The humble coin flip game is a tool every agile practitioner has used or experienced at some point. It allows people to experience the benefit of small batch sizes to the point of single piece flow.
I love it. But it's promise is flawed.
It breaks my heart to say it, however the panacea makes assumptions which are unicorns in the world of software development.
As an agile consultant I'm consistently being asked to help organisations that are striving for faster flow, yet, very rarely do I see the perfect storm of:
A stable, dedicated team
Each member of the team should be 100% allocated to a single team, indefinitely. Context switching is the single biggest cause of waste in organisations today. Check out the research from Gerry Weinberg, it's a powerful graphic I use often to expose the impact of asking people to work across multiple teams.
The right skills and tools for the job
Having dedicated team members is one thing, but having the right team members is another. When you try to plug square holes with circular pegs there are still gaps!
An equal backlog
Whether you use t-shirt sizes, fibonacci, #noestimates, or something else, every product backlog contains a wide array of complexity. Being able to create consistent, predictable single piece flow requires every item to be pretty much the same. Even when this is a possibility, it takes time to reach this maturity in the structure of the backlog.
A consistent, sequential, repeatable process
Pick your framework of choice (Scrum, Kanban, XP, etc), respect it, be disciplined, and trust the process. If you're constantly changing the process, skipping steps, or finding ways around it, then it's not going to serve you as intended.
Clear roles and responsibilities
This should be simple, after all scrum defines just 3 roles - right?
Organisations are more complex than this. A team is often required to interact with multiple business stakeholders, software providers, development/delivery partners, change management / PMO and so on. There may be project managers that are required to play a role.
The bigger the org, the more messy this gets, hence the need to articulate, document and socialise who does what & when. Note - this should be created by or with the team and not by someone external.
Focus
According to study on interruptions lead Gloria Mark, research from University of California Irvine concluded that when considering knowledge work “it takes an average of 23 minutes and 15 seconds to get back to the task.”
Reduce meetings, create agreed break times, and block time for focussed work.
In conclusion
Should we stop trying to optimise our flow? Do we give up on switching from a project to a product mindset? Are we wasting time on transformation programs? Is agile dead?
Of course not.
Small, incremental improvements are always possible. Experiments should be welcomed. Bottlenecks can be removed. Marginal gains compound over time to create big wins.
You may get frustrated along the way. You will need to exercise patience before having meaningful data that can drive decision making. But no, don't ever stop trying.
Comments