Agile Development & Scrum at Portland General Electric

Great presentation this morning at AgilePDX Westside, by Mark Menger of PGE. Here’s my notes…

 

  • We had to start by defining what our product is.
  • A team mission was not accidental. Our vision is:
    • Keep IT Application Teams fully occupied by delivering quality solutions to the most valuable work for PGE in a predictable fashion.
      • Fully occupied – meaning time not wasted (and not overuse – forced o.t.), and considering creative/career needs
      • Quality – Improve breadth and depth of regression and functional testing (manage projects with a focus on quality)
      • Most valuable – driven exclusively by the business (with IT as a business partner)
      • Predictable – Make commitments to deliver and fulfill them
  • We embedded BA’s on the team.
  • Retrospective training was extraordinarily helpful. Some of their retrospectives focus on particular aspects of software development.
  • They do CSPO (product owner) training and Community of Practice meetings.
  • They do rolling annual capacity plans
  • Metrics for assessing planning are Avg Story Age, Planning to Working Ratio.
  • They’re beginning Hack-A-Thons. This is where you find 10 people, and you have 1 day to build whatever you want.
  • Technical Scrum of Scrums are important. This is done 3 days/week. A tech rep from every team – including infrastructure – meets and discusses planned deployments, versioning, services, etc.
  • We are still trying to improve quality – not just by rejecting bad results – but on preventing bad delivery. (In other words versus fixing bugs go further up the chain and improve the processes that put them there.)
    • QA testing embedded
    • Automated regression testing
    • Communities of Practice (Java/.NET/ testing/scrum master, etc)
    • Technical Debt Management with Senior Leadership
  • The senior management involvement on technical debt was not a small factor. Executives had to be taught that debt is like a car – sometimes its acceptable to incur that debt, and make payments, to meet the needs of the moment – as long as you make your payments. Otherwise you get repo’d! So incurring technical debt in the short term is OK, as long as that technical debt is not deferred too far in the future.
  • We’ve had to put more of an emphasis on speaking the language of the business – if your team can’t do that, you need to find someone that can articulate your application in business terms for you.
  • Planning Poker stinks (see!) – instead we use card sorting exercises, which use relative sizing. Really mature teams leave behind the whole idea of estimating, and make sure their sizes are all of a relative size.(Mark doesn’t believe this)
  • we have a pool of 10 scrum masters we pull from – “This is your buddy to help you learn about the scrum process” – is useful for projects that map across multiple teams.
  • Success does follow a Gaussian pattern – some teams will be high performing, others low performing, most will fall somewhere in the middle.
  • Language is a powerful tool.

 

So all in all, it’s a success story – and they started small. “In the beginning, IT was not perceived as a partner. We had on operations budget, and our mandate was delivering more with less – this forced us to change. The AMI project was a successful pilot of scrum – the project was way behind, really in crisis, and agile delivered fantastic transparency – it really saved the project.”

 

 

In other notes, I have a friend in the cube next to me that’s going to work for another company. He’s worked here for six years. When I talked to his manager, he mentioned that he wouldn’t try to stop him even if he could – that it’d be good for his career. Another friend echoed this – he said, “I have conflicting opinions about people that have worked for the company too long.

 

I tend to think that, for most people, effectiveness degrades with years. It seems to me that some people’s job security is tied to how loud they can be. In fact job churn is linked to increased GDP. (http://www.economist.com/node/21547224) This does not mean that an employee can’t be effective for 5 years or so, but after that you are pushing it unless you are moving to different departments or changing responsibilities.”

 

If you get a chance, read that article cited. It’s interesting… and the 8% payoff on average explains why the life expectancy of most of my jobs has been 18-24 months. (That and my schtick paling with time!) The article cited mentions:

In the typical job-to-job move (that is, without any intervening stint of unemployment) an American worker can expect a rise in wages of over 8%. This gain represents, at least in part, an improvement in productivity. As workers obtain skills and find better job matches, their output and earnings rise. And as firms obtain ever more suitable labour, they can afford to pay higher wages. In this way, the churning of the labour market contributes to growth in the potential output of the economy.

Aly Chiman

Aly Chiman is a Blogger & Reporter at AlyChiTech.com which covers a wide variety of topics from local news from digital world fashion and beauty . AlyChiTech covers the top notch content from the around the world covering a wide variety of topics. Aly is currently studying BS Mass Communication at University.