Phased versus Big Bang

Implementation is best done in phases. Period. This used to be an important topic to write about but I think most of us stopped writing about this because we thought we got our message across. I suspect it’s time to start writing about this, again.

I was recently listening to someone telling me about the terrible list of failed LIMS and ELN projects out there and I suspected that these projects were those that did not properly manage their projects. I can only just suspect that without knowing more about those projects. We talk about using the phased (rollouts of manageable chunks) over the big bang (everything at once) approach.

With that said, one of the biggest factors to failure that I have seen over my career has been the inability of a project to prioritize what they need. Along with that, those types of projects tend to insist they put every single feature they need as one large rollout. There are a variety of reasons for this. It is not only the inability to prioritize but also, sometimes, the fear that they won’t get any future features rolled-out for them. They believe that they must put everything in, at once, or they won’t get it.

The unfortunate outcomes tends to be:

1. First of alll, it is never 100% correct in that first rollout. There are always some errors and rework to be done. If you rollout just what you really need, you have less to fix. And, possibly, the whole thing is so bad that it has to be pulled-back and postponed. UAT (User Acceptance Testing) addresses some of this, but it does not catch everything, either. Thus, if you rollout chunks of features, you have fewer interconnected things to fix.
2. While the big bang approach sounds as if you can just get everything out at one time and be done, even the best projects with the largest staffs don’t tend to have the coordination and skills that would be necessary to make something like this work. And, maybe it’s a myth that it really could work. Sometimes, we hear talks at conferences from customers who claim this worked for them. Consider that some might be a fluke and some actually aren’t telling the full story of the horrible implementation they’ve had. I know this is true because I’ve heard some projects give these talks — projects that I happen to know were horrible disasters.

Most companies know all this — you’re not reading anything new, here. However, I felt it was time to remind the few holdouts that are still trying to get this big bang working and struggling with it to realize that it’s almost certainly going to be a disaster.

Gloria Metrick
GeoMetrick Enterprises
http://www.GeoMetrick.com/