Another former ‘Softie Evangelist, Alex St. John is blogging about the DirectX days.  Lot’s of interesting ideas and history in his blog, but one topic caught my eye: Evangelize early.   I couldn’t agree more.  Alex talks about the DX team working with the game industry before anything in DX was really working, making it participatory and getting the game developers to invest long term by working along side the team on the same timeline release to release.  Great tactic if you can pull it off, and was in the Microsoft Evangelism Playbook for years.

This sounds pretty early doesn’t it?  Well, I’ve got an example that is even more extreme.  Believe it or not, I evangelized Windows NT for Microsoft when it was just a stack of papers.  I flew around the world with some of the leaders of the original Windows NT development team conducting design reviews of Windows NT *paper specs* with our key ISVs before any code was running, not even a demo.   Why? Remember at the time Windows was moving from 16 bit to 32 bit, adding real multitasking and virtual memory as well as the ability to implement protected subsystems running things like POSIX  compatible API shells.  We were interested in whether ISVs wanted Win32 to track the semantics of Win16, or could we “do the right thing”.  Pretty universally the answer was, do the right thing and don’t saddle us with a legacy design from Win16.  Whew, that’s exactly what we wanted to hear.

Doing things this early is quite risky, you tip your hand to the competition.  You may be in fact wasting your partners time if the product either doesn’t ship or is radically different in the end.  But in this case, it was warranted and we were rewarded with both early support and an architecture and an API pattern that served the industry for 20 years.

Our BlackBerry 10 evangelism effort implemented some of these best practices.  A full nine months before launch we provided more early seed units to the worldwide developer community than BlackBerry or anyone else in the mobile industry has before, to the best of my knowledge.  While we didn’t review paper specs, (that would have been at least a year before I joined the company), app developers did incrementally familiarize themselves with our early builds of BlackBerry 10, provided feedback along the way and ultimately rewarded us with an industry record level of early support for a new platform at launch.

When is the right time for you to begin evangelizing your product?