...It would take far more time to build the platform than you think...
Based upon my experience I could have developed the SQL database, the web page survey, and the reporting code in 2-3 hours. I agree that conditional and branching logic would add time but this last survey had none of these type questions.
I understand your point about 'reinventing the wheel' but my point, and the topic of this thread, is about changing the way APS thinks about IT.
I am pretty sure I know how the survey decision came down. It is pretty apparent that APS has several million dollars in cost overruns with the new building. This, coupled with declining membership, are causing a significant money pinch. So when the desire for a survey cane down, the 'inexpensive' choice was to use Survey Monkey.
I have no idea if APS sprung for a Survey Monkey subscription. There are some thing which suck about Survey Monkey. Their subscription model is not very flexible; for example you cannot remove their free advertising on the pages unless you pay. The final page redirect is also annoying to users. You often need to subscribe for an entire year even if you just want to do a single survey. They offer no refunds. And I won't post how it is done but it fairly easy to work-around the limit that only allows you to take it once. In other words, if I wanted to skew the results a person could have taken the survey multiple times.
But again, the point of this thread is about the confluence of APS and IT; not about specific IT tasks such as a simple survey.
Thanks for the reference on MarketSight, I'll check them out.