Your CRM Tracks the Pipeline. It Can't Tell You What You Actually Made.

A Dubai brokerage can tell you, to the hour, how many leads it took last week, which portal sent them, and which agent sits top of the board. Ask the same brokerage how much of the commission it booked last quarter has actually reached the bank, and the precision disappears. You get an estimate, or a promise to check with finance. The pipeline is instrumented to the minute. The money is not. That gap is the most expensive thing in most brokerages, and almost nobody is looking at it.

For years the accepted answer — and I held it too — was that a brokerage rises or falls on how well it manages leads. Capture faster, route smarter, chase harder. That is where the budgets went, and where the conversation still goes. Dubai now has more real estate CRMs than a market this size needs: portal-integrated, WhatsApp-connected, RERA-form-ready, and inexpensive.

So lead management is, broadly, handled. Brokers are not short of a place to put a lead. What changed my view was spending time with the numbers behind the pipeline rather than the pipeline itself. The leads were captured. The deals were closing. And the owner still could not answer three questions that decide whether the business actually makes money: which clients are profitable, which developers are sitting on unpaid commission, and what cash will land over the next ninety days. The CRM had no opinion on any of them, because a CRM is a sales tool. It was never built to read money, and asking it to is how agencies end up running the most important part of the business on a spreadsheet.

The scale is what makes the blind spot costly. Dubai recorded AED 917 billion across more than 270,000 property transactions in 2025, up roughly 20% on the year, according to the Dubai Land Department. Brokerage commissions reached AED 13.59 billion across 96,440 broker-led transactions, earned by 32,294 registered brokers working out of 9,785 offices — most of them small agencies in the one-to-thirty-agent band.

That is an enormous amount of commission moving through a large number of small businesses, and the overwhelming majority measure it with a monthly spreadsheet export and a feel for how the quarter went. The bigger players have a finance function to lean on. The small agency — the bulk of those 9,785 offices — has the owner, a part-time accountant, and a deals tracker that lives in one person’s laptop. Three numbers go missing in that gap, and each of them changes how the business should be run.

Net margin by client. A CRM reports gross commission and deal counts. It does not step gross down to net: gross commission, less the 5% VAT due to the Federal Tax Authority, less the agent split, less the lead and advertising cost attributable to that client. Do that arithmetic across a year and the ranking usually inverts. Take a worked case. A repeat investor closes six deals on AED 612,000 of gross commission, costs almost nothing to service, and is sourced through reputation rather than paid leads. A one-off tenant generates AED 88,000 of gross on a single lease, sourced through a portal that charged for the enquiry, on a higher agent split. After the split and the acquisition cost, the investor’s net margin runs near 49%; the tenant’s lands closer to 21%. The brokerage feels both as “a closed deal”. Only one of them is worth chasing more of. Most agencies are busiest on their least profitable work and have no figure that tells them so.

Realisation rate by developer. This is the one that surprises people, and it is the sharpest. In off-plan, the developer pays the commission — and rarely on time. Commission is booked at the SPA and then waits, commonly 60 to 120 days, sometimes tied to the buyer’s payment milestones, sometimes simply slow. A portion is never collected at all. Yet I have not met a brokerage that tracks realisation: booked versus invoiced versus actually banked, aged 0–30, 31–60, 61–90 and 90-plus days, per developer. Without it, “booked” gets treated as “earned”, and the difference is working capital you have lent a developer interest-free. Picture two developers on your books. One pays inside 45 days at a 98% realisation rate. The other averages 96 days, leaves 30% of the commission ageing past 90 days, and writes off a slice every year. On the sales board they look identical — both move units. On a realisation view they are not remotely the same business partner, and that should change which inventory your agents push. You cannot act on a number you have never calculated, and right now that number is uncalculated almost everywhere.

Compliance and tax exposure. Trakheesi advertising permits are mandatory on every listing and every channel, including social, and fines start at AED 50,000 for a single non-compliant advert. Permit expiry, Form A authorisation, Ejari and Oqood status sit one lapse away from a penalty that erases a deal’s margin. Add the UAE’s 9% Corporate Tax on taxable income above AED 375,000, and the agency profit-and-loss now carries lines that no Dubai sales CRM was designed to model. These are financial questions wearing a compliance label, and they belong in the same place as the margin and the realisation figures, not in a separate folder that someone remembers to check.

So the answer is not another CRM. It is the financial layer that sits beside the one you already have. We build it on the Microsoft stack: Power Apps and Dataverse hold the records — Lead, Contact, Property, Deal, Developer, Commission and Payment, the seven entities the whole picture rests on. Microsoft Fabric and Power BI carry the analytics. Power Automate chases the receivables and flags a permit before it expires.

The starting point is deliberately ordinary — the deals tracker the brokerage already keeps in Excel, the developer commission terms from its agreements, and the agent split table. Three imports, and the history populates. Inside the first working session the owner sees what the spreadsheet never totalled: gross commission income cut by developer, agent, area and transaction type; net margin per client and per client type; and the realisation gap, with the slow developers flagged in figures rather than adjectives.

The realisation view is the one that earns the conversation. Each developer shows its booked commission, how much has been realised, the outstanding balance aged into thirty-day buckets, the realisation rate as a percentage, and the average days-to-pay. Fast, full payers sit in one corner; the laggards sit in another, with their unrealised exposure stated in cash. Alongside it runs a commission forecast — the weighted pipeline projected forward — and a cash-flow view that accounts for the developer payment lag, so the owner sees not only what should close but when the money actually arrives. A receivables chase fires automatically when a developer crosses 30, 60 and 90 days overdue, so the analytics turn into action rather than observation.

None of this replaces the sales pipeline. The CRM keeps doing the part it does well. The financial layer answers the questions it was never built to answer. Here is a screen-by-screen walk through of that layer — six views, illustrative figures, the structure I would build on a real estate brokerage’s own deal and payment data.

A piece that only described what works would not be worth your time, so here is what does not.

The hard part is the data, not the engineering. Brokers’ deal trackers are inconsistent, and developer commission terms are often handshake arrangements that shift per project and per promotion. A structured import template and a guided onboarding absorb most of that, but not all. Realisation analytics are only as good as the payment dates the brokerage recorded — so for some agencies the realisation view sharpens over the first few months rather than being complete on day one. Pretending otherwise sets a client up to distrust the first dashboard that disagrees with their memory.

Forcing marketing cost down to the individual deal is the other trap. Multi-touch attribution will stall the build, so margin is calculated with cost held at source and agency level first, refined to the deal later. And the forecast is forward-looking by nature: narrowing the reporting window shrinks the historical actuals, but it does not, and should not, shrink a pipeline that describes the future. There is a platform limit too — portal lead data in Dubai is gated, so the layer is built to work from the brokerage’s own deal and payment data first and to enrich from portal feeds where access exists, rather than depending on a feed that can be switched off.

For an owner, a general manager or an operations lead, this is not a reporting upgrade. It is the difference between managing on instinct and managing on the number. It tells you which clients to keep and which to reprice. It tells you which developers deserve your agents’ attention and which are financing themselves on your unpaid commission. It tells you what cash you can count on next quarter, after the payment lag, not before it.

Those are operating decisions, made every week, and most owners make them blind because the only tool on the desk was built to chase leads. Dubai’s brokerages do not lack data. They lack a single version of their own financials they can trust — and the commission they have earned but not yet collected is the clearest place that costs them.

We work the way the rest of our practice does: operations-first, fixed-scope, and accountable to one person rather than handed to juniors. The shape is four phases — Discover, Prototype, Deploy, Expand — and the promise on the first of them is specific. First value in six weeks, not a fifty-slide roadmap.

In Discover, we agree the data model and import your last two quarters of deals and developer terms. In Prototype, you get the working financial views — margin by client, realisation by developer, the commission forecast — built on your own numbers, not a demo dataset. Deploy adds the automation: the receivables chase, the permit-expiry alerts, the agent-versus-target tracking, and the VAT and Corporate Tax lines on the agency P&L. Expand is where AI earns its place — scoring which developer commissions are most likely to pay late, once there is enough of your own history to make the prediction honest rather than decorative.

Unify the data. Predict with it. Act on it. In that order, always.