Examples of productivity measurement
On this page, two practical examples of software productivity measurement are given. The examples are based on effort hours, because productivity is expressed in hours. Multiplying effort by a relevant hour rate transposes the examples to cost.
Practical example of software productivity measurement 1
An organization requested its software supplier to realize a new application in the Java programming language. Although the contract was based on a time-and-material basis, an agreement was made that the supplier would carry out the project at least at a ‘market average’ productivity. Analyses of the ISBSG repository ‘New Developments & Enhancements’ made it clear that for this type of project a market average productivity of 11 hours per function point was applicable.
After project completion, the productivity was measured. The effort spent on the project by the supplier:
- Functional design: 2000 hours
- Technical design: 3000 hours
- Coding and Unit Test: 4000 hours
- Systems test:3000 hours
- Support Acceptance Test: 200 hours
- Implementation:200 hours
- Overhead (Project management, Quality Assurance): 1000
Total: 13.400 hours.
After project completion, the functional size was measured by a NESMA certified analyst of the supplier and reviewed by a NESMA certified analyst of the customer. The functional size measured was 1000 FP (NESMA 2.2).
The productivity realized in this project turned out to be 13400 / 1000 = 13,4 hours/FP.
The project was not carried out in an equal or better than market average productivity. Depending on the exact terms and conditions of the contract, the customer may be able to decide to pay only for the 1000 * 11 = 11.000 hours that are supposed to be a market average price.
Practical example of software productivity measurement 2
An organization decided to outsource all the new releases for a specific Microsoft .Net application from their IT department to one specific external software supplier. To get a grip on the performance of this supplier, the supplier had to implement a productivity measurement process in order to show improvement over time. The following agreements were made:
- The first 3 releases of the application are measured and converted into a baseline productivity;
- All next releases are measured and productivity has to be reported by the supplier;
- The supplier has to conform to productivity improvement targets: 10% after 1 year, 20% after two years, 25% after three years.
- The customer organization has the right to periodically review the size measurements (by an independent party) and audit the effort administration system of the supplier.
After three releases, the baseline productivity was set: 20 hours/FP.
In the next figure, the trend is shown. Please note that when it comes to productivity expressed in hours per FP, the lower the value, the better. A downward trend is therefore good!
Although the supplier did not make the agreed targets the first 2 years, productivity improved over time and after three years exceeded the target. Focus on measurement and productivity improvement resulted in faster and cheaper delivery of new functionality for the customer.