Tuesday, April 22, 2008

AgileScotland - “Extremely Profitable Programming with COBOL" - by Clarke Ching, Edinburgh, 12th of May, 2008.

Hi everyone,
Last year I managed a very successful Agile project, despite far from ideal circumstances. 
I want to share this story with you at the next Agile Scotland meeting because it shows how Agile can be used in everyday organisations, by everyday people, working on everyday technology to do better business.

On the first day of the project, I met with our project's key customer - a senior manager with one of the UK's largest companies - he said to me, "I don't know you Clarke, but we simply don't trust your client.  We've had too many bad expereinces".  Fortunately, my client had warned me to expect this and, in fact, my "promise" to them was specifically to "rebuild trust with their customer".  Three months later, mid-way through the project, they flew some of their experts down to visit us, to find out how we'd achieved such a turnaround.  A few days later we discovered – and fixed - our first defect.  By the end of the project we had rebuilt trust considerably, we'd profitably delivered the project as promised, and in doing so the customer made tens-of-millions of pounds in extra profit.  However, we very nearly screwed up the whole thing right near the end – I'll tell you all about that on the night.

The project was a success despite doing many things which some niave but enthusiastic agilists say cannot or should not be done in "agile" projects.  For instance:

  • My team were working on an old legacy system made up of COBOL, ancient Oracle and rubber-bands;
  • Our customer consisted of around 50 people spread across five sites, based in three countries. 
  • We promised (and then delivered) to an aggressive, fixed-scope, fixed-price, fixed-date contract.
I'll repeat it: I want to share this story with you all because it shows how Agile can be done in everyday organisations by everyday people working on everyday technology.  This project finished in 6 months - it would have taken between 10 and 14 months, if done the "old" way.  That delay would have cost my client the contract and it would have cost their customer tens-of-millions of pounds in profit.  Agile, Theory of Constraints and good-old-fashioned project-management did that. 
In this session I'll share a few of the "secrets" (things you don't read about in the currently published agile books) which saved the project, such as:
  • How to use fear, intimidation and genuine concern to force your customer to collaborate so that they can be successful (I am serious about this: this project would have failed if I hadn't spoken one very calculated, very sincere, sentence just at the right time);
  • How to build trust with your team by being lazy and getting them to do your job;
  • How to do TDD with COBOL;
  • And (as a special bonus!) I'll share with you the one rule we used to deliver more work in each iteration than anyone had ever thought possible.  Seriously.
At the risk of sounding like an infomerical …
... But wait, there's more!  Since I'm feeling rather pleased with myself about my "publishing deal" I would like to buy pizza and a few beers for everyone who comes along on the evening. 
If you would like to learn some new stuff from this story then I only have 15 spots so please send me an email: Clarke.Ching@SpiceUpIT.com.  We will meet in Edinburgh, starting at, 7:15 on the 12th of May, 2008.
Clarke Ching