Setting Up a Dealing Desk for a New Broker: The First 90 Days
2 November 2026 · 9 min read
A new broker needs dealing coverage from the day the first live account trades, not from the day the first incident happens. The good news: a day-one desk does not need to be large. It needs correct platform configuration, a short list of written risk controls, defined coverage hours, and someone accountable for watching the book. This article lays out what to configure, what to document, and how to decide between building, outsourcing or mixing the two.
If you are earlier in the journey, our broker launch dealing and risk setup page covers the full launch scope. This article focuses on the desk itself.
When a new broker needs one
The trigger is not headcount or volume; it is exposure. The moment client positions can move your P&L, someone needs to be watching: net open position by symbol, margin utilisation across the account base, and the routing between your books and LPs. For most new brokers that moment is launch day.
What changes with scale is the coverage model, not the need. A broker with fifty live accounts needs monitoring, written limits and an escalation path, which one competent person can provide during one session. A broker with five thousand accounts across sessions needs a roster, tooling and procedures that survive individuals. Planning the second model while running the first is the mistake to avoid: coverage holes appear exactly when growth outpaces the informal setup.
There is a common belief that a new broker can defer the desk because the platform automates everything. The platform automates execution and margin maths; it does not decide your routing policy, notice a coordinated account cluster, or tell you that a symbol's configuration no longer matches your intent. Those are desk functions, and at launch scale they are part-time functions, but they are not optional ones.
Day-one configuration
Before the first live trade, the platform needs a deliberate configuration, not defaults. On MT4 and MT5 that means account groups with considered leverage and margin settings, symbol lists with correct sessions, markups and swaps, and routing profiles matched to your intended book model. The same discipline applies on cTrader, Match-Trader, TradeLocker and DXtrade: every default you accept is a risk decision you did not make.
The bridge deserves equal attention: LP connections tested, routing rules documented, failover paths verified rather than assumed. Configuration done in a hurry at launch has a way of becoming permanent, which is why platform administration is worth treating as a function from day one rather than a task someone does between other jobs.
A practical tip for launch week: export the full configuration, platform and bridge, and save it as the approved baseline. Every later audit then has something to diff against, and the question of when a setting changed has an answer. Brokers that skip this step spend their first audit reconstructing what the original intent was, usually from the memory of whoever did the setup, if that person is still around.
Finally, test the failure paths before live flow does it for you. Disable a feed briefly in a quiet window and confirm the alerts fire. Simulate an LP disconnect and confirm the failover route actually carries flow. These tests take an afternoon and feel paranoid until the first real incident, at which point they feel like the cheapest insurance you ever bought.
Minimum risk controls before launch
The day-one control set is short and non-negotiable:
- Written exposure limits per symbol and in aggregate, with a defined action when each is breached
- A routing policy: which flow is internalised, which is hedged, and who can change the classification
- Margin and stop-out monitoring with an escalation path that names actual people
- An event calendar process: who checks the macro calendar, and what changes before high-impact releases
- Daily reporting, even if the audience is two people: exposure, actions taken, anything unusual
- Access hygiene: named manager credentials, no shared logins, and a record of who can change what
None of this requires a large team. It requires that the controls exist in writing before they are needed, because the first time a limit is breached is the wrong time to decide what the response should be.
Keep the control set deliberately small at launch. Six controls that are actually monitored beat twenty that exist only in a document nobody opens. You can, and should, add controls as the book grows and as incidents teach you where the gaps are. The launch set has one job: make sure the first surprise is a written report, not a P&L shock.
Build, outsource or hybrid
Building in-house gives maximum control at maximum fixed cost: illustratively, a genuine 24/5 rotation requires four to five dealers plus tooling and management, which is a substantial payroll before the first trade. Outsourcing converts that to a scope-based retainer, commonly 20 to 40 percent of the in-house equivalent, with coverage starting in weeks. The full cost comparison walks through the arithmetic.
The hybrid pattern is common among new brokers: an outsourced desk covers all sessions at launch, and an internal hire grows into a head-of-dealing role who owns policy and manages the provider. That gives you institutional knowledge in-house and continuous coverage without a payroll cliff. Whatever you choose, decide before launch; the worst coverage model is the one improvised after the first incident.
Budget honestly for whichever model you pick. The in-house line items are salaries, recruitment, tooling and management time, and they arrive whether or not the book grows. The outsourced line item is a retainer sized to your scope, plus the onboarding effort of documenting your procedures. Illustratively, a new broker's outsourced coverage at launch scale runs a small fraction of one dealer's fully loaded cost, which is why the build decision usually waits until the book justifies a full rotation.
First 90 days checklist
By day 90 the desk should run on documents rather than memory: written limits, written procedures, written reports. That is the foundation every later decision, more coverage, more tooling, more headcount, builds on.
A useful day-90 test: could a competent dealer from outside your firm read your documentation and run the book for a week without asking questions? If yes, the desk is an operation. If no, it is still a person, and your next investment should be documentation before anything else, because every other improvement depends on it.
- Days 1 to 14: verify live configuration against the intended setup; confirm limits, routing and escalation work in practice with small real exposure, not just on paper
- Days 15 to 30: establish the daily reporting rhythm; run the first configuration audit; document every procedure that currently lives in someone's head
- Days 31 to 60: review routing classifications against actual client behaviour; test the escalation chain with a drill; baseline execution quality per LP and symbol
- Days 61 to 90: audit bridge configuration for drift; review limits against observed volatility; formalise the change log so every platform and bridge change has a reason and an approver
- Throughout: track coverage gaps honestly. Every session nobody watched is a data point about your real risk, and the list of gaps is the business case for your next coverage decision
Frequently asked questions
Can we launch without any dealing coverage?
You can, and some brokers do, but it means accepting that exposure breaches, margin cascades and configuration errors will be discovered after the fact. Even a minimal setup, written limits, one session of monitoring and an escalation path, is dramatically better than none, and costs little.
What is the cheapest credible day-one setup?
A correctly configured platform, a short written control set, and outsourced monitoring sized to your book. Illustratively, an outsourced retainer at launch scale runs a fraction of one dealer's salary, and Deltar engagements start with a fixed-scope Health Check at $1,000 that documents your configuration and coverage gaps.
When should we make the first internal dealing hire?
When the role is policy and oversight rather than coverage: someone to own the risk document, manage providers and make discretionary calls. Hiring a single dealer for coverage first usually creates a single point of failure without removing the need for external coverage of the other sessions.
Which platforms does this apply to?
All of them. The examples here use MT4 and MT5 because that is where most new brokers launch, but the same day-one discipline applies on cTrader, Match-Trader, TradeLocker and DXtrade: deliberate configuration, written limits, defined coverage and daily reporting.
What is the most common launch mistake?
Accepting default configuration and planning to fix it later. Defaults become permanent under launch pressure, and the audit that finally reviews them usually happens after an incident. The second most common is coverage that exists on paper but not at 3am.
Deltar sets up dealing operations for new brokers: launch-scope dealing and risk setup, ongoing platform administration and the outsourced dealing desk that covers every session from day one. Start with a Health Check of your launch configuration.