Define the question precisely
Identify the variables that matter
Gather relevant historical data
Look for patterns and trends
Separate signal from noise
Use probabilistic thinking
Build multiple scenarios
Test assumptions against reality
Update predictions as new information appears
Monitor leading indicators
Consider second-order effects
Account for uncertainty and randomness
Compare forecasts with actual outcomes
Learn from prediction errors
Repeat the process continuously