The handover problem
The people who negotiate an arrangement spend months building a shared understanding: why it exists, what was traded for what, which points each side conceded and why. Almost none of that is written down, because it is context rather than obligation.
Then the document is signed and the work passes to operating teams who were not in any of those rooms. They receive a document, a target and a counterpart they have never met. Predictably, they interpret ambiguous clauses in their own favour, not out of bad faith but because they have no other information.
The cheapest remedy is a deliberate handover: a session in which the negotiators explain the intent behind the terms to the people who will operate them, on both sides, together. A short written summary of what the arrangement is for and what each party expected to get from it is worth more in the second year than any clause.
Sponsors
An arrangement needs a senior person on each side who wanted it and is accountable for it. That person's job is not to attend meetings. It is to make the arrangement's claims real inside their own organisation: to get it into plans, to make sure someone's objectives reference it, and to arbitrate when it collides with a departmental priority.
Sponsor departure is one of the most reliable predictors of decline, and it is worth treating as a scheduled event rather than a surprise. When a sponsor moves on, the arrangement should be formally reintroduced to their successor, who is free to conclude it is not worth continuing — a clean decision that is much better than a slow fade.
The operating layer
Below the sponsors, working arrangements are usually simple and should stay so. A named counterpart on each side who can get things done without escalating. A standing group where the people doing the work meet, sized to the actual joint activity rather than to the ambition of the announcement. A shared list of what is in flight, visible to both sides.
Cadence should match the work. Monthly is right for arrangements with continuous joint activity; quarterly for arrangements that produce a few substantial things a year. The common error is a heavy meeting calendar around thin activity, which trains everyone to regard the partnership as a burden.
What a review meeting should actually do
A review that consists of both sides reporting good news is worse than no review. Four questions make one useful.
What did we say we would do since the last review, and what actually happened? What is blocked, and which of us is blocking it? What has changed in either organisation that the other should know about? And is this still worth doing — with a genuine willingness to answer no.
The last question is the one that gets omitted, and its omission is why so many arrangements continue for years after everyone privately knows they have stopped producing anything.
Escalation
Escalation routes exist so that disagreements move upward before they harden. A workable route names the levels — working contact, sponsor, senior decision-maker — and puts a time limit on each, so that an unresolved issue rises automatically instead of sitting with someone who cannot solve it.
The cultural point matters more than the mechanism: escalating must be normal rather than hostile. Where escalation is treated as an accusation, people avoid it, and problems are instead solved by quietly doing less. That is how arrangements die without anyone recording a dispute.
Information flow
Partners cannot plan around each other without warning. A supplier changing a specification, a partner reorganising its sales team, either side entering an adjacent arrangement — these are the events that damage trust when discovered late and cost almost nothing when shared early.
Most arrangements would benefit from a short standing list of things each side undertakes to mention: changes to the relevant product, changes to the people running the relationship, changes in strategy that touch the overlap. It is a modest commitment and it prevents the most common category of avoidable grievance.
Where this leads
Governance decides whether an arrangement is run. Measurement decides whether anyone can tell if it is working.