True flow requires a cultural shift (DevOps) and technical rigour (XP, TDD) to achieve "One-Piece Flow". It’s about moving work continuously without queueing.

​Read my full deep dive on identifying and fixing uneven flow here:
matthew-shaw.github.io/blog/2025/11...

Watch out for anti-patterns disguised as Agile. "Scrum-fall" (mini-waterfalls), weaponised Story Points (playing Tetris with velocity), and big-batch planning. These practices institutionalise the very unevenness and inventory we need to eliminate.

The solution is shifting from Push to Pull. WIP limits in Kanban aren't constraints on productivity; they are countermeasures to unevenness. They force the team to stop starting and start finishing. Queueing theory is clear: as utilisation approaches 100%, wait times skyrocket. Slack creates flow.

DORA metrics are a detection kit, not just a scoreboard. High variance in Lead Time is your warning light. If one task takes 2 days and another takes 40, your process is chaotic. We must optimise for consistency and reduced variance, not just raw speed.

I’ve mapped the 8 Lean Wastes (Muda) to modern software delivery, backed by evidence from Accelerate.

Stop optimising the wrong things.

Full breakdown here:
matthew-shaw.github.io/blog/2025/10...

The most damaging waste isn't in the code—it's 'Unused Talent'.

Treating developers as ticket-takers rather than problem-solvers kills innovation. DORA research proves that autonomy is a prerequisite for high performance, not a luxury. Trust your teams to design the solution.

'Overproduction' is the most seductive waste. We feel productive shipping features, but if they don't solve a user problem, we've just burnt cash to increase our maintenance burden.

Building the wrong thing efficiently is still a failure. Outcomes > Output. Always.

Take 'Inventory'. In a warehouse, it’s unsold stock. In dev, it’s that unmerged branch or the PR rotting in review.

​Unfinished code is a liability, not an asset. It blocks flow and hides defects. Until it’s in production, it is just risk.

​Stop starting, start finishing.

I've published part 2 now, which dives into the systemic cause of waste: matthew-shaw.github.io/blog/2025/11...

In case you missed part 1, looking at the eight wastes: matthew-shaw.github.io/blog/2025/10...

Of all 8 wastes, which one do you think is the most damaging in our industry right now?

Here's one with WireMock for comparison: github.com/matthew-shaw...