Which spend produced company dollar, and which produced agents?

A brokerage earns on the spread between what agents produce and what they keep. A single productive agent recruited this year may generate company dollar for five, which dwarfs the value of any individual client lead the same budget could have bought.

Yet recruiting spend is almost never attributed. The agent signs months after they first read something, the first closing is months after that, and by then nobody remembers which campaign started it.

Matching an agent's production back to the recruiting enquiry that produced them is a long-lag join on a person — exactly the case this product exists for, and one where the numbers are large enough that being roughly right is not good enough.

Brokerages routinely report sales volume, which is the agent's number rather than the firm's. Ranking channels on volume rewards whichever source produces the largest transactions, regardless of the split behind them.

Exporting company dollar instead ranks channels on what the brokerage actually kept. On a firm with a mix of splits and caps that can reverse the ordering entirely.

Recruiting is measured, where it is measured at all, by agents signed. That is a countable event with a date, it happens within a reasonable time of the campaign, and it makes a satisfying chart. It also says nothing about whether the brokerage made money, because an agent who joins, takes a desk, consumes onboarding time and training, closes two transactions and leaves for a competitor inside a year has cost the firm considerably more than the company dollar they generated. Signing them was not the achievement; keeping them would have been.

Retention differs sharply by where the agent came from, and every brokerage funds all of it from one line in the budget. An experienced producer poached from a rival office behaves differently from a newly licensed agent who answered a social campaign about a career change, and both behave differently again from somebody introduced by an existing agent who will feel personally responsible for them. Those are three distinct acquisition channels with three distinct survival curves and three distinct costs per productive year, and no report the firm currently runs distinguishes between them at all.

Adding a departure date to the agent roster makes the whole picture visible at once. Production is then ranked by the channel that recruited each agent across their entire tenure rather than at their first closing, so a source that reliably produces agents who stay two years is separated cleanly from one producing agents who stay four months and take their pipeline with them. That separation usually reorders the recruiting budget more decisively than any change to the creative, the offer or the split on the table, and it does so using dates the back office already records.

A brokerage with an affiliated mortgage desk, a title interest or an insurance arrangement earns from the same client more than once. The commission is the visible part and the only part marketing is ever shown. Behind it sit an origination fee on the loan, a share of the title and settlement work, sometimes a referral on the home warranty or the moving arrangement, and each of those lands in a different ledger, on a different date, under a different entity, reconciled by a different person. Nothing routinely brings them back together against the client who generated all of them.

Marketing therefore gets judged on the first line only, because the first line is the one sitting in the transaction system and the others are somewhere else entirely. That understates the return on every channel by whatever the attach rate happens to be, and the attach rate is not uniform across sources. A first-time buyer who needs financing arranged, title work done and insurance placed behaves very differently from a cash purchaser downsizing with their own solicitor, and the two arrive through completely different campaigns at completely different costs.

Exporting each revenue line as its own row against the same client contact detail solves it without any reconciliation work and without asking the affiliated businesses to change anything. Every line credits the channel that produced the client originally, so the report shows what a client from each source is worth across the whole enterprise rather than what their commission alone was worth to one department. For a brokerage arguing internally about which entity funds acquisition and at what level, that is precisely the number the argument turns on, and it has usually never been assembled.

Company-generated lead programmes are common and their economics are awkward in a way that rarely gets stated. The firm pays the vendor, distributes the enquiry to an agent on the rotation, and receives its split of whatever commission eventually results, sometimes at a referral rate higher than that agent's usual schedule and sometimes not. So the brokerage carries the entire cost up front and receives a portion of the benefit much later, which makes the return dependent on the split arrangement at least as much as on the quality of the leads themselves.

That means the same lead source can be comfortably profitable in one office and loss-making in another, purely because of where the agents working it sit on the commission schedule that year. A ranking built on gross commission produced will show the source as strong in both places and hand the finance director nothing he can act on. The problem is not that the data is wrong; it is that the column being ranked belongs to the agent rather than to the firm paying the invoice.

Exporting company dollar per transaction, with the agent named on the row, prices the programme properly for the first time. The report ranks sources on what the brokerage actually retained, and the agent column shows whether a disappointing source is genuinely weak or is simply being routed to agents who capped in August and have been keeping everything since. Those two diagnoses call for opposite decisions, one about the vendor and one about the rotation, and they are completely indistinguishable without that column present.