Case Study: Rapid Development of Custom, Mission Critical Software

by Pete Carapetyan


Starting Where Others Had Failed

This fortune 500 company had already invested millions in a 7 year project to integrate a large vendor's payroll system. 5 years in, they discovered it wouldn't do one mission critical piece, and not only that, but the mammoth, very capable application vendor had already tried to create this piece of functionality - and had given up after failing.

Since dataFundamentals had already established a track record of pulling rabbits out of hats for this department, could we look at the idea of creating a piece of software to integrate on the data side and put this functionality into a separate piece of custom software ?

Iterative Start

Tensions were high and the company was already in the throws of re-organization after the 2008 market melt down, so we prototyped at just a data and test level for a couple months with little more than myself and occassional management meetings to establish ever evolving requirements, and help from one of the company's programmers to furnish the external data feeds.

After three months of successful 2 week Agile iterations to prove that would work with real data and real results - we had an established and proven domain model. It was decided to bulk up the team, create a UI and bolt it on to the domain model we had created.

Four months later the small team reached substantially feature complete status - with a fairly impressive UI and back end feature set, and after dozens of serious requirements changes.


The corporation's applications were being outsourced to a firm in India just as this application was being declared 'almost ready'. Like many projects which failed in that outsourcing arrangement, this project spent over a year in continued development, and was then mothballed.


Hard knocks.