Agile Development Blog

I believe that Agile project management in small, co-located teams crossed into the mainstream back at the sold-out Agile 2007 conference, but Agile program management at scale is just now heating up. Last week’s article on eBay’s Agile adoption in Business Week (combined with other recent news) shows us that organizational agility is becoming a mainstream topic at the highest levels.

The market is no longer asking, “Can we scale Agile across the enterprise and large distributed teams?” and instead is asking, “How do I get there?” and “Who can help me?”

BusinessWeek asks, “Can eBay Get Its Tech Savvy Back?”

Watch the video of eBay's Mark Carges on BusinessWeek.com

Author Douglas MacMillan says: “Carges’ plan for eBay is to take the “agile” method of software development epitomized by the daily deal widget and expand it to other areas of the site. New product pages will be customized to better accommodate different categories, such as jewelry and clothing. And the company is helping third-party developers create applications for eBay’s site such as a UPS-branded terminal for monitoring shipments” Read the full article >>

Author Note: Mark Carges was my boss and mentor at BEA.  I know a good bit about his current efforts and they are really going for it at eBay and PayPal.  It is great that their enterprise agility efforts will unfold in public eyes.

In other mainstream signs of Agile…

I was excited to see a couple of other great articles on Agile this week, including:

• Marketers often say you’ve reached the mainstream market when you notice your peers are doing it, and you feel behind enough to move. InfoWorld’s Paul Krill noted in his article Software Companies Jump on Agile Programming Bandwagon how many providers are “eager to hop on the agile development train.” (Clearly, we have an early mainstream market now.) See my post about traditional providers, including IBM, entering the space.
• Gantthead’s Bob Weinstein handily made the case for transitioning to Agile development in his article on Making a Case for Agile Project Management. He says:

“If ever there were an ideal time to make the leap from a traditional to an agile project management approach, it’s now. In this tense, uncertain, cost-cutting environment where CIOs are watching their bottom lines like hawks, the concept unfailingly proves successful. It not only delivers consistent, excellent results on time, but often under budget.”

• Finally, David Rubenstein from SD Times tackles the issue of Agile Development Built to Scale. I agree with Robert Holler that scaling anything across an entire organization is tough, and that sometimes Agile just gets a bad rap for something that is universally a challenge. It does take commitment from the team, a bit of training and a lot of inspecting and adapting to be the best Agile organization you can be.

The question these days is: how good do you want to be, by when, and who’s the best partner to get you there?

On our recent webinar “Demystifying Cloud, The Next Generation Architecture” we had a number of thoughtful and tough questions related to security, intellectual property and risks. We provided answers to these questions in the recording, but I found the recent SD Times article “Cloud Providers Answer Tough Questions” an even better source. In this article, a number of experts on specific platforms from Microsoft, Google, Salesforce as well as Rally’s own Zach Nies answer questions about security, lock-in and IP.

Henry Ford didn't foresee the impact of the first car - do we foresee the true impact of the Cloud?

Daryl Plummer from Gartner also did a great job recently describing the real point of cloud computing as he reviews Russ Daniels recent Forbes article. Russ says:

“In my view, the ability to facilitate innovation and entrepreneurship in this new model is one of the most promising ways to ignite the next wave of economic growth. We can no more see the full impact of the cloud than Henry Ford foresaw the impact of his desire to produce more cars in less time.”

As a result of SD Times’ tough questions and our desire to “ignite the next wave of economic growth,” we decided to talk in our next webinar with Global Logic and IBM about how to go to the cloud and mitigate risk along the way. As with any pilot, the goal is to enter wisely, learn fast and then move forward.  Given the iterative and incremental method of Agile is best suited for this fast-learning approach, we will title our next talk “Going to the Cloud – the Agile Way.”

We are structuring the content now, but I would love to hear your ideas, questions or feedback on this topic. I will also post a registration link for the webinar as soon as I have it.

Thesis: Taking a learning-first approach to your cloud efforts can help you avoid the risks of vendor lock-in, IP security and a spectacular failure

Proposed Agenda:

1. Review the innovation, benefits and risks
2. Typical approach – Choosing early, over selling, dramatic big bang
3. The Agile/Lean approach – Set-based, scientific and learning-based
4. Case study
5. Close

Jean in the mini-kitchen at Rally headquarters.

Jean has been doing a series of posts (What is with all the Agile process overhead? – part 1,  What is with all the Agile process overhead? – part 2, and How do you plan for unplanned work? – part 3.) on what goes in your backlog.  We have both noticed that it is really tempting to put the kitchen sink in your backlog.  At Rally, we only do progressive elaboration of stories based on subsequent levels of planning, which ends up looking a bit like this:

• Ideas and research topics managed by product line on the wall and adjusted across product lines every quarter
• Epics at the roadmap level (multiple release level) managed in a roll-up project in Rally’s Agile project management solution and estimated every eight weeks for release planning
• Stories at the release level managed in individual projects in Rally elaborated every eight weeks at release planning
• Tasks at iteration level managed by team member in Rally and elaborated every two weeks at iteration planning

As a result, we keep our backlog well groomed and do not waste time (muda) managing excess inventory or over-investing in stories that are not guaranteed to get built.  I have noticed it is really tempting to try and load all your ideas, suggestions and half-baked stories into a product like Rally.  We built Rally to make that uncomfortable – to try and discourage that.  Rally focuses you and the team on the iteration and not a big list of defects and bugs, while our coaches and technical account managers try to help folks move away from thinking of our system as an inventory manager.  We think of Rally and design it as a real-time decision support system.  (See my 2006 article in SD Times on “Kill your inventory manager” for more on this topic.)

As Jean and I were talking about this in our kitchen, I was noticing something  (see the picture). In this picture, there are no dirty dishes in the sink.  This is a relatively new occurrence here at Rally.  It happened when we moved from our old location, which had a huge kitchen and two dishwashers, to this new building. We doubled our square feet but shrunk our kitchen space in half with no dishwashers as a result of an in-house cafeteria that we share with Oracle and Quantum.

With almost no counter space and no dishwasher to batch up dirty dishes, you have to clean your dishes as you use them.  For the 150+ people who work at this office, that means a real just-in-time process of dirty it and then clean it.  As a result, we have less dishes, less wasted time and less wasted space due to managing the dirty dishes.  We also lost the typical sign that says, “Your mother does not work here! Please clean-up after yourself.”

Excess inventory leads to wasted time, wasted effort, and friction.  If you keep your backlog well groomed and focused on the valuable items, as Jean illustrates, you can keep from wasting valuable time.

This post is in honor of Greg Macaluso and Kevin Mindenhall, both manufacturing and distribution process consultants from Coopers & Lybrand.  In 1993, they taught me the way of JIT and Lean while working on an MRP project at Robinson Brick and a remittance processing project at U S WEST Communications.  In addition to tutalage, they had me read and flip the wonderful pictures of JIT, kanbans and manufacturing facilities with no horizontal space.  By removing the horizontal space to store stuff, the raw and work-in-process inventory went down and the throughput went up!