How finishing what you start makes teams more productive and predictable

Lucas F. Costa's data-driven argument for why teams should finish tasks before starting new ones. Using Little's Law and practical examples, he demonstrates that concurrent work increases cycle times - 'working on features concurrently causes features to take longer.' Context switching has high costs, disrupting engineers' flow state and reducing productivity. Little's Law mathematically proves that 'cycle times elongate when work in progress increases and throughput remains the same.' Teams improve predictability by delivering smaller, consistent batches rather than large, irregular releases. While some tasks require batching due to transition costs, in software 'the cost of holding software tasks' increases exponentially compared to manufacturing. His core recommendation: 'limit work in progress' and focus on completing one task before starting another. This approach maximizes team efficiency, reduces variability in cycle times, and creates more reliable performance forecasts - making teams both more productive and more predictable.