Limiting work in progress, seeing where the queue forms, and the theory that explains why both of those work.
Lean and Kanban start from the same observation, which is that work sitting in a queue costs more than work being done slowly. This list is for delivery leads, Scrum Masters and product people trying to make a busy team faster without asking anyone to work harder. It opens with the method and the board designs, moves to the measurement that shows whether a change helped, and finishes with the manufacturing and product development theory explaining why smaller batches and shorter queues win.
Anderson's original account of the method, covering work in progress limits, explicit policies, classes of service and an evolutionary approach that starts from the process you already have. Begin here, because everything below assumes it.
DeGrandis names five thieves of time, including too much work in progress, unknown dependencies and unplanned work, then gives a board design that exposes each one. The book for a team that is visibly busy and still slow.
Cycle time scatterplots, cumulative flow diagrams and throughput histograms, with the reasoning from Little's law that connects them. This is how you demonstrate that the work in progress limit actually did something.
Reinertsen supplies the economics, including queueing theory, batch size and the cost of delay that tells you what to work on next. Dense, and it is the theoretical backbone under most of the practices above it.
Goldratt's novel about a failing manufacturing plant, and the source of the five focusing steps for managing a constraint. The clearest explanation anywhere of why local efficiency and system throughput are different things.
Martin and Osterling take the technique past a single team to work that crosses functions, with a current state map, a future state map and a plan between them. The way to show a leadership group where the waiting really happens.
Constraint thinking applied to schedules, replacing task level padding with a pooled buffer you can monitor as the work proceeds. Relevant to any Kanban system where dependencies rather than capacity set the pace.
How lean thinking survives contact with governance, budgeting and audit, which is where most adoptions stall. Practical about funding a stream of work without committing to fixed scope in advance.
The three ways and the theory of constraints in narrative form, applied to an IT organisation rather than a factory floor. Useful for colleagues who will never open Anderson but need to understand why the queue matters.
Cutler's collected writing on the feature factory, including the signals that you are in one and the practices that move a team out. The sections on small batches and on making the cost of delay and the queue length visible are flow arguments aimed at a product audience.