Monday, December 12, 2005

Breakfast in NY with Ken Schwaber

Finetix hosted a breakfast for about 30 senior people from NY banks who were interested in hearing about Agile in Capital Markets. Ken Schwaber came along to add further weight to the Scrum argument.

We took a reasonably Agile approach to the session: we did a basic introduction to Agile, Scrum and XP including how they dovetail and what the benefits are when building Banking systems. We then opened the floor to very interactive Q&A.

Most of the questions were not a surprise:
  • Can you do this if your QA is offshore? [yes, and...but...]
  • Do you mean do the architecture in 30days? Then the design in 30days? etc [no...]
  • Can you do Agile without XP? [yes, Scrum...but...]
  • Do you mean get the QA involved early to sort the tests out? [kind of but more...]
  • Can you do this with junior people, you've all done it with good people? [no, done it with all sorts of people...]
And most of the statements were challenges that were straightforward to handle:
  • Developers won't want to Pair Program [not 100% true, and...]
  • You need documentation for when the system gets taken over in 3 years time [you ain't gonna need it if you don't deliver a system PDQ]
  • You need to design an architecture first! [er, no you don't if...]
  • 40-Hour Week, hah! I work that a day! [empirical data 40 useful hours...and you can do other things...]
It was great having Ken there to lend examples and data from other large, successful organisations outside Banking. Thanks, Ken.

We did something right as I ended up visiting several institutions as a result of this breakfast to further the Agile discussion in their particular contexts.

And in each place I ended up going through the traditional conversation along the lines of:
They: "That's what we did when [things went wrong]. We had [horrible deadline X] and we worked as a small team, met every day, did what it took and delivered"
Me: "Yes. Did it work?"
They: "Yes - but it nearly killed us working at that rate"
Me: "So what happened afterwards?"
They: "We went back and took time to rearchitect the system properly over [n>10] months"
Me: "Did that succeed?"
They: "Sort of - it was late. And it wasn't as good. And the users weren't happy"
Me: "So why did you change your successful way of working?"
They: "Errr"
Furrowed brows
They: "We couldn't keep working at that pace!"
Me: "I agree. You clearly know what it takes to succeed you just keep going back to old habits when the pressure is off. Why not work that way all the time, but at a reasonable pace, with reasonable deadlines?"

(Its becoming unamusing to repeat this conversation with people)
The Banking IT world knows this model of delivery probably better than any other! The hard, high pressure deadlines are often externally imposed by regulations and law. It is amazing that Financial institutions haven't embraced Agile development more.

I found there was a lot of misunderstanding about what Agile Delivery entails. Everything from fear of XP and believing its just about programming, to the "it's a license to hack" attitude. Agile seems to be a dirty word? Why I wonder? Is it the fear of change and not understanding the impacts, or is it something more?


At Monday, December 12, 2005 2:02:00 pm, Blogger Crispin Rogers Johnson said...

I think fear of change has a big part to play in this, mostly driven by poor/painful memories of past change resulting from failure to manage risk, etc. Something of a Catch 22. The only tool I am aware of to address this is an open retrospective, with emphasis on the "open" part.

My current client hates change but yet has started a Ruby on Rails project out of fear that they may lose an edge. One occaison that fear has worked in my favour.

At Monday, December 12, 2005 3:37:00 pm, Anonymous Crispin Rogers Johnson said...

I am working on the first XP project our company has undertaken and fear has been a very big issue for us. Creating a bubble of project safety is essential to any agile project. Any tips, hints or useful URLs that provide techniques would be much appreciated.


At Monday, December 12, 2005 6:09:00 pm, Blogger John S Nolan said...

I wish there were an easy answer.

Primarily, is there a timebox of effort aimed at achieving a quick win, where the team has commited to do the-whole-XP without question? (Including stand-up and retrospectives)

"Project Safety" is a term I dislike. XP helps you manage the risk properly in a project, the feeling of safety is about how the people in the team feel. You need to get people to address their personal worries and ensure that they are addressed directly. That's mainly management, and XP is not a mangement mechanism.

- get someone who has done it before to help. Primarily a change agent / scrum master / facilitator type (not necessarily an XP Coach, who to me concentrates on the practices)
- if you can;t get someone else (say due to resource restrictions) set time aside for the team to talk about these things (i.e. retrospect more often, but quickly 1/2 hour sessions) (This is a bit like self-managing teams which I think is an oxymoron and bad risk management but sometimes necessary in a vacuum of good management)
- recognise its a change management issue not just an XP issue and address it as such (look at Kotters change failure points and things like High Performing Teams)
- look at the Scrum ideas and use them for managing the goals (Scrum is managment, XP is delivery)
- Facilitate discussion of fears

A few 'easy' questions that need a shared view:
- is there an urgency to do XP, a real need? What is driving you?
- is there a commitment from the team to do XP, at least for a timebox of say 30-days before analysing?
- is there a simple vision of why your doing XP and what the world will be like when you're doing XP?
- is there a facilitator (probably external) helping to identify and break you all of your learned habits? (difficult to do the thing and look at the process of the thing)
- are you aiming at a few short-term quick wins that will make-or-break your efforts?

Of course the real easy answer is pay for some of my time ;-)

(and plese don't go for the easy joke here, Crispin!)

At Wednesday, December 14, 2005 4:14:00 pm, Anonymous Crispin Rogers Johnson said...

Thank you for the suggestions Johnny. As a priority I will look at the questions you suggest. As much as I'd like to hire you (or ANO Agile consultamt) we don't have any budget, so we are reduced to DIY ;) (at least I didn't go for the most obvious joke!)


Post a Comment

<< Home