Desmond File

Blog archive

Four Steps To Save a Dying Dev Project

If you've been reading Redmond Developer News lately, you've probably seen the new DevDisasters page written by Worse Than Failure publisher Alex Papadimoulis. His accounts, submitted by readers, illustrate the high price of botched development. There's no doubt that "train wreck" projects can destroy budgets, crater business plans and ultimately ruin promising careers.

Now, Forrester Research has just released a report that aims to help developers dig their way out of trouble. The research firm interviewed more than 20 application development professionals and came up with a four-step plan to help dev shops recover wounded programming projects. The steps provided in the Forrester report are:

  • Halt all work and declare a reset
  • Fix the root cause of the problem
  • Re-plan the project
  • Execute against the new project plan

This doesn't seem like rocket science, but Forrester does a nice job of digging into each of these steps to help dev managers effectively redirect project efforts. One theme I noticed: focus. In the first stage, Forrester recommends that shops "paint a target" on the root cause of the failure, while in the next stage it calls for a single person (a "fixer") to lead the effort while also increasing the clarity and depth of fuzzy project requirements. Across each stage, dev managers need to focus, streamline and componentize aspects of the project.

Ultimately, the research itself notes that the toughest challenge may be deciding to stop work and start over. Too often, projects that desperately need to be reset are allowed to roll forward even as dev managers futilely apply minor course corrections or changes.

You can read an excerpt of the Forrester Research summary here.

Do you think Forrester is on to something? What secrets have you discovered that can help save a dying project? E-mail me at [email protected].

Posted by Michael Desmond on 08/29/2007

comments powered by Disqus


Subscribe on YouTube