Deployment is an Activity, Not a Milestone
Take a look at your project plan, particularly at the part that says "Deployment". How many times is this shown as a task? The chances are that there is only one - right at the end of the project - and this is on the plan as a milestone rather than a task.
So, according to that plan, you only deploy your application once and no time has been allowed for this - the underlying assumption is that it will just work.
But what about the other deployments that happen all the way through the project lifecycle?
- developers building and deploying to their sandboxes;
- deployments into the test environment(s) for testing;
- deployments into the staging / UAT environments;
these happen more than once - they are regular occurrences in any well-managed development project.
Deployment or Integration?
There will probably be an activity on the plan marked "Integration", and this will have an amount of time associated with it.
You know, from experience, that integration is something that is costly and unpredictable. It is the point on the plan when you start to bring all of the various parts of the application together. Like, for example, the first time you put all of the code into the test environment to be tested.
Time is spent, and wasted, debugging the code drop and making it work. This is normally put down to "integration issues".
But it is not integration. It is deployment - the bringing together of the application and an underlying infrastructure.
It is a thief, all the more so because it is concealed. It masquerades as "integration".
Intangible Costs
What is the cost of not delivering on time?
What is the cost of your customer losing confidence in the quality of your software?
What is the cost of resources being sucked into debugging and troubleshooting when they are planned for other activities?
The Tangible Costs
You know the hourly cost of all of the resources on the project. If a deployment takes one hour, then what is the total cost of those resources being unproductive?
Even worse, what is the opportunity cost of these resources being pulled into debugging the deployment when they should be getting on with other work? What does this do to your timeline and your milestones?
Multiply this cost by the number of times your application will be deployed.
This is the amount of money that is simply being thrown away, being wasted.
