When Motorola created and implemented the Six Sigma quality assurance methodology, from 1986 to 2001, they reported saving $16 billion, a 4.5% savings, in revenue. This was attributed to Six Sigma’s ability to cut out waste and optimize efficiency within the company. The figures are also impressive because many methodologies promise results, while some only provide more red tape. Like in manufacturing, the need for a formalized methodology in performance engineering remains the same.
In performance engineering, we have our own testing methodology and best practices. A performance testing methodology can be a hybrid of the best parts of other methodologies such as CMMI (Capability Maturity Model Integration) and Six Sigma. Before testing begins, start by understanding requirements. Afterwards, you can turn those requirements into deliverables, and those deliverables lead to standardized load tests that you can perform. After those load tests are performed, you analyze the data to determine your Key Performance Indicators (KPI). From here you can quantify your results as binary (success or failure) or as a scorecard (1 – 10). From there as goes in layman’s terms: “Lather, rinse, repeat.” Performance testing is an iterative process.
The importance of methodology comes from the modern workflow. As mentioned in the introduction previously, process methodology is aimed at reducing waste, reducing cycle time, and improving overall performance. In performance testing, it’s important not to waste time in the long term to save a few minutes. Hastily performing a load test that will turn out to have defective results and can waste hours of company time. Allocate some time toward planning ahead appropriately. Measure twice, cut once.
