The biggest issue for software schedules

Every time you create a program you are creating something new with new challenges, tools and

requirements.  At the same time business requires schedules and costs to figure out how to bring a product to the customers.

Scheduling and estimates is a never-ending topic in software. Different estimation techniques have been employed over time.  From a simple one like “take whatever engineer tells you and multiply it by two” to Scrum points and velocity calculations.

My team spends about an hour after every sprint doing a “Sprint Review” meeting where we discuss things that accelerated progress or derailed our plans.  The finding of what has screwed up the schedules the most has been quite surprising to me.

I was expecting that it would be one of the usual suspects such as “engineers are naturally optimistic” or  “no one anticipates the level of detail”, however the most common source of delays have been… waiting on dependencies!  Engineers might have been absolutely correct about anticipating the level of effort and might have taken into account the complexity of requirements and even time to do bug fixing, but still missed their anticipated delivery date.

What are some of those dependencies?
1) waiting on a bug fix from another team
2) working with an external vendor to do localizations
3) waiting on access to legacy source code
4) waiting on a customer to give additional data for a bug repro

This all made me think about how valuable program managers can be.  I remember working at Microsoft a decade ago where some engineers said: “developers develop, testers test and program managers – stay out of the way”, but it turns out that these busy bees of tracking dependencies, facilitating conversations and documenting schedules are really quite impactful.  Of course when everything goes well and all dependencies are there just in time no one notices that, so to compensate for it find a Program Manager near you and give them a hug! ☺