Value Optimisation · EST. 2011 · APAC

Features taking months to arrive. The bottleneck wasn't the team. A 50% reduction in delivery timeframes at the ATO.

The Situation

A major ATO programme had been operating as a product organisation for two years. Business owners had invested in the transition. Demand management was working. Prioritisation against customer value was in place. The flow of work from idea to release was improving — but features were still missing committed release dates, and business owners wanted to know why.

ZXM was engaged to examine the feature readiness process and identify where value was being lost before work reached delivery teams.

What the diagnostic found

Before any recommendations were made, ZXM mapped the complete value stream — tracking each step from feature concept through to release readiness, measuring cycle time at every stage and batch size at every handoff point.

The finding was clear. Delivery teams accounted for approximately 5% of total feature lead time. The bottleneck was not in the teams at all — it was in the upstream process that determined whether a feature arrived at planning ready to be worked on.

Three structural conditions were producing the delay. Features were developed in silos with sequential handoffs rather than parallel collaboration, which extended development time independently of team capacity. Senior executive endorsement was required mid-feature rather than at completion, creating mandatory wait states that had no relationship to delivery readiness. And features were presented for planning consideration too late to be adequately refined — arriving in the final weeks before a planning interval boundary, when the opportunity to sequence and slice them had already passed.

The product roadmap did not yet provide sufficient transparency to allow upstream contributors to plan their work against delivery timelines. The result was a system that produced features on its own schedule rather than the programme’s.

ZXM mapped the equivalent processes across the other major product verticals in the ATO and confirmed the pattern was consistent.

What changed

The intervention targeted the structural conditions, not the teams.

Feature endorsement was shifted from a mid-development checkpoint to a single approval at completion, eliminating the rework triggered when approved features were subsequently split or refined. A cross-functional pipeline team was formed across solution design, business representatives, and subject-matter experts to work collaboratively on features from brief to ready — replacing the serial handoff model. Feature briefings were introduced at the start of development to align scope and identify slicing opportunities before writing began, rather than after.

The programme also established a transparent product roadmap visible to all stakeholders and introduced a weekly demand assessment cadence, replacing the pattern of features being presented in bulk at the close of each planning cycle. A feedback mechanism was put in place to flag features at risk for the next release, with reasons, early enough to act.

What held

Feature development timeframes improved by 50% for value streams that adopted the new process.

For standard features — those operating within established patterns and using intentional design — lead time from concept to planning readiness dropped to four to six weeks. Complex or novel solutions, where new thinking was required, ran to three to six months: a range that was now predictable and visible rather than variable and opaque.

The re-approval requirement — which had been adding one to two months to features that were split or refined after endorsement — was eliminated. Product management gained the ability to assess incoming features and demand signals weekly rather than processing everything in a late-cycle rush. Collaboration between business owners, architects, and delivery teams measurably improved.

The 50% improvement in cycle time was a consequence of addressing the system. The teams had not changed. The work they received had.