Predictable Agile Delivery
Business leaders say they want Agile Delivery, but that's not the whole story. They actually want Predictable and Agile Delivery.
Large organisations need to make forecasts so they can make prioritization decisions, make plans and mobilize resources accordingly.
The traditional model is optimized for (mass) production at scale of identical things which does not suit customized services or knowledge work, in which the deliverables are always different. Neither does it allow for the deliberate non-optimization of secondary (supporting) teams.
In this talk Russ describes the mechanisms that make the output of delivery teams predictable, so they can be trusted to produce reliable cost estimates. He provides examples of teams that have succeeded and reveals the symptoms of those who are yet to reach this milestone.
He then describes how the organization can be re-engineered towards a value-delivery model, citing the inverted pyramid as one example.
Outline/Structure of the Keynote
The traditional "machine model" as is and why it doesn't work for agile
The ideal to-be model, with Haier as the case study example for inverted pyramid
Feedback mechanism - using delivered value to predict capacity
What good looks like
Symptoms of avoidance and getting the feedback wrong
Advanced teams situation using noestimates
Current model is optimized for (mass) production at scale of identical things. This does not suit services or knowledge work in which items are always different.
In such cases, the operating model has to be changed to allow self-organization and ownership of production planning, design and development. This is basic agile, but it takes time for large organizations to realize this disparity and act upon the learning.
Trust is needed and can be earned by the team through getting better at understanding and predicting their work. Eventually, when trust is established, they may transcend the need to estimate.
executives, managers, VPs, change agents, team members