Information Age: News, analysis & insight for IT & business leaders

Effective global software sourcing involves getting the price/quality ratio right, say consultants Michael Bragen of SPR and Paul Michaels of Metri

For some time now, India has been recognised as the global centre of choice for offshore software development, driven by the twin attractions of a highly educated population and low labour costs. This model, however, is beginning to change, driven by several factors.

Despite its reputation for innovative technology, India struggles to find a sufficient pool of talent to fill the tens of thousands of software engineering and IT services jobs being created.

In addition, wages in the country have been rising rapidly while Western demand has been cooling in response to gathering economic headwinds. All this indicates that India’s rate of growth for IT services in the next five years is likely to slow in relation to other emerging economies, which means more sourcing options are entering the frame.

In the quest for optimum low cost/best practice centre of production excellence, who are the emerging candidates? One of them is China, which has a flourishing IT outsourcing industry that is available to both domestic and foreign clients, and growing in parallel with the country’s industrial and corporate expansion.

Other contenders could include Brazil, Hungary and South Africa. And ironically, given the current weakness of the dollar, the US may itself become a low-cost/high expertise outsourcing centre for software development and other IT services to today’s emerging economies.

Wherever in the world they are sourced, purchasers are increasingly reluctant to leave their customised software development projects to chance. Many have had their fingers burned in the past, and now they want to know before they begin: a) what they are paying for; and b) what they are going to get.

For a European CIO faced with the decision of whether to buy a customised application project from India, China or anywhere else, they want to be able to apply a set of guidelines in order to make the right choice. But given the intangible nature of software, this is not straightforward to achieve.

Fortunately, the Software Engineering Institute (SEI) framework for evaluating and rating IT development organisations – Capability Maturity Model Integration (CMMI) – provides a set of international best practice standards against which to measure the main elements of an effective development process, from initial project planning to software testing and validation. Most software development suppliers, certainly in India, are CMMI certified, although the extent to which these standards are adhered to may vary.

CMMI standards and measurements can also be used at the planning stage of a project to get an accurate estimate of how well the end results will meet expectations and how fast the software can be produced.

Productivity metrics

When planning a project and selecting a supplier, two crucial things to consider are productivity per unit of paid programmer time and the quality of the finished product relative to the rate of productivity.

Unless both of these components are benchmarked against universal best practice industry guidelines, it is hard to make an apples-with-apples comparison or to judge whether a particular outsourcing centre represents fair

value or has adequate skills for the job. Measuring productivity must be relative to something else. A programmer may produce 1,000 lines of code in a given timeframe, but this doesn’t necessarily mean it is accurate to requirements.

To get a true assessment, we must focus on the number of function points that can be produced per hour or per month. What is a function point? It is a standard unit for measuring the size of a software application project. Once this function point size has been established, the number may be used to derive the anticipated effort required and the quality of the project through all the defined phases of the production life cycle.

Focusing on these function points takes into account the most challenging part of a project, which isn’t the coding process itself (in which some international development centres excel); it’s about the way an application is built, designed and tested, and how well this reflects the actual business needs of the customer.

In fact, the programming component accounts for about 10% of the job, while the other 90% should be spent in mapping the project’s life cycle to allow for all of these stages, allowing for adequate testing cycles to ensure a bug-free delivery.

The degree to which this is achieved is a measure of its quality. Therefore, benchmarking productivity is about far more than comparing the technical aspects of code output – it is also about measuring what is delivered in terms of its functionality.

Further reading

Offshore 2.0
Organisations are now looking to their sourcing partners for technology and business process innovation