Okay. Not really a tip or technique, but something that I was reminded of over the weekend that some people might find interesting.

One of the important parts of a product manager's job is to identify and prioritize features that go into a given software release. This is often done using a spreadsheet of features. With limited development resources, you can never get to all the features, and so scoping of a release is a process of prioritizing and sizing features, and matching that to available time and resources.

Each release is a balance of 3 types of features: features that will help you generate new sales, features that make existing customers happy, and features to ensure the long-term viability and supportability of your products. Here is an example of each type:

• PeopleSoft Internet Architecture --> A big feature, yes, but it was driven completely by new sales.
• Change Assistant --> Existing customers wanted a better way of managing software updates and fixes.
• Refactoring of the component processor code --> although expensive to do, making this investment in the architecture would make it easier to support the products and easier to add new enhancements in the future.

So, as you can imagine, this can be a pretty involved process with lots of levels of complexity.

So, here's the trivia: With all the features being identified, priorized, and scoped, we in PeopleTools Product Management found an interesting trend. You could call it the Source Control rule. For every release since 1997, adding Source Control to PeopleTools was the highest ranking feature "below the line." We'd go through the process, haggle about resources and release timing, and boom... there it was... Source Control and everything ranked below it was out of scope.

Some people would joke: if you want to kill a feature, rank it after source code control.

Finally, the EVP of products and technology made an ultimatim: PeopleTools X would not ship without source control. Guess what... he was right because Oracle bought the company. So, the ultimate question was: could we have used the source control rule to predict the Oracle acquisition???

Labels: PeopleCode, PeopleSoft