We lost a client we had held for three years, at a point when the campaign metrics were the best they had been. Nobody on our side saw it coming, which was itself the problem.
They agreed to a proper exit conversation, which not every client does, and what came out of it changed how we run accounts.
They were not measuring what we were measuring
The central finding.
We reported on cost per acquisition, reach, engagement and channel efficiency. Every one of those had improved over the period.
Their board was looking at revenue growth and market share, neither of which had moved much.
So every month we presented improvement and they heard a report about activity that was not producing the outcome they cared about. Over three years that accumulated into a settled belief that we were optimising a machine that was pointed in the wrong direction.
They were probably right, and nobody had told us, because our monthly meetings were with a marketing manager who was measured on the same things we were.
We never spoke to anybody who could cancel us
The structural error underneath it.
Our relationship was entirely with the marketing team. Good relationship, genuine mutual respect, and none of them made the decision to renew.
The person who did had seen us present twice in three years and had no independent sense of what we contributed.
When the budget review came, we were a line item defended by somebody junior to the person cutting it.
The lesson is not about politics exactly. It is that if you cannot articulate your value to the person who decides, somebody else will articulate it for you and they may not do it well.
We had stopped proposing anything
The finding that stung most.
In the first year we brought ideas constantly, several of which were rejected and two of which worked well.
By year three we were executing an agreed plan competently and proposing refinements to it.
From the client's side that read as a supplier who had run out of things to say. They described it, fairly, as us having become an execution resource.
What had happened internally is that the account had been handed to a team who inherited a working plan and had no mandate to question it, and the senior people who might have had been moved to newer accounts.
Which is the standard shape of how agency relationships decay and it is entirely predictable if anybody is looking.
The reporting had become ritual
A smaller point that they raised and I have thought about since.
Our monthly report ran to about forty slides. It was thorough and it was largely the same forty slides with different numbers.
Nobody read it. The meeting consisted of us presenting it and them waiting for the end.
What they wanted, and had never asked for because it seemed rude, was three slides and a discussion about what to do next.
We had confused thoroughness with value, and the volume of reporting was partly a way of demonstrating effort. They experienced it as an hour of their month.
What we changed afterwards
Five things, applied across every account.
We now agree at the outset what business outcome the work is meant to affect, and report against that alongside the channel metrics, even when it is uncomfortable because the business outcome has not moved.
We insist on a relationship with somebody senior enough to make the renewal decision, with at least two conversations a year that are not a performance report.
Every account has a standing item for a proposal — something we think they should do that they have not asked for — with the explicit expectation that most will be declined.
Reporting was cut to a page and a conversation.
And we run an annual review where we ask directly what we would need to do to be indispensable, which produces awkward answers and is the single most useful thing on this list.
The uncomfortable summary
We lost the account for doing exactly what we had been asked to do, well, for three years.
The mistake was never checking whether what we had been asked to do was still the right thing, and never speaking to anybody positioned to tell us it was not.
That is a failure of curiosity rather than of competence, and it is considerably harder to notice.
The warning signs, in order
Having gone back through the three years, the indicators were all present and none was acted on.
The meetings got shorter and more people stopped attending.
Questions became administrative rather than strategic.
Requests started arriving as instructions rather than as discussions.
The senior sponsor stopped appearing entirely.
And nobody asked us what we thought about anything for about eight months.
Any one of those is unremarkable. All five together, in sequence, is a relationship that has already ended and is waiting for the renewal date.
Whether we could have saved it
Probably, if we had noticed a year earlier, and only by changing what we were doing rather than by defending it.
By the time we were told, the decision had been made and the conversation was a courtesy. That courtesy was worth more than the account, which is a strange thing to conclude and I think it is true.