Tuesday, December 8, 2015

An initial problem with Large Scale Scrum (LeSS)

I've noticed a bit of a problem with Scrum: it's comes across as hopelessly idealistic and simplistic for many people hearing about it for the first time. If you work in an organisation of any complexity or size you will have felt this gut reaction to hearing a Scrum expert explain the framework: it seems to ignore the reality of the current organisational complexity, existing roles and management thinking.

I was in Sydney last week attending the Certified LeSS Practitioner training by the wonderful Bas Vodde.

At the training I found LeSS triggered an even worse gut reaction: as a direct application of Scrum at organisational scale LeSS seemed too simplistic to address the challenges across multiple teams, systems and functions. The challenges seem much more difficult at organisational scale compared to adopting Scrum at a team level.

I happy to say that after 3 days of stories, exercises, insights and some hilarious videos from Bas, I feel I got past this initial "it's too simple" gut reaction and gained a new perspective on Scrum and some new ways to think about agility at organisational scale. Here is one of the best bits from the training:

Actually creating real feature teams.  Feature teams done properly are for me the central idea for organisational agility in LeSS. This includes the restructuring into fully cross-functional feature teams that work across components in the architecture and implementing the related changes to reporting lines, job descriptions and engineering practices.

One of my favourite parts of the training was how Bas positioned feature teams as a way of forcing organisational agility through continuous learning. He suggested we welcome the problems that specialisation brings. Forming feature teams allows us to deliberately set up a conflict between specialisation and prioritisation. You could also see this as a conflict between efficiency and agility.

If a Scrum team's purpose is to always work on implementing the highest value customer features then there is often going to be a significant imbalance between skills and features. The current specialisations and systems knowledge available in a given team will often not be suitable to implement the selected highest value customer feature.

Bas suggested we exploit this imbalance by challenging teams to learn new systems, skills and domain knowledge and to use this imbalance to force this learning. On the other hand, when there is an available team that already has the skills to implement a feature we should take maximum advantage of the skills in that team by implementing that feature in the best team possible.

This idea of treating team members as expert learners, not as expert specialists was repeated throughout the training and was a challenging idea with many ramifications.

"I used to be a BA, but now I'm a team member"

Many of us had fears about the willingness of staff to be proactive learners and their willingness to work outside their specialisation. One helpful topic was a review of McGregor's Theory X and Theory Y management beliefs, and a reminder that these beliefs can be diabolically self-reinforcing: if you believe staff need a clearly defined job role to be effective this will create narrow work behaviours in staff. Bas pointed out shifting management beliefs takes lots of time, and probably lots of beer.

There were a lot of other gems coming out of the LeSS training, and I'll save these for a future post: scaling the Product Owner role, the feature team adoption map, differences from SAFe and new perspectives on team level Scrum.

If you are interested in scaling your agile adoption outside a single team feel free to get in touch, I'm happy to give an overview of LeSS or discuss how LeSS might be useful in your context.