What a small brokerage's tech stack really costs

Ask a broker what their software costs and you will get a number that is somewhere between a third and a half of the real figure. Not because anyone is being careless — because the question is harder than it sounds, and the parts that are easy to count are the parts that cost the least.

This post is about how to count the whole thing. It is worth doing once a year, and it is worth doing before you evaluate anything new, because otherwise you are comparing a new tool's full price against an existing stack's partial one.

Start with the number you already have

Pull every recurring charge from the business account and the card, for twelve months. Not the ones you remember — the ones that actually posted.

Two things reliably turn up in this exercise:

Charges for tools nobody uses. Every small brokerage has at least one. A product bought for a specific problem that got solved another way, a seat for an agent who left, an annual plan that renewed past the point of usefulness. These are pure recovery: cancel them and the money is back this month.

Charges nobody can identify. A line item with a company name that means nothing to anyone in the office. Chase these down. Sometimes it is a tool under a brand name you did not recognise; occasionally it is a subscription started by someone who is no longer there.

This gives you the subscription total. It is the honest floor, and it is the smallest part of the answer.

Then count the per-seat multiplier, properly

Per-agent pricing is where a small brokerage's costs stop being predictable, and the trap is not the price — it is the direction.

Most per-seat tools bill on seats you have provisioned, not seats you are using. An agent who left in March is a charge in October unless someone removed them. In an office with normal turnover, this is a standing leak, and it is invisible precisely because the total only moves a little each month.

Count it two ways and compare: what you are paying for, and how many licensed agents are actually working. If those numbers differ, you have found real money before you have evaluated a single alternative.

The second question is what growth does. Going from ten agents to fifteen is half as many people again — and on a fully per-seat stack it is half as much software cost again, on every tool at once. That is worth knowing before you recruit, not after.

Now the part that is not on any invoice

Here is where the real number lives.

Double entry

Count the number of times the same fact gets typed into two systems. A new listing that goes into the transaction platform and the CRM and the website. A closed deal that gets recorded in the transaction system and again in the accounting system. A new agent set up in five places.

Then estimate, honestly, how long each one takes and how often it happens. You do not need precision — you need the order of magnitude. Most small offices find this is somewhere between two and six hours a week of someone's time, and that someone is usually the office manager, who has other work.

This is a real cost. It is just paid in hours rather than dollars, which is why it never appears in a stack review.

The reconciliation tax

Worse than double entry is double entry that drifts. Two systems hold the same fact, they disagree, and somebody has to work out which one is right.

The cost here is not the fixing. It is the doubt. Once the office has learned that the commission number in one system does not always match the other, every number gets checked, and checking is slower than trusting. An office that trusts its own numbers moves noticeably faster than one that does not, and the difference has nothing to do with the software's features.

Integration work

Anything you paid a person to connect. A one-time setup fee, a consultant, a contractor who built a sync, a plugin with its own subscription. And — this is the part people forget — the maintenance. Integrations break when one side changes, and someone has to notice and fix it.

Count both: what you paid to build it, and what it has cost to keep alive.

Onboarding and training

Every tool has a cost per new agent, paid in the office manager's time. Setting up accounts, walking through the basics, answering the same questions. Multiply by your actual turnover, not your ideal turnover.

A stack of six tools has six of these. This is one of the clearest arguments for consolidation, and it is almost never the argument people make.

Switching costs you have already paid

Look back at the last tool you changed. How long was the office slower? How much data came across badly? What never came across at all?

This is a sunk cost, so it does not belong in the annual total — but it belongs in your estimate of what the next change will cost, and most brokerages estimate that number far too low because they remember the decision and not the month after it.

Put it together

You now have five numbers: subscriptions, per-seat drift, hours lost to double entry and reconciliation, integration build and maintenance, and per-agent onboarding time.

Convert the hours at a rate that reflects what the person doing them actually costs the business. Then add it up.

The total will be substantially larger than the invoice total. That is not an argument for buying anything — it is an argument for knowing the figure, because every decision after this one depends on it.

What to do with the number

Three things, in order.

Cancel what is unused. Immediate, free, and it usually pays for a meaningful fraction of whatever you do next.

Fix the seat count. Also immediate. Set a rule that offboarding includes removing seats, and put it on the same checklist as collecting keys.

Then, and only then, look at consolidation. Fewer systems means less double entry, fewer integrations, less onboarding and fewer places for a number to disagree with itself. It also means a real switching cost, paid up front, and a genuine risk of being worse at one thing in exchange for being better overall.

That trade is a real decision with real downside, and you cannot make it well against a partial cost figure — which is the whole reason to count properly first.

If you want to see what a consolidated version of this looks like before you price anything, our pricing page lays out how Patio is structured, and the demo below shows the part that is hardest to evaluate from a feature list.