This tendency to leave the checks until the last minute explains the well-known 'demo-effect'. Your programming team has worked long and hard on the new application and next week there's the big trade show, where you're going to present the prototype to future customers. You ask your team to make sure everything is ready, and even the quality is good. Unfortunately they only have the Friday afternoon to do any checking and in the hurry they don't really find many problems. On the other hand, everybody realizes that should they find something critical, there wouldn't be enough time to fix it before the trade show, and that would be a catastrophe. Even the psychological setup is against a proper check. On top of that it has been long since the last check was done and the crew really only has time to scratch the surface with their check.

They find a few problems, which are probably masking at least as many, which probably prevent the testing of certain features sure to contain even more problems. All the found problems are fixed and everybody overly confident that all will go well. You pack the release and leave for the trade show.

The first two days on the show go extremely well. On the third day, brimming with confidence you invite the visitors to your demonstration. These are the ones you have been waiting for, these are the biggest prospects you've had in your mind. The demonstration better go as well as before. But even as you are thinking these thoughts, tragedy strikes. The demonstrated feature collapses without warning leaving you with nothing. You try to explain that the system is still under development and how this kind of thing is quite normal in software. The impression left on the prospects is not convincing, though. This demo-effect might have cost you half-a-year's sales. Afterwards, you curse your bad luck.

Of course, it wasn't bad luck, neither was it a coincidence that the application crashed just as it did. It didn't sense the presence of the highly esteemed visitors, or your nervousness. Nothing supernatural happened. As you packed the application on Friday, you also had packed wrong confidence based on the results of your crew's quality checks. For months, the application had accumulated complexity and errors, which nobody had had the time to remove or even check for. For months, the application had expanded in features but deteriorated in internal quality at the same time. The debt that the application had accumulated in internal quality, was what you had to pay back in the form of the demo-effect.

Have you ever wondered what created the idea of a demo-effect? It isn't just a myth that just when you need a piece of software to work, it doesn't. In the excerpt from Chapter 3. "Root Causes", I draw a picture of how it could still be going on in organizations today.