The Register has followed up on their interesting poll on the “off-shoring backlash” that I referred to in my earlier blog item last week, with an article of the challenges of distributed software development. You can read their full article here.
The problem areas and their impact on the different models for distributed software development will come as no surprise to anyone involved in such projects, with communication being the top problem area, followed by “software quality issues due to the variable level of skills between locations” and “political issues with the way the organisation is structured”. The Register concludes that the main message of their survey seems to be that if you distribute development, quality will suffer.
Given that this second survey includes a lot of organisations which do not off-shore some/all work, but distribute the work between different locations within their own organisations, in my mind, the survey confirms that the off-shore model – when used properly – can work as well as any equivalent model based on staff working locally. The key is to get the right model, and to ensure that the organisational roles and procedures for software development are strongly stated and adhered to.
What I find most interesting is the correlation of the perceived success against the reasons for distributed development. In this area, 65% of respondents who believed their distributed development approach was primarily motivated by cost savings reported that their development was out of control. Those focussing on resource optimisation and/or management strategy responded best, with only 12% out of control and around 50% good or great.
In my previous blog item I noted that:
Another type of poor customer is one who has had a poor methodology in his IT operations to date (poor/non-existent documentation, poor statements of requirements, non-existent specifications, etc ...), and expects to continue in the same way when working off-shore – a 5,000 mile gap between customer and supplier will only accentuate the problems of this approach.
The Register’s article supports this – it highlights that organisations that take an ad-hoc or mixed approach to distributed development have the highest dissatisfaction with their level of control – whilst those having formalised hub-and-spoke or peer-to-peer approaches have better control.
I can but repeat that this is no doubt that off-shoring can and does work - but only if it is done properly – with the right management expectations, relevant splits of roles and responsibilities, firm methodologies and strong communication.