A start is an attractive event in a services business. The account team can tell a client that work has begun, the project has a team and the operating report gains another active item. The difficulty is that a start also makes a claim on attention already supporting something else. Review capacity, specialist judgement and production support may be shared across accounts even when the delivery teams are separate. I want the decision to begin to include those consequences, because they do not become less real when the start is commercially important.
This is not a preference for leaving work untouched. Clients have legitimate reasons to want progress, and an organisation that avoids every new commitment would fail in a different way. The question is whether the work can move through the conditions required to finish, rather than simply become active. Starting a package whose next dependency is unavailable may create activity without creating a credible route forward. It can also interrupt work that previously had such a route. The operating account should make that exchange visible.
The local plan does not contain the whole constraint
An account manager can prepare a reasonable schedule based on the people and conditions known within that account. Problems emerge when several such schedules draw on the same resource without a common view. A specialist may be named for a short review in each plan, yet the combined demand may exceed the attention available in the relevant period. Recovery support may be assumed rather than reserved. These are operating relationships between plans, and they need a decision process that can see more than one account at a time.
The answer is not necessarily to centralise every detail. A central office can become another source of waiting if it requires routine local choices to pass through a crowded meeting. I want a clear distinction between decisions an account owner can make within an agreed boundary and conflicts that require a wider view. The boundary matters more than the number of approval steps. If an apparently local start consumes shared capacity already committed elsewhere, the conflict needs to be examined before either account treats its date as settled.
Our 2013 review of forty-one active work packages made this problem concrete. Each package had a reason for being active, but the combined workload was moving through shared constraints inefficiently. We paused eleven through explicit client discussions and continued thirty, with JPY 63 million in transition and resequencing expense. That number does not establish that thirty was an ideal workload for every circumstance. It describes a particular decision about a particular set of commitments whose conditions we had finally examined together.
The objection from some account leaders was that a visible pause would damage confidence more than the less visible delay already occurring. I understood the concern. Announcing a start creates an expectation, and reversing it can be uncomfortable. But concealing congestion inside many apparently active packages does not protect the client’s underlying interest in completion. It mainly delays the conversation about what the organisation can credibly do. I would prefer to have that conversation before the start, when more choices remain and fewer people have arranged their work around an unsupported sequence.
A pause still needs an owner and a commercial account
Removing a package from the active list does not remove its obligations. The client still has a purpose, the account still has a relationship and the company may still have commitments that require attention. I therefore want a paused package to retain its owner, reason and conditions for reconsideration. Without that account, reducing concurrent work can become a cosmetic exercise in which difficult items disappear from the operating measure while continuing to create uncertainty elsewhere.
The commercial consequences should be explicit. A changed sequence may affect client plans, contractual discussions, staffing arrangements or expense. Those matters require the relevant owners and, where necessary, specialist advice. An operations team should not describe a pause as though it unilaterally changes every associated commitment. In our case, the eleven packages were discussed with the affected clients and retained their context. The JPY 63 million expenditure was reported as the cost of transition and resequencing, rather than converted into a presumed future margin benefit.
The continuing work also needs attention. A smaller active set does not automatically solve its underlying technical or operating problems. If a package is blocked because a decision is missing, reducing the number of other packages may help the decision-maker attend to it, but it does not supply the decision. I want the review to identify those conditions directly. Otherwise management can treat lower concurrency as a universal remedy and become disappointed when the same unresolved question remains inside a less crowded plan.
It is also important to avoid using people as interchangeable capacity units. The review should establish the relevant skills, authority and knowledge needed at each point. That does not mean the COO takes over staffing policy or technical judgement. It means the operating plan acknowledges the conditions those disciplines identify. A headcount total can provide context, but it cannot establish that the particular review, support or client interaction required next is available. The work’s dependencies need their own account.
Make the starting decision worth examining
The best improvement would be to move this reasoning earlier. Before agreeing a new start, I want to know what it will require from the shared operating environment and what existing commitment may be affected. The answer need not be an elaborate document for every small item. It should be proportionate to the consequence and clear enough that the relevant participants understand the choice. A short, accurate account is more useful than a large plan whose most important assumption remains implicit.
Management incentives can make that difficult. A start is easy to count and can look like momentum. The cost imposed on other work may appear later, in another account or another reporting period. The operating leader needs to connect those effects without treating every account request as unreasonable. I would ask which decision the measure encourages and whether the organisation can see the completion consequences of the starts it celebrates. The purpose is a more balanced account of progress, not a new slogan about starting less.
I also want the original decision preserved when the plan changes. If every revised schedule replaces the earlier one, the organisation loses the opportunity to learn why the starting assumptions were wrong. Some changes arise from legitimate new circumstances. Others reveal a constraint that was visible but not examined. Those explanations deserve different responses. A useful history helps management improve the next decision instead of treating each new crowded period as an unforeseeable exception requiring another round of urgent coordination.
My lesson from the forty-one packages is that concurrency is a management choice, even when nobody has explicitly made it. Allowing every local start can be as consequential as selecting a smaller set. I would rather own that choice early, with the affected people and clients involved, than ask delivery teams to absorb its contradictions later. Progress should be judged through work that can move towards an understood completion condition. A new start earns its place when the organisation can explain how it fits with the promises already in motion.