When delivery slows down, most technology leaders respond the same way: add people (or AI Agents), restructure teams, improve tooling, or introduce a new framework. Activity increases. Meetings increase. Reporting increases. Throughput often doesn’t.

The uncomfortable truth is this: in a complex system, optimising teams does not optimise the system.

For over 20 years, I’ve applied Theory of Constraints (ToC) to large-scale, high-stakes defence systems where improving the wrong thing had real consequences. The same systemic laws govern software organisations today. At any point in time, a single constraint determines your organisation’s throughput. Everything else is noise. Improving non-constraints may create the appearance of productivity, but it will not increase outcomes.

In this session, I’ll challenge several common leadership assumptions:

Why utilisation is often the enemy of flow
Why adding capacity frequently reduces throughput
Why reorganisations rarely solve systemic delay
Why your constraint may not be where you think it is

We will explore a practical, leadership-level application of the Five Focusing Steps to product and engineering systems — including constraints created by architecture, governance, decision latency, and executive behaviour. You’ll leave with a disciplined framework for improving throughput without defaulting to headcount changes, restructures, or tooling initiatives.

If you’re responsible for delivery performance, this session will help you stop optimising parts — and start improving the system.

Your Teams Aren’t the Problem — Your System Is: Why Most Software Leaders Optimise the Wrong Things

Simon White