Problem: the portfolio grew faster than outcomes

Many initiatives started simultaneously. Each looked important in its own context, while the overall system filled with waits for decisions, analysis, another team, system access, or approval.

Dependencies lived in manually maintained diagrams and messages. Leaders saw active work but not the cost of parallel starts: more Aging, longer queues, and a lower chance of finishing the most important work on time.

What changed

RFC became the portfolio unit and Epic its deliverable part. Links between initiatives and teams moved into one system, while manually maintained dependency views were replaced with a more durable structure.

Portfolio reviews combined WIP by stage and work class, current Aging, recurring blockers, Throughput, and P85 System Lead Time.

Manage starts as well as finishes

The conversation moved from “how do we accelerate every team?” to “what should the system not start now?”

Signals triggered decisions: delay a new start, reduce scope, change sequence, assign a dependency owner, or stop work whose expected value was unclear. Probabilistic scenarios showed the consequence of current WIP versus alternatives without making the priority decision for leaders.

Result

Average Epic throughput rose from 16 to 54 per month, while RFC System Lead Time at P85 fell by 64%. The long tail and cross-team waiting became portfolio decisions rather than local team problems.

Business impact

Commitments include probability and dependency context; low-value work can stop before costly implementation; leaders see which relationship delays several initiatives; and headcount is no longer the first answer to weak throughput.

Diagnostic questions

  • How many initiatives are active, and who authorises a start?
  • Which dependencies repeat across quarters?
  • Is there a rule for stopping or delaying work?
  • Can RFC, Epic, team, and expected business outcome be connected?
  • Which constraint has the greatest portfolio-wide effect?

A delivery diagnostic can reconstruct these relationships from a small initiative sample before a large transformation programme begins.