Most distribution leaders know their top ten producers by name. They know which are warm, which are due a call, and roughly where next month's submissions are coming from. The top ten aren't the problem. The other forty, or sixty, or hundred are.
At that size the usual measure is volume. How many threads, how many submissions, how much premium last year. Volume is easy to get and it's what everyone reports, and on its own it tells you very little, because several quite different relationships produce the same volume number.
Here are five kinds worth telling apart. They're descriptions rather than a framework, and nobody needs to adopt the names.
The steady one
Sends you roughly the same thing every month. Three or four threads, mostly submissions, occasionally a renewal query. Has done for years.
Never a problem and never a priority, which is precisely the risk. A steady producer generates no urgency in anybody's inbox, so they get answered last on a busy day, every busy day, for a very long time. The first sign that a steady relationship has stopped being steady is usually that it's already gone.
The lumpy one
Nothing for six weeks. Then four submissions in a fortnight, all at once, all urgent.
This is a producer whose own book renews in clusters, often because of how they acquired it, which is not a thing anyone at your end has reason to know. Their quiet spell is a cycle rather than a decline, and it reads as a decline in every system that compares this month to last month.
This is the case for looking further back than four weeks. A month tells you they've gone quiet. Two months tells you they go quiet every year at about this time.
The one warming up
Two threads a month becomes four. Then five, with a first quote request in there.
The most valuable relationship on this list and the hardest one to see, because nothing about it is large. Going from two to four is a hundred percent increase and it looks like ordinary noise on a desk handling two hundred emails a week. Nobody has ever walked into a Monday meeting and said the Hendricks account went from two to four.
By the time a warming relationship is obvious it has usually stopped warming, because the broker's read of whether you're interested was formed some weeks earlier, in how fast the early ones came back.
The scattered one
Sends you three lines through three different underwriters, so appears as a middling relationship in three inboxes at once and as your busiest producer in none of them.
The mechanism gets its own article, because it's the one that costs most and the one people find hardest to believe until they've counted. The short version is that brokers email people rather than mailboxes, and the better the relationship the more people they email.
The busy one that isn't producing
High volume, all of it servicing. Endorsements, mid-term adjustments, queries about documents, the occasional chaser about a claim that isn't yours.
Feels like an engaged relationship. It's active, it's friendly, somebody there emails you most days. And it hasn't carried a new submission since March.
We can't reliably separate this one from a relationship that's simply mature, where the book is placed, it renews, and servicing traffic is what a healthy year looks like. Both produce the same shape. Anyone who tells you they can tell those apart from the outside is guessing.
Four firms a volume report would rank in roughly the wrong order. Each one is measured against its own history, never against the others.
What volume was hiding
Five relationships, and a volume report ranks all of them by a number that doesn't distinguish between them. The scattered one is under-counted. The lumpy one is mis-timed. The warming one is too small to register, and the busy one that isn't producing sits at the top of the table looking like your best account.
What separates them is movement against their own history rather than size against each other. Is this firm doing more than they normally do, less, or the same, and how long has that been true for. A firm sending two threads a month is not underperforming a firm sending twelve. They're different firms.
That comparison is straightforward arithmetic and nobody does it, because doing it means holding a normal for every producer in the book and checking this week against it. Sixty of those. Every week. It isn't a reasonable thing to ask of anyone with a technical workload, and it's the sort of dull repetitive comparison that software is actually good at, which is roughly the argument for what we build.
There's arguably a sixth kind, the producer who only ever emails one person at your end. Whether that's a relationship or a single point of failure depends entirely on whether that person is still there next year.
Something to check this week
Take your producer list, sorted by last year's premium, and look at the top fifteen. For each one, write down from memory whether they've sent you more or less than usual in the last month.
Then go and check three of them properly against the mail.
Ten minutes, and the useful part isn't the score. It's noticing which ones you had no view on at all, because those are the relationships being managed by whoever happens to open the email.