How do you decide to pull the plug on a project?

Over the past six months, I’ve been involved with two projects that, to put it politely, didn’t go exactly as planned.

Project A was cancelled, after a significant amount of preparatory work had been done, and project B was just released, after a much more significant amount of effort was expended.

What caused one to be cancelled and the other to be continued?

• Timing: A had a tight schedule, with a deadline that made sense from the business’ perspective.  When it became clear that the deadline would be missed, cancellation was a logical option.  B had goals but no deadline, which made it harder to make a go/no go decision
• Risk: A required some server rejiggering that would have made it extremely difficult to roll back.  B, on the other hand, would have been a simple software re-release to roll back to previous, somewhat working, software.
• Business impact: A was a ‘nice to have’ project, whereas B was addressing something that had caused employees and users pain for years.
• Sunk costs: A had fewer sunk costs than B, which paradoxically made it easier to cancel (paradoxically, because you should not consider sunk costs when looking at an investment).

What is the difference between pushing through to finish a difficult project and polishing a turd that you really should abandon?  It’s a fine balance, and as project B dragged on, I wasn’t sure at times which path we were headed down.  As someone who loves to ship, it’s hard for me to give up on a project that I had a hand in building, but no business writes software in a vacuum, and those business needs can and do serve as valuable checkpoints on the software process.  Because of the huge business value, it made sense to push through on project B.

Sometimes business priorities change, or, as in the case of project A, deadlines impose a different set of priorities (beyond the purely technical).

Regardless, I think that the more I consider any projects I have dragging on through the lenses of timing, risk, return and sunk costs, the easier it will be to make a go/no go decision.