The operating plan needs to show how its commitments fit together.

In 2013, my application services remit contained forty-one active work packages. Each had a plausible explanation for having started. Taken together, they were competing for review attention and specialist support in ways that the individual account plans did not show. Teams were repeatedly moving between work rather than bringing a smaller number of packages to a usable completion point. I had supported some of those starts and could not reasonably describe the congestion as a failure of effort by the delivery colleagues.

We reviewed the dependencies with account and technical leads and proposed pausing eleven packages while thirty continued. Clients affected by the pauses needed an explicit discussion about the sequence and the implications for their own plans. Some account leaders opposed the change because a visible pause could damage confidence more immediately than the less visible delay already occurring. I supported the smaller concurrent set and accepted JPY 63 million in transition and resequencing expense.

The paused packages retained their owners, commercial context and conditions for restarting. We did not remove them from responsibility or describe a pause as cancellation of an obligation. The continuing work received a clearer account of shared review and support requirements. Technical approval remained with the relevant specialists; the operating decision did not make an unfinished technical condition disappear. We also kept the original dates in the management history so the revised sequence would not be presented as though it had always been the plan.

The immediate result was an agreed change in the work allowed to proceed together. I do not claim a universal productivity improvement from reducing forty-one to thirty, because the packages were not interchangeable units of output. The useful change was a more credible set of concurrent commitments and a visible account of what had been deferred. I would now examine those constraints before authorising the starts. Waiting until work was active made the correction more expensive and required clients to revisit plans that should have been discussed more carefully at the outset.

A 2024 review examined twenty-seven proposed production-change windows across several managed-service accounts. Eighteen had the required operating support aligned. Nine contained unresolved conflicts involving shared recovery specialists or client-side availability. The individual change records were not necessarily wrong; the problem appeared when their conditions were considered together. A calendar that looked complete could still ask the same person to support incompatible activities or assume a client participant would be available without an agreed commitment.

I supported moving the nine conflicting windows, with JPY 42 million in additional scheduling and support expense. Commercial colleagues were concerned about explaining another change to clients, and delivery teams had already organised parts of their work around the original dates. Proceeding would have avoided that immediate inconvenience but left known conflicts unresolved. I put that choice into the approval discussion rather than allow the presence of a date to stand in for evidence that the arrangement could be performed.

The teams confirmed participant availability, clarified recovery support and agreed the revised sequence with the appropriate client contacts. Technical and security approvals remained separate requirements. We retained eighteen windows and moved nine; we did not classify the moved changes as cancelled or technically defective merely because their operating conditions were incomplete. The account recorded the reason for each move and the expense of the revised arrangements, including which other commitments had to be considered.

The outcome was the resolution of identified calendar conflicts before those changes proceeded. I resisted describing that as a quantified number of incidents prevented. We did not have evidence for such a counterfactual claim, and it would have made an otherwise defensible decision less credible. I would bring the shared-support view into the calendar earlier next time. Late reconciliation turns a normal operating choice into a correction that interrupts scheduled work, even when the people examining the individual change records have done their own work carefully.

In 2018 we reviewed twenty-two service incidents across the delivery operation. Eight included a delay in assigning response coordination because teams disagreed about where the incident had originated. The discussion of cause was necessary, but it was happening in the wrong place in the operating sequence. Colleagues were reluctant to accept coordination when doing so could be interpreted as admitting that their team had caused the problem. That left the response less organised while the technical investigation still had work to do.

I proposed a clearer separation between coordinating the response and determining causal responsibility. Some technical leaders worried that a coordinator without the deepest system knowledge might direct inappropriate action. Their objection was valid if coordination were treated as universal technical authority. We therefore defined the role around assembling participants, maintaining the decision account and escalating questions, while diagnosis and specialist approvals remained with the authorised people.

Management allocated JPY 24 million to operating guidance and rehearsals. The exercises practised assignment of a coordinator without requiring a prior finding of fault. They also tested how that person would obtain technical decisions and how the role would transfer if the scope changed. We preserved the later investigation process and did not describe the new arrangement as removing accountability. The aim was to prevent an argument about accountability from delaying the work required to understand and respond to the incident.

The result was an amended coordination process and rehearsed role boundaries, not evidence that every future incident would be resolved faster. The original eight cases explained why the change was needed; they were not reused as a performance claim about the revised approach. I would have involved the technical leads earlier in drafting the role. Their concern exposed wording that initially gave coordination too much implied authority. Correcting it helped us preserve specialist judgement while making the operating response easier to organise.

A 2022 review of twenty-six weekly delivery reports found nine weeks in which the baseline had been revised after urgent work arrived. The revised reports were internally consistent, but the original commitments had largely disappeared from the comparison. Management could see what the teams had done without seeing how often they had been interrupted or which decisions had changed the plan. I had accepted those reports as a reasonable account of delivery and needed to correct that expectation.

I asked the reporting team to preserve the original weekly commitment alongside the authorised changes. Some account leaders worried that this would make necessary responsiveness look like poor performance. I agreed that an interruption should not automatically be treated as a failure. The purpose was to expose the decision and its consequence, including situations in which accepting urgent work was the right choice. A report that concealed the change could not support that more useful discussion.

We reconstructed the nine altered weeks and identified fourteen client milestone changes linked to the interruption decisions. Nine revised dates had been accepted by the relevant clients; five were still under discussion. We reported those positions separately. JPY 31 million in additional out-of-schedule support expense was recorded without assuming that every yen represented avoidable waste. Some work had been necessary. The account needed to explain which choice had been made, who had authorised it and what other commitment had moved as a result.

The outcome was a reporting method that retained both the plan and its changes. It did not establish that urgent requests were inherently bad or that all fourteen milestones had been resolved. The five still under discussion remained visible. I would now define how to report interruptions when establishing the operating cadence, rather than ask for reconstruction months later. The retrospective work showed that a tidy weekly report can be truthful about completed activity while concealing the management choices needed to understand how that activity came to replace the plan.

When I joined Seikoma, I tried to improve consistency by bringing more delivery-sequence decisions into central meetings. By 2011, four account teams were attending twelve recurring coordination meetings each week between them. Some matters involved shared capacity and belonged in a collective discussion. Others could have been resolved by the account manager within an agreed boundary. My intervention had made routine work wait for central attention without consistently improving the quality of the decision.

I proposed removing ten of those recurring meetings and retaining two for unresolved cross-account dependencies. The change was not permission for account teams to override technical, security or client approvals. We defined which sequence decisions could be made locally, when another account’s commitment created a conflict and what needed escalation. Some colleagues preferred the protection of a central meeting because it made responsibility feel shared. I had to explain that a clear local remit offered a more usable form of accountability than requiring a committee-shaped conversation for every adjustment.

We spent JPY 9 million over four weeks on facilitation, guidance and rehearsal with the four teams. The local decision account retained the reason for a sequence change and its boundary. The two remaining meetings focused on dependencies that the account owners could not resolve independently. We examined examples during the transition so that an apparently local choice would not quietly consume support already committed elsewhere.

The direct result was a reduction from twelve recurring meetings to two, with explicit local decision boundaries. I do not convert the removed meetings into a claimed delivery saving or assume that all released time became productive output. The important correction was to my own belief that more central attention necessarily produced better coordination. I would now identify the decision that truly needs a shared forum before creating the forum. Operating discipline should make responsibility easier to exercise, not create an additional queue for decisions that responsible colleagues are already capable of making.