How can productivity and quality in software development be managed effectively and sustainably controlled? A KPI-based management model as response.
In my last article “Three levers for higher productivity in software development“, I used empirical values to show how industrial methods, such as standardization and automation, affect productivity. Thus, software development by industrial means does not aim at mass production of similar products; on the contrary, it enables a high degree of individualization by a small-scale standardization of business-related and technical components.
Blog series: Productivity of Software Development
- Individual Standard: Mass Customization in Software Development
- Increasing Productivity of Software Development – a Book Launch
- New Book: Increasing Productivity of Software Development – Part 2
- Three Levers for Higher Productivity in Software Development
- Productive Software Development Requires a Management Model
- Software Development: Every Error Is an Opportunity
- Productive Insight: Monitoring in Software Development
Selective measures are no guarantee for sustainable improvements
Experience has shown that individual improvement measures are not sufficient to increase productivity significantly and sustainably. The introduction of a modeling tool alone does not guarantee the advantages that model-based development has to offer: it must be easy for employees to use. If modeling requires both business and technical knowledge, other roles or skills may be required. The creation of the models must have a direct benefit for the subsequent process steps (preferably without media breaks).
Sometimes supposed improvement measures decrease productivity. This may only become apparent after the completion of a new development, if the corresponding product is further developed. Therefore, there is a need for measurements to reliably determine the effectiveness of the measures over a longer period of time. These measurements should at least consider quality and productivity, as it would be detrimental to improve productivity at the expense of quality and vice versa. In addition, it has been found that the better planned in advance, the more effective the improvement measures are. When planning, you should be aware of their costs and leverage. Another important aspect is how different measures influence each other.
The theoretical functioning of the model
The management model used at the PASS Consulting Group for several years is based on three key performance indicators (KPIs) and the corresponding measurement methods. The productivity of a development process is calculated by measuring the functional size of the developed software and the required personnel expenses. Because of the dependencies, an analytical analysis of productivity also requires a simultaneous measurement of quality – and thus another measurement method. Finally, not only personnel costs but also license and infrastructure costs should be taken into account. This results in the three KPIs productivity, costs and quality.
This model is based on the cyclic application of measuring methods, the analysis and evaluation of their results and derived optimization methods. In the optimization phase, the so-called Key Performance Areas (KPAs) play an important role. These are areas that have varying degrees of influence on the KPIs and thus represent areas of action for possible improvement measures. KPIs reveal whether productivity, costs and/or quality need to be improved compared to the baseline or compared with other organizational units. KPAs provide guidance on which areas of action to improve these KPIs are particularly effective. Such a consideration is usually supplemented by a calculation of the expected improvement based on the scope of a measure and possible mutual influences of several measures.
After the implementation of improvement measures, the KPIs collected and evaluated in the next cycles show whether the assumptions about effectiveness are correct or whether control measures need to be taken, e. g. a recalibration of the assumed leverage effects in the KPI/KPA model. This information is an important basis for controlling the organization and also serves to check compliance with the company’s goals.
PASS Consulting Group has been using this management model for several years. The core of the system is a self-developed method for measuring the functional scope, the data interaction point method. The application of the measurement methods, their evaluation and the optimization phase are carried out in a monthly cycle. The results are presented to the relevant development managers and strategic management at the Quality Controlling Meeting (QCM) and optimization decisions are made in this committee.
This management model and its practical application are described in detail in the second book of the book series “Increasing Productivity of Software Development”.
Picture credit: shutterstock