Institutions rarely fail inside one discipline. They fail at the joins.

Consider how a large institution typically changes itself. A strategy firm produces the plan. A systems integrator, chosen by a separate procurement, builds the platform. A security consultancy is engaged after the design is fixed. A design agency is asked to make the result pleasant to use. A communications firm is retained when the launch date nears, or when something goes wrong. Each adviser is competent. Each delivers what its contract describes. And the programme fails anyway.

It fails because the hard problems sit between the contracts. The strategy assumed a capability the system cannot provide. The system was designed without the controls the regulator expects. The service was redesigned around a process the strategy had already abolished. The announcement promised a date the delivery team never agreed. Nobody was paid to notice, so nobody did.

Where accountability goes missing

The pattern is familiar to anyone who has sat through a programme review. Every workstream reports green. The whole is plainly red. The reason is structural. When advice is divided among firms, accountability is divided with it, and the gaps between the pieces belong to no one. The client is left to integrate work it hired experts to do, often without the staff to do it.

When advice is divided among firms, accountability is divided with it. The gaps between the pieces belong to no one.

Procurement rules make the problem worse. Breaking a large engagement into separate lots can look like prudence: more competition, lower prices, less dependence on a single supplier. But each lot is scoped in isolation, and the joins between lots are rarely written into anyone's terms of reference. The institution saves on fees and pays in failure.

What integration actually requires

Hiring one large firm for everything is a tempting answer, and a flawed one. Scale brings its own risks, including conflicts of interest when the same adviser designs a system and then audits it. What matters is that one person is accountable for the whole problem, that the team includes every discipline the problem needs from the start, and that someone outside that team checks the work before it reaches the client.

In practice this means three habits. First, the decision and its delivery are planned together: no strategy is final until the people who must build it have tested its assumptions. Second, security, design and communication enter at the beginning, when they are cheap to accommodate, rather than at the end, when they are expensive to retrofit. Third, the engagement is organised around the outcomes the institution cares about, and judged by them.

Questions for leaders

Leaders who suspect their programmes are failing at the joins can test the suspicion with a few questions. Who, by name, is accountable for the whole programme, and do they have authority over every adviser? When the strategy changed last, how long did it take the system design to follow? Was the security team consulted before or after the architecture was fixed? Has the launch announcement been agreed with the people who must deliver it?

If the answers are uncertain, the risk lies in the spaces between the workstreams, and that is where attention should go first.