Sweet (meant sarcastically)

Okay, when I was with Citibank ATMs from 2001 - 2007 we released about 4 major updates for the United States.  Four in  he United States while release hundreds of software releases around the world including: Chipped Cards for Malaysia and Taiwan.  Different chipped cards for Europe (UK, Germany, Greece and Belgium).  Conversion to the Euro from Italian Lira, Belgian Francs, German Marks and Greek Dracma.  Acceptance of Cash Deposits and immediate counting / crediting to the account. Ability to set transaction preferences.  Ability to transfer between 3 different currencies and withdraw cash in all three (Chinese Remembi, HK Dollars, US Dollars). ADn so many many more.
.
So let's say about 1 US release every other year.
.
In 2007 they moved to a new vendor model and outsource development to "save money".  The outsourced team would be about 33% cheaper (projected) than combined US / Indian team that was working on the ATMs.
.

Haza!

This sign went up this week.  Which means the "new" team has made 1 release from 2008 - 2013 (giving a February release credit as 2013).
.
So let's just assume the cost per year was $1,000,000 for in house development.  It wasn't but let's just use round numbers.  That means that a US Release (ignoring everything else) cost about $1,750,000 (7 years / 4 releases * $1 Million per year).
.
Now let's say that outsourcing saved everything they thought it would (ha ha).  The cost of a US release is now (given the same constraints) $3,300,000 (5 years * ($1Million per year * 66%)).
.
That, obviously, doesn't include the "Cost of Lost Opportunity" since your release cycle has moved from 21 months to 60 months.  Nor the "Cost of Non-Conformance" since the new ATM software doesn't do what the old one did (Ability to Set Account Preferences).
.
Great Job Brownie.
.
Because, of course, outsourcing always saves money.  The Brainiac that decided this has long since been given a big ass bonus and left the company.