Follow the work.

Fictional insurance customers need four work types. Solid pieces are reusable; hatched pieces are customer-specific. Each organization has four equivalent work slots. Customer teams divide those slots equally; specialist teams use one per discipline. The hybrid first uses its four slots to build shared cores, then moves that same capacity to customer pods. Setup and component transfer are included; no extra platform staff is hidden. Context caches start equally warm and misses add a fixed delay. All three have the same approval gates. The slow-team example uses four customers and slows one quarter of work capacity: customer B, Workflow, or the hybrid’s Workflow core builder (then customer pod B).

The hybrid is fastest in the default shared-needs scenario, not in every setting. It trades platform investment and a shared dependency for reuse with local delivery ownership. Long-term maintenance, staffing changes, and real approval queues are not modeled. The organizational pattern is informed by Team Topologies: stream-aligned and platform teams; all timings are invented for this illustration.

The top scorecards show the completed scenario, independent of animation progress. “Best here” means the shortest time to deliver to every customer, compared at the displayed precision of 0.01 time units; ties share the badge. Build work counts common and custom implementation effort, including repeated cores. It excludes context, setup, approvals, and assembly; their delays are reflected in delivery time. The bars use the same scale across all three approaches for each metric.