Successfully Delivering a Software Project

You have decided to replace a key system, you have defined what you require and have selected the solution best able to meet those requirements so how do you deliver that project successfully to time and budget?

The trouble with technology, or process improvement, projects is that, even when buying packaged software, no two projects ever seem to be the same and there will be requirements that are unique to your organisation, particularly if there is some level of customisation   

However there are some sensible measures that can be taken to ensure project success: 

1. Have a plan. It sounds obvious but it is amazing how often projects don’t and if you don’t know where you are going you will end up somewhere else.  Your plan needs to be understandable and manageable so as with requirements gathering ‘big is not always beautiful’. The plan should be appropriate to the size of the project but make sure you have a plan and that it is understood. A 1000 task project plan is not something to be overly proud of as most people will not understand it and so something will go wrong.
2. Manage the Plan. The plan is not just a deliverable to be filed. It has to be managed as it will not look after itself.  Things will happen that may cause you to review the plan and possibly to adjust it. This doesn’t mean micro-management but it does mean regular project meetings both for the project team and the senior stakeholders and there needs to be a log of real risks and issues, i.e. ones that really will impact the project not hypothetical ones that fill up the log.
3. Manage Change: The plan and the project are not set in concrete. It is a fact of project life that requirements change. This can be for a whole bunch of reasons, the original requirement was misunderstood or a new requirement as a result of a strategic change in the organisation.Whatever the reason you need a process to manage change, so that those involved in the project understand how much it will cost, why it is necessary and the impact on the overall project. Then the decision can be made as to whether the change should be approved and when it should be implemented.
4. Be realistic with the Go live date. One of the reasons why software projects fail is because of unrealistic timescales. There are a number of reasons for this. Firstly there is the perception of it being packaged software, but with all packaged solutions there will still be configuration and testing as well as some elements such as interfaces and data migration that will be unique to that project. If there is an operational or legislative reason why the solution needs to be implemented by a certain date then that needs to have been incorporated from the beginning of the project. There is the tendency to be too optimistic but issuing deadline after deadline will only lead to distrust and aggravation on the part of the client and it takes a strong project manager to propose a realistic plan.
5. Manage expectations. Know who your stakeholders are and ensure that they are involved and informed. They all want the project to be a success but they want to know what is happening, when and how it will impact them. In particular nobody likes surprises. Ensure that there is regular communication appropriate to the stakeholder and their level of involvement.