Yup, I’m pretty familiar with this stuff… McConnell’s “Demystifying the Black Art of Software Estimates” was a big thing for me. :) In this *particular* niche of development, I am super-good at estimating… actuals on both budget and timeline tend to be within +/- 10% when I have a decent spec up front and work with a team I am familiar with.
In this case, the team has NEVER worked with the development tools other than a 1 week bootcamp 6–12 months ago, with 2,200+ screens even a 20% overage will be a deadly mistake, etc. etc. etc.
So my proposal was to implement the initial framework, which I SWAG’ed at 200–250 hours for a team of incompetents (and then said “400”), then have the team that will use it for those screens do 10 screens on that framework so we can get a more-accurate velocity using the actual team doing the actual work.
And then after all of this, with an estimate that came in WELL under $1m for 2,200 screens…
“Oh, that’s too much.”
Umm. What? I gave you a MIRACLE estimate, anyone who has ever done this work would thing I was a mad man to give a project that cheap…
And that’s what I mean when I say no one ever even thinks about ROI, never mind any other 2nd or 3rd order calculation. All they care about is the upfront and recurring sticker price.