Commission Reconstruction: Why a Brokerage Runs on Four Systems, and What the Fourth One Costs.
Every commercial brokerage closes a deal twice. The first closing is the one everybody celebrates: the lease is signed or the sale records, the client is happy, the broker sends the email. The second closing happens a week or two later in the back office, when somebody has to work out exactly who gets paid what, and discovers that the deal does not exist as one thing anywhere in the company. It exists as four things, in four systems, and the job is to rebuild it.
That rebuild has a name inside most brokerages, even if nobody writes it down. We call it commission reconstruction, and it is the single most expensive routine task in a brokerage's operations that never appears on a budget line. This post is about why it happens, what it costs, and what actually ends it.
The four systems.
Walk through where a deal lives from first contact to payout and you find the same four records almost everywhere.
The CRM holds the relationship. The property, the contacts, the activity, the stage. If the brokerage has done the work to model a deal with two sides, the listing side and the procuring side, it lives here. Often it does not, and the CRM holds one deal with one broker on it because that is what a generic CRM ships with.
The listing or marketing platform holds the property. Square footage, asking rent, the flyer, the tour history, the offers. Some of this is in the CRM too, re-keyed.
The commission spreadsheet holds the money. The gross fee, the co-broke, the referral, the house split, the broker splits, the deductions for marketing spend, the payout timing. This is where the deal is actually defined financially, and it is usually the property of one person in operations.
Accounting holds the cash. The invoice to the client, the receipt, the payable to each broker, and the entries that make the month close.
None of these four is wrong about the part it owns. The problem is that nothing forces them to agree until the money has to move, and at that moment someone opens all four and starts reconciling by hand.
What the reconstruction actually involves.
Commission is not one number. It is a chain of decisions applied in order, and every link is a place the four records can disagree.
Start with the gross fee, which the spreadsheet says is one figure and the signed agreement, filed in the CRM or a drive, may say is another after a late concession. Then the co-broke: was there an outside broker, at what percentage, and is that recorded against the deal or only in an email? Then the referral fee, which someone promised in a phone call in March. Then the house split, which changes with the broker's annual production, and production is a rolling total nobody maintains outside the spreadsheet. Then the split between two internal brokers who worked the two sides, which is where the arguments start, because the CRM only shows one of them. Then the deductions: the marketing package charged back, the sign, the photography. Then timing: the fee is paid in two instalments and the second is contingent on occupancy, so half of every calculation above has to be done twice.
When the deal is one record with two sides and a commission plan attached, this chain runs in seconds. When it is four records, a person runs it, and a person runs it after the fact with whatever evidence survived.
What the fourth system costs.
The cost is not the spreadsheet. The cost is what the spreadsheet forces everyone else to do.
Delay. Payouts wait for reconstruction. In the brokerages we have worked with, the gap between a deal closing and a broker being paid correctly runs from two weeks to a quarter, and the variance is the reconstruction queue, not the bank.
Disputes. Every reconstruction is an argument waiting to happen, because the evidence is spread across systems that were never designed to be evidence. A broker remembers a verbal split; the spreadsheet has the default; the CRM has neither. The managing broker ends up adjudicating from memory.
Errors that only surface at tax time. A referral fee paid but never recorded against the deal; a house split applied at last year's tier; a co-broke paid out of the wrong entity. Each one is small. Each one is found late, by the accountant, when it is expensive to fix.
A person you cannot lose. The spreadsheet has an owner, and that owner is the only person who can close the second closing. Their vacation is a backlog. Their resignation is a crisis.
No management view. Because the money is defined in a spreadsheet, the questions a managing broker most wants answered, which brokers are profitable after splits and chargebacks, which property types carry the fee, what is owed and when, cannot be asked of any system. They are asked of the person.
Why a fifth system is not the answer.
The instinct is to buy commission software. There are good products, and for a brokerage that only has a spreadsheet, one of them is a step up. But commission software bought as a fifth system inherits the problem it was bought to solve: it needs the deal fed into it, with both sides, the plan and the concessions, and that feed is the reconstruction again, done earlier and by a different person. The fee engine is not the hard part. Getting one true deal into it is.
We have made this argument before about commission splits and payouts for CRE brokerages, and it holds for every version of the fifth-system idea. A tool that computes commission from a record the company does not have will compute the wrong commission, quickly.
What ends it: one deal record with two sides and a plan attached.
The fix is a data model, not a purchase.
One deal, two sides. The deal record carries the listing side and the procuring side as first-class parts of the same record, each with its broker, its share and its source. A deal a brokerage works from both ends is one record with both ends, not two deals or one deal with a note.
The commission plan is applied when the deal is created, not when it closes. Gross fee, co-broke, referral, house tier, internal split, deductions and instalments are fields on the deal, defaulted from the plan and edited when a concession is made, with the edit logged. By closing, the payout is already computed and has been visible to everyone on the deal for the life of the deal. There is nothing to reconstruct because nothing was ever taken apart.
Accounting posts from the deal. The invoice, the receipts and each broker's payable are generated from the deal's own figures. The deal is the system of record for the money as well as the relationship, and the ledger is downstream of it.
The property lives once. The listing data the marketing platform needs comes from the deal's property record, not a re-keyed copy, so the flyer and the CRM cannot drift.
We built exactly this for a commercial brokerage that came to us for a CRM cleanup and left with a solution; that story is here, and the model is described in more depth in our piece on commercial real estate CRM development. The short version is that once the deal was one record, the commission spreadsheet emptied itself. Nobody was asked to give it up. It simply had nothing left to reconstruct.
How to tell whether this is your problem.
Three questions, answerable in an afternoon.
Pick the last five closed deals and ask how many days passed between closing and the final broker payout. If the number varies by more than a week between deals, the variance is reconstruction.
Ask who could compute the commission on a live deal today, without the spreadsheet's owner. If the answer is nobody, the spreadsheet is a system of record wearing a disguise.
Ask whether the CRM shows both sides of a two-sided deal. If it shows one, every internal split in the company is being decided outside the record.
If two of the three come back badly, the second closing is costing you more than the software you keep not buying, and the fix is an integration of the systems you already own rather than a new one.
What we do about it.
We start with discovery you pay for only if you go ahead, and we come out of it with a data model for the deal, a commission plan mapped from the way your splits actually work, and a guaranteed estimate for building it into your CRM and connecting it to your books. If we estimate low, we absorb the difference. Retainer clients see the full project plan for each release before committing to it. That is the commercial shape of the same idea the post is about: one record, agreed up front, that nobody has to reconstruct later.
Find out where your deals, listings and commissions fall between systems.
In a no-risk discovery we look at how a deal moves from listing to commission across your CRM and accounting today and show what one connected system would change. You pay only if you proceed. Or see how we approach it.
More on the same problem:
Comments