Leverage iteration

• Make early and meaningful contact with the problem
• Use high quality, rapid iteration to address uncertainty
• Focus early efforts around areas of risk and uncertainty
• Use iteration to deepen understanding successively
• Favor experimental results over expert opinions
• Use the desired end state as a guide
• Reevaluate assumptions based on iteration results
• Reduce the stigma of initial failures
• Rank knowledge by how it enables new capabilities

Increase iteration speed

• Increase adaptivity
• Increase decentralization
• Fail quickly, cheaply, and in diverse ways
• Combine local overconfidence with global stability
• Prioritize rewards for success over penalties for failure
• Prioritize iteration speed over defect rate
• Create solutions with resources immediately available
• Leverage models and simulation
• Increase selection pressure
• Define and measure short-term criteria correlated with long-term goals
• Increase overall speed with a mix of long duration, high intensity, and recovery activity

Increase iteration quality

• Define value and organize each iteration to maximize it
• Iterate to make value flow without interruptions
• Complete more iterations
• Do the hard parts first
• Use the scientific method when possible
• Prioritize a rough representation of the overall system over detailed representation of a single element
• Evaluate iteration methods frequently
• From more to less useful: Products, prototypes, mock-ups, models, diagrams, concepts, words
• Incrementally iterate abstractions rather than features
• Iterate in a series of stable, but progressively improved products
• Freeze features progressively
• Eliminate defects progressively
• Cherish and learn from exceptions and failures
• Reduce clutter and simplify
• Inject some randomness into exploratory processes
• Use more randomness in early iterations, less in in later iterations

Monitor and test

• Reflect after each iteration
• Consider what didn’t happen
• Temporarily eliminate safety margins to reveal problems and constraints
• Test each iteration in an real world environment
• Use both independent and in-system monitoring
• Identify and monitor thresholds that trigger instability