Agile

Last week I attended the Agile Business Conference in London. With four concurrent tracks and well over 200 delegates this must of necessity be a report from a personal perspective.

In the opening keynote Diego Lo Giudice from Forrester encouraged the thinking that we are moving beyond IT centricity and IT ownership into a Business Technology world. This was supported by other speakers. Kevin Herry and James Yoxhall gave an outstanding talk on how IPC Media integrated agile, incremental project delivery with evolving business priorities by using value based backlogs and metrics. However it did leave me wondering whether the IPC Media situation was very specific because the business value could be measured with readily available web site measurement tools. In less well instrumented business situations measuring business value is a hard task, and this was reinforced in the special session on value based contracts.

Diego also used the metaphor of the Crown Jewels, a very British concept I guess, to advocate using Agile selectively for maximum benefit. Being an industry analyst of course he was almost bound to show some industry adoption numbers, and these certainly supported the assumption that most organizations are using Agile very selectively. This intelligence was also corroborated in my own discussions with numerous delegates who agreed that Agile is hard to scale out.

However to rebut this to some extent, there was an outstanding talk from John Flenley of SITA (Société Internationale de Télécommunications Aéronautiques). Together with four colleagues, he explained how the endeavour to replatform and modernize common systems and services used by airlines and airports worldwide had been transformed by a highly structured, very large scale Agile approach. In the process John demolished a number of sacred cows (apologies for the mixed metaphor) and described a functionally baselined modernization process using pre-sprint knowledge discovery and specification processes which fed multiple concurrent, distributed (i.e. not co-located) development sprints! In addition they contracted with no less than three outsourcing providers using fixed price contracts based on Function Points. Key to the program’s success was a tight focus on automation and metrics and measurement across the entire life cycle, and the use of rework measurement as a key to governance and program management. My own notes on the session were heavily annotated with the comment “back to the future!”

Another outstanding talk from Dave West of Tasktop, and perhaps better known as the ex Forrester analyst supposedly responsible for coining the term Water-Scrum-Fall. His theme last week was automating ALM, and here he was (see comments below on WSF) on firmer ground. I have long railed against the Agile community that is so heavily People and Process oriented, and so little interested in architecture and automation. So it was a breath of fresh air to hear someone evangelizing full life cycle automation across tool and participant ecosystems. He advocated the concept of the software supply chain in which ecosystems collaborate using tool integration, buttressed with defensive programming mechanisms and automated testing to establish separation of concerns. Yes! Sounds like a service architecture to me. Dave agreed there is a real need for better standards and meta models in this area, which are currently woefully inadequate, and it’s great to see this topic getting air time at last.

Talking about Water-Scrum-Fall, Manav Mehan from TCS recounted his experiences of using the hybrid method and argued that a) using Agile delivery together with largely unchanged planning and requirements  processes is a given for most enterprises and b) that hybrid is merely a stage in a longer term adoption and maturing process. He also described how they gradually evolved the planning and requirements processes. However he explained he received so much negative feedback about the name Water-Scrum-Fall that he changed the name to I³ (Interaction, Incremental, Interactive). Speaking to Manev later I asked whether in retrospect he wouldn’t have been better off just referring to the approach as agile? I’m afraid in my opinion, Dave W’s contribution in this area is less of a step forward and more like a legacy.

And finally to the topic that was almost absent from the conference – architecture. For me architecture (and technology) is at least as important in delivering agile business as people and process, and because I was manning the Everware-CBDI stand I had great opportunity to engage with delegates on the topic. And I can report wide agreement on this point. I attended one session on architecture and it was an interesting user experience, but to me this is an area that, not just conference organizers, but the agile community at large need to engage with.

There’s much more. I haven’t mentioned governance, which was in evidence, but to be honest didn’t impress me too much. More work to do methinks, particularly demand management and architecture governance.

I will leave the final word to Stephen Grafton who ably chaired the keynote and panel sessions. He said, “We are still struggling to define agile with a big A or not!!” And this summarized the conference. There is really good experience with Agile in smaller, tightly focused, or larger highly standalone situations. But for the typical enterprise looking to deliver more broadly based agile business, there’s a great deal of learning required.