Continuous Improvement – From Theory To Reality
I attended a conference last week that covered the topic of “Continuous Improvement”. The speaker in question recommended project managers to conduct post-mortems after every assignment to identify process changes that could be adopted in order to improve future projects.
Dozens of books have already been written on continuous improvement, and I can't add much to the subject – so I won't. But I do want to clarify why continuous improvement isn't the silver bullet most people think it is.
Welcome To Reality
Reality is no two projects are alike. And while you might, in retrospect, identify things that you could have done better in your previous project, it doesn't necessarily mean that applying your findings to your next project will automatically result in faster, cheaper, better software.
Let's assume you're a Project Manager in charge of developing ezines (electronic magazines delivered via a Web site or an email newsletter). You've just completed managing the development of The Project Mangler (www.projectmangler.com) and decide to have a post-mortem to determine how you could improve your process for your next project: Gantthead (www.gantthead.com).
Your Project Mangler post-mortem conclusions are clear. Ezines should be developed according to the Waterfall approach:
- Start with a prototype
- Develop a static HTML template based on the approved prototype
- Wrap the template around the dozens of existing articles
- Develop an index page that links to the articles
- Test your pages for broken links, missing images, and email addresses
Excellent. You now have a proven strategy and a step-by-step action plan that you can apply to the development of Gantthead. Or do you?
Gantthead is different than The Project Mangler. Now only does it contain >20 times more articles than The Project Mangler, it also (a) requires users to authenticate to the system, (b) has various memberships levels (some free, others at a premium), and (c) runs on a CMS (Content Management System). Do you sincerely believe that the lessons learned in the Project Mangler post-mortem apply to the Gantthead project?
If you've ever tried to customize a CMS system and integrate it to a user authentication system and a third-party e-commerce solution, I think you already know the answer. While the Waterfall model was appropriate for developing the Project Mangler ezine, you'd probably want to adopt an incremental software development process for Gantthead given its increased complexity and associated risk.
The Point Is…
The point is, in theory, continuous improvement sounds like a silver bullet. Conduct a post-mortem, learn from your past experiences, and modify your process for future projects. However, reality is, no two projects are alike. Consequently, findings from your post-mortem don't necessarily apply to your next project. They might not “improve” your process.
When Does Continuous Improvement Apply
Different projects have different needs. So how do you determine if your post-mortem findings apply to your next project? By examining each project individually, and asking the following questions:
- How well do I understand the requirements at the beginning of the project?
- How likely is it that those requirements are going to change?
- How well do I understand the technology?
- How likely is it that the system architecture will need to change midway through the project?
- Have the software quality attributes, also known as "ilities", been properly defined?
- Which software quality attributes are more or less important?
- Is the project constrained by time, functionality, or budget?
- How reliable does my project schedule need to be?
- How much business and/or technical risk does the project entail?
- Do I need to be able to make midcourse corrections?
Before you start applying results from one project's post-mortem to another project, make sure the two projects have similarities.
Do I believe continuous improvement is an effective method to systematically improve software development? Yes. Do I recommend project managers to conduct post-mortems? Absolutely. Do I think lessons learned from post-mortems should be applied to your next project? It depends.
This article was originally published on www.gantthead.com.