information that is going to be communicated. These measures can, for example, demonstrate in the form of a dashboard, that have been detected :

  • 8 out of 10 non-conformities are critical in the context of a project's architecture
  • 1 non-conformity out of 2 was detected by the customer and not by the agile team
  • 45% of the effort of a software product release was spent on testing

All this has a much greater impact and contributes to your management reacting and taking the necessary corrective measures.

Example of a dashboard for the Management

Dashboard for Management

What are the advantages of quality metrics?

We can list, among others, the following benefits:

  • Identify and prioritize problems;
  • Measure the progress of your quality program;
  • Make forecasts on the quality of your products;
  • Make informed and preventive decisions;
  • Compare with past projects (history);
  • Convince or reinforce the need to implement a quality program;
  • Reduce technical debt;
  • Reduce the cost of non-quality

In general, metrics are used to discover trends in a project in an objective way. Indeed, if the metrics are collected in a relatively stable context, it becomes possible to interpret the trends as an improvement or deterioration of a quality aspect. Any change in the trend of metrics should be examined and can trigger corrective actions.

Anything can be measured, but you must make a sensible choice. It is best to start with simple, yet impactful metrics for management, for example:

  • The number of errors per deliverable,
  • Categorization of errors,
  • The severity of errors,
  • The cost of a defect on the project or product,
  • etc.

We must also measure things that are useful and easy to maintain with the least amount of human manipulation, for example

  • Customer satisfaction,
  • Costs of controls (which can include software testing, code reviews, etc.)
  • Percentage of correction,
  • etc.

Metrics should not be hidden. Make sure they are visible, especially to those who have decision-making power, so that things change.

Many quality measurement models exist

There is a whole range of predefined measures available in ISO/IEC 25010 or SQuaRE (Software QUality Requirements and Evaluation) which is an evolution of the ISO 9126 standard, IEEE 982.1 and 982.2, SEMA (Software Engineering Measurement & Analysis) and in deliverables or Internet sites under Copper Jones, Halstead, McCabe, Boehm, Albrecht. There are also many agile and LEAN metrics that measure the ability of the agile team to deliver solutions such as the progress of a Sprint or the progress of an Epic or Epic, the lead time or the cycle time.

What are the pitfalls to avoid with metrics

In practice, a simple way to determine what to measure is to simply ask your management what they want to see measured. The needs are often limited to simple, yet effective measures. In particular, the time and cost consumed in correcting non-conformances, as well as their number per software product or project.

When implementing a measurement program, one must avoid falling into the trap of wanting to measure everything (measurement madness), or using metrics for purposes other than continuous improvement. For example, using them to evaluate the performance of a resource or a team.

Continuous Improvement and Conclusion

Metrics allow you to understand the areas of weakness in your organization and to provide solutions to improve them. Take the time to measure and track the progress and effectiveness of your quality program. Every organization should have a dashboard to illustrate quality metrics and focus on them for continuous improvement to gain efficiency and quality.

While ensuring quality assurance is not always straightforward, there are tools and methods to achieve it. In particular, by defining in advance the level of quality expected and how it will be measured. Ensuring rigorous quality monitoring, whose key elements are understood by all stakeholders, allows for fewer and less costly adjustments to be made in favor of continuous improvement. By making the entire organization aware of quality, so that it is present at all stages, from needs assessment to delivery, we maximize the chances of offering users what they expect and, at the same time, contribute to achieving business objectives. Each time, the organization will gain in maturity and will be able to distinguish itself from others or stand out from the competition.

To go further :

Quality assurance: create your software quality and testing dashboards

Contact us

To learn more about our new services or to talk to us about your skills development needs, contact Cyrielle Renard at 514-380-8237 or by email: crenard@technologia.ca.

Similar articles

See all our articles