Road to the Perfect Project
Ever since software development projects have been around, people have been coming up with ways to help ensure they come out well. Unfortunately, history shows us that there is no process, methodology, or toolset that can guarantee project success. But there are some practical techniques, many of them non-technical, that can dramatically improve a project’s chances.
In this, and in the next couple of postings of Business Analyst Times, I will lay out techniques based on observations from my own personal experience that I know can reliably produce excellent results. In combination, they can result in what is experienced as a “perfect project”. I present from more of a project management perspective rather than pure BA one, because my experience has been that management activities and decisions tend to impact projects, for good or ill, more than technical ones. The ideas that follow are in approximate order of importance. They are necessarily quite brief, and each one is expanded in future postings.
1. Keep the project as small as possible. A project’s size is inversely proportional to its chance of success. As people are added to a project, the potential number of interactions increases pretty much exponentially, and things fall through the cracks, often catastrophically.
2. Carefully select a leadership team comprised of a project manager and key business and systems analysts. They must be highly qualified technically and must work together well as a team.
3. Create a partnership relationship with the customer. To improve the project’s chances of going smoothly, these two organizations need to work in concert, where mutual trust, joint responsibility, transparency, and good will are operative.
4. Limit formal processes and documentation to a minimum. “Big Process”, as exemplified by CMM/CMM and the IEEE standards can get in the way of a small sized project with high quality, motivated people.
5. Work intuitively. Operate at least partially in the mode exemplified by the Agile Methodology. By being highly flexible and willing to change as users are exposed to the emerging system, the probability of system acceptance is greatly increased.
Anyone who has been around a while knows that the conditions for achieving the perfect project are not commonly present. But, if they do exist where you now are, or you can influence things to cause them to exist, then enjoy your perfect project.
Keep Project Size Small
My experience has been that project size, in terms of the number of people working on that project, is inversely proportional to its chance of success. I have been around several projects that had 100+ staff. While none of them outright failed, none of them came out particularly well. In one instance, a literal handful of people (myself included) redid one of those large systems in a year of intense work, and turned a mediocre system with poor performance characteristics into a roaring success.
Why do large projects have problems? In a nutshell, the active principle is project communications. As people are added to a project, the potential number of interactions increases pretty much exponentially. The result is that important communications don’t always get to all necessary recipients, and things fall through the cracks. Mistakes are made, and they can prove to be disastrous. People feel isolated, morale tends to be poor, and turnover is high. Big projects try to mitigate this by requiring huge amounts of formal documentation and by holding many, many meetings. The sheer mass of material makes it nearly impossible for individuals to read everything they need to read to stay up to speed on everything that might affect them. It also becomes almost impossible for any one individual to comprehend the full documentation set as a whole, meaning that there may be no one individual who understands the entire project.
Any project that has more people than can fit around a large conference table is going to experience degraded performance due to communications issues. Conversely, with a small team, any communications issues can be addressed face-to-face and be observed by the entire project team. Much can be done on an informal basis that would otherwise require formal process and documentation. For instance, the two persons responsible for both ends of an interface can sit down at a whiteboard and work out the interface, and then, instead of creating a formal Interface Control Document, they just send an email summarizing the design. When it comes time to test the interface, both parties can be present and can troubleshoot and fix any problems on the spot. I know this works, because I have done it.
A small, closely knit team of sharp people can work miracles. I have been lucky enough to have worked a few myself. I claim that with such a team, working informally and without undue process and documentation overhead, I can build a given system faster, with better quality, and with better user acceptance than can be done by your average 100-person team, working under what is considered best practice processes.
Some projects are just too big functionally to do them with a single small team. This is especially true if hardware and communications must be procured and integrated, facilities must be built, and national security data is involved. But it may be possible to cleanly partition the problem into a set of functions that can then be implemented using a set of closely cooperating small teams, with a top-level team comprised primarily of team leads. Each team then operates independently on its functional area, and the top-level team ensures that cross team issues such as interfaces get handled efficiently and effectively.
© Copyright John L. Dean, 2008.
Look for Part 2 of Road to the Perfect Project in the next Business Analyst Times
John L. Dean is an IT professional whose long and broad experience base gives him an unusual perspective on what does and does not work. He has 40 years experience, mostly in government IT contracting, including systems engineering, systems analysis, requirements analysis, systems architecture and design, programming, quality assurance, testing, IV&V, and system acquisition, equally split between technical and management. He has worked on both the contractor and government sides of the fence, has been an employee of large (IBM, CSC, Booz-Allen, MITRE, ACS) as well as very small companies, and more recently has worked as an independent consultant. John has a Masters degree in Operations Research from Clemson University. He can be contacted at