Jumping on the metrics bandwagon
“If you cannot measure it, you cannot control it. If you cannot control it, you cannot manage it. If you cannot manage it, you cannot improve it.” - H.J. Harrington
In today’s competitive world, the yardstick of success is not just ousting peer organizations or products, but also outperforming your own self. Metrics are an integral part of product and process improvement. However, identifying a suitable set of metrics for your organization needs thorough consideration, as you might land up having far more on your plate than you can chew.
1. Do not seek information of which you can’t make use of - Anna C. Brackett
The “Practical Software and Systems Measurement Guide” and the “SE Leading Indicators Guide” offer a comprehensive set of metrics. It might be a good idea to use the “Goal-Question-Metrics” approach to choose the ones most suited to your business objectives.
First identify your immediate goals, which usually stem from current concerns or problems affecting the organization. “Releasing the software to the customer” might be one of your current concerns for example. The next step is to ask questions that focus on achieving these goals like here: “How reliable is my software?”
Metrics are nothing but a set of data associated with each of these questions, in order to answer them in a measurable way. The metrics associated with the reliability of our software-to-be-released would be: the testing effort in terms of man hours, number of failures etc.
2. All the data that you collect is only as useful as you present it – Anonymous
Once you identify the metrics you need, the associated data such as how, when and by whom measurements are carried out, stored and accessed must be clearly defined. Once collected, this data needs to be analysed to understand and extract the information lying within. Visual representations like charts and graphs facilitate analysis. Analysis and interpretation of measurement data -often termed as Indicators – depends on the context of the process and product. For example, fewer failures detected during testing might either indicate a mature Design and Implementation process or an incompetent Testing process if defects were reported after the product was released.
3. Data is a precious thing and will last longer than the systems themselves – Tim Berners-Lee
Historical data for all projects across the organization must be collected and maintained in a suitable manner, so that it is available for identifying trends. This factual data is worth more than a thousand guesses while planning projects or process improvements. For example, lessons learned from previous projects contribute towards better planning for future projects, which in turn improves an organization’s processes in itself.
Did all of that sound intimidating? Not to worry, having helped you define and manage your processes, “Stages” helps measure and improve your processes with the following inbuilt metrics features:
- A compact set of process and product metrics recommended in the “Systems Engineering Leading Indicators Guide”
- Easy-to-use interfaces with your organizational processes for collecting measurement data
- Visually appealing and easy-to-interpret integration with BIRT for reports and graphs