≑
Think Read Stream All Posts
Subscribe
01 Jul
Why Hiring the Right People Won't Fix a Broken System
Most reorgs replace people. The good ones replace the system around them.

Most reorgs replace people. The good ones replace the system around them.

The classic example is the NUMMI plant. Before Toyota's involvement, GM's Fremont factory was infamous for strikes, absenteeism, and sabotage. Workers would deliberately leave a Coke can rattling inside a finished car to annoy the customer. Same plant, same people, terrible output.

Toyota didn't replace anyone. They changed the system. New processes. Quality over quantity. Every worker could stop the production line by pulling a cord. Lights stayed on at night to prevent theft. Within months, the same workforce was building high-quality cars.

At Sberbank, we restructured without replacing staff. Cross-functional teams with clear product focus cut time-to-market by four times. Same engineers. Different system.

The same principle applies to platform engineering organisations today. A platform team that struggles with adoption, slow delivery, and burnout often isn't a hiring problem. It's a system problem β€” incentives misaligned, ownership unclear, processes designed for output rather than outcomes.

Jay Galbraith's Star Model breaks organisational design into five elements:

  • Strategy: mission, goals, markets, products
  • Structure: how authority and roles are organised
  • Processes: how divisions and roles coordinate
  • Rewards: how employee goals align with company goals
  • HR Policies: how people are hired, developed, retained

Culture isn't a direct part of this model β€” it emerges from how the five elements interact.

Change one element, and the others must adjust. Misalignment between them weakens the organisation.

A platform team optimised for delivery speed but rewarded for ticket-resolution metrics will struggle. A cloud team responsible for cost discipline but with no influence on procurement decisions will fail predictably.

The system sets the path of least resistance.

If the system favours individual contribution over team outcomes, people will follow the easier path. If the rewards measure output (tickets closed, features shipped) instead of outcome (adoption, satisfaction, business impact), behaviour follows the metric, not the intent.

  1. Explore the system. Don't redesign from a slide deck. Pick one team. Spend ten days with them. Run interviews and workshops. Map what actually happens, not what the org chart suggests.

  2. Visualise the five elements. Draw out current strategy, structure, processes, rewards, and HR policies. Look for misalignment. The conflict points are usually where dysfunction shows up.

  3. Plan changes as a system. Identify the smallest set of changes that move multiple elements into alignment. Avoid changing one element in isolation β€” that's how reorgs break things further.

  4. Implement gradually. Start with one team. Monitor. Adjust. Then roll forward. In my experience, companies that try to flip the entire org simultaneously don't make it through the transition intact.

The temptation in any underperforming organisation is to point at people. Fire the underperformers. Hire stars. Bring in a new leader. It's almost always cheaper and more effective to look at the system first.

Continue reading