How to Start a Prop Firm: The Operator’s Playbook

Three things gate a prop firm launch, and only one of them is technology. Most founders working out how to start a prop firm begin with the platform, because the platform is the part they can picture. It is also the part that takes the least time.

Here is what has to be true before you take a single payment. A legal entity that can contract with traders and appoint providers. A payment provider that has underwritten your business and will process evaluation fees. A rulebook written precisely enough that software can enforce it without an argument. The rest of the build sits behind those three, and none of them is bought from a vendor in an afternoon.

We build the technology and operations side of prop firms, and our sister company City Traders Imperium is a live prop firm under the same owner. So this is written from both chairs: the vendor shipping the platform, and the operator who has to explain a margin rule to a trader who has already passed. Below is the sequence, with a dependency and an owner for every stage, and the honest split between what runs on our clock and what runs on somebody else’s.

How to start a prop firm: the order the work actually happens in

Run it as two tracks, not one list.

Track one is the technology. Programs, portals, rules engine, risk, payouts, integrations. That work is sequenced and it moves at a speed a vendor controls.

Track two is yours. The entity, the legal paperwork, the payment provider. Those move at the speed of a registry, a law firm and an underwriting desk. No vendor controls them, and any vendor who implies otherwise has never launched a firm.

Founders who launch late are almost never late on technology. They are late because track two started after they signed with a technology vendor, when it should have started first. Read the table as a dependency map rather than a calendar.

Stage	What has to be true	Who owns it	Depends on
Model decision	Programs, prices, target markets and payout policy stated on one page	Founder	Nothing. This gates everything else
Legal entity	A company that can contract with traders, hold funds and appoint providers	Founder, with local counsel	Model decision, which sets the jurisdiction
Legal paperwork	Terms, trader agreement and policies match the rules you will enforce	Founder’s counsel, from your rulebook	Entity and rulebook
Rulebook	Every target, drawdown and prohibited strategy rule written so software can decide it	Founder, with risk input	Model decision
Payment provider	An underwritten merchant account in your entity’s name	Founder	Entity, plus every document the underwriter names
Platform and portals	Trader Portal, Admin Portal and the platform connection live, under your brand	Prop FinTech	Rulebook
Rules and risk engine	Evaluation logic, breach detection and payout gate running automatically	Prop FinTech	Rulebook
Integrations	Platform, payments, KYC and attribution connected, then tested	Prop FinTech, with your credentials	Provider and KYC accounts existing
Payouts Management	Banking details stored, approval rules set, exception path defined	Your team or Trade Guard	Payment provider, rulebook
Operations cover	Named cover for risk, support, payouts, KYC and disputes across your trading week	Founder decision	Everything above
Demand	Tracked campaigns and attribution that reach the funded trader	Founder	Site, checkout and payments live

Two rows in that table decide your launch date: the payment provider and the rulebook. Everything technical depends on the rulebook. Everything commercial depends on the provider. Start both in your first week and the rest compresses around them.

What ten days covers, and what it does not

Our claim is ten days, and the boundary matters more than the number. Ten days is everything related to us: the technology side. Anything that is a legal setup or a payment provider is third party, and outside our control.

Inside the ten days:

• Trader Portal and Admin Portal live on your own domain, under your own brand
• Programs configured against your rulebook: targets, drawdown model, consistency and prohibited strategies
• The rules and risk engine that decides breaches automatically, instead of a person reading equity curves
• Payouts Management: stored banking details, approval workflow, escalation path
• Integrations: trading platform, payment provider, KYC and attribution connected, then tested with real flows
• Affiliate Portal and Automation: referral tracking, commissions, onboarding triggers and reporting
• Your team trained on the Admin Portal, monitoring running before the first trader registers

Outside the ten days, and outside our hands:

• Incorporation, and whatever your registry asks for on the way
• Terms, trader agreement and disclosures signed off by counsel in your jurisdiction. Our Legal Paper Work service guides that paperwork. It cannot make a law firm move faster
• Merchant account underwriting, which is the long one
• A KYC provider account in your entity’s name

Ten days is our build. The entity, the paperwork and the payment provider are your homework, and they run on other people’s clocks.

Say that on your first call with any technology vendor and watch the answer. Anyone selling a ten day launch that includes underwriting does not own that timeline, and you find out in week three with a finished platform and no way to take a payment.

What the launch actually costs, line by line

There is no single honest number for starting a prop firm. The range is too wide to be useful, and most figures published on the question are marketing rather than accounting. What is useful is the list of lines, what moves each one, and which ones founders get wrong.

Government fees are the predictable part, because they are published. Hong Kong charges HK$1,545 to incorporate a local private company with share capital by electronic filing, or HK$1,720 on paper, and a one year business registration certificate is HK$2,350 including the levy, for certificates commencing between 1 April 2026 and 31 March 2027. Cheapest and most certain lines on the list. Your counsel’s invoice for the terms and the trader agreement is published nowhere, and it is the larger number.

The lines, roughly in the order you commit to them:

• Legal paperwork. Terms, trader agreement, disclosures, privacy and refund policy. Driven by how unusual your rules are. A standard two phase model is cheaper to paper than a scaling plan with a bespoke payout split.
• Technology. Platform and portals, rules engine, risk, payouts, affiliate and automation, plus the trading terminal your traders already know and the market data behind it. Driven mostly by whether you take a configured product or commission custom work and a migration.
• Payment processing. Setup or underwriting fee, per transaction rates, and a rolling reserve held against refunds. Driven by risk classification far more than by volume.
• KYC and AML checks. Priced per verification, so this line scales with marketing rather than with revenue. A campaign that brings unqualified buyers still bills you for every check.
• Payout capital. The money that leaves your account when traders withdraw. Not a launch cost, a standing one, and the line that decides whether a good month for your traders is survivable.
• Acquisition. Creative, media spend, and the tracking that ties a sale back to the ad behind it. Judge it on reporting that follows the buyer to a funded trader, because cost per lead flatters you and cost per funded trader does not.
• Operations cover. The hours somebody has to be awake for, and the line most first time founders leave at zero.

On the rate itself, mainstream published pricing is the wrong baseline but a useful one to know. Stripe publishes 1.5% plus 20p for standard UK cards, 2.5% plus 20p for cards issued in the EEA and 3.15% plus 20p for international cards, with a further 2% where a currency conversion is involved. Across the firms we work with, prop firms rarely sit inside pricing like that. Evaluation fees get classified as high risk, so budget for the category rather than the headline.

Four lines get underestimated. The rolling reserve, your own cash sitting with a provider for months. Dispute handling, a process and a person rather than a fee. A second payment route, which looks like duplicate spend until the morning your first route pauses. And the cost of rewriting rules after launch, paid in trader relationships rather than money.

Payments are the real critical path

Getting a merchant account can take as little as one week, and it can run on for a month or two depending on which provider a firm goes with. That range is our own experience across the firms we work with, and it is the widest variable in any launch we have been part of. In most launches it sets the date. Not the technology.

The spread has causes, and most of them are yours to fix:

• A brand new entity with no processing history. Underwriters price uncertainty, and pricing uncertainty takes time
• A model the desk has not seen. Some understand evaluation fees immediately. Others read the word trader and stop reading
• Incomplete documents. Every round trip for one more certificate costs days, and you rejoin the queue behind other applicants
• A jurisdiction mismatch, with the entity in one place, the directors in another and the target market in a third
• Volume forecasts you cannot support. Underwriters compare the forecast to your history, and a number with nothing behind it invites more questions
• Applying to one provider and waiting for an answer before applying to a second

Four things shorten it. Apply before you commission the technology. Apply to two providers in parallel and accept paying for redundancy you hope not to need. Ask the underwriter at the start for the full document list, then send it complete rather than in instalments. Give a forecast you can defend, and take the reserve terms rather than negotiating for weeks over a percentage.

The consequence of treating this as an afterthought is not a delayed launch. It is a launch with one payment route, which puts a single point of failure between your marketing spend and your revenue. Routes pause. Reviews happen. When yours does, a firm with one route stops selling that morning, and the ads keep spending.

Writing the evaluation rules so they survive contact

Two rules cause most of the disputes in this industry: one sided betting and margin rules. Write both to be machine enforceable before you launch, because rewriting them afterwards means renegotiating with traders who already passed under the old wording.

One sided betting, in practice, is a position that can only pay in one direction. Hedged exposure split across two accounts so one of them is guaranteed to hit the target. Grid or martingale structures that survive on the size of the next position rather than a read of the market. Risk parked ahead of a scheduled release with the offsetting leg somewhere your rules do not look. None of it demonstrates the skill your evaluation claims to measure. If your rulebook simply says prohibited strategies are not permitted, you have written an opinion rather than a rule, and you will be arguing that opinion with a trader who has read it more carefully than you have.

Margin rules generate disputes for two reasons, both fixable in the drafting. The first is ambiguity at the moment of breach. Is drawdown measured on balance or equity. Intraday or at the close. Against the starting balance, the trailing high or the highest closed balance. What happens to floating loss in the seconds before the platform reports it. Every one of those is a fork, and a rulebook that leaves any of them open has two readings.

The second is that enforcement happens after the fact. A breach is noticed, reviewed, then acted on, and the trader stopped just past the limit will argue about the margin of the breach. If the wording allows two readings, that trader is right to argue.

One test, before launch. Hand the rulebook to somebody who has never traded, give them five borderline cases, and ask them to decide each using nothing but the document. If they cannot, your software cannot either, and the gap gets filled by whichever support agent is on shift, differently each time.

Detection is the other half of the same problem. A rule you cannot see being broken is not a rule you have. One sided betting rarely shows inside a single account, which is why it is caught by comparing entry timing, direction, size, device and payment details across the whole book rather than by watching one equity curve. And a flag decides nothing on its own. Someone has to read the evidence, make the call and defend it to the trader waiting on a payout, which is the work a trained operations desk does.

The cost of getting this wrong arrives late. You run two rulebooks at once, grandfather the traders who passed under the old one, explain the difference in every ticket, and watch a screenshot of your own terms circulate with the ambiguous clause highlighted. No platform fixes that.

Covering the roles from day one, not month three

Day one staffing is a shift coverage problem, not an organisation chart. That is why founders under hire: they count job titles, when the load is measured in hours that have to be covered.

Five functions need cover from the day your first trader registers:

• Risk monitoring, across the trading week. The forex week runs from Sunday evening to Friday evening and your traders will use most of it. Someone has to see a breach or a platform fault while it is happening
• Trader support, weighted to your traders’ hours rather than yours. Buyers from a dozen countries do not queue politely inside one office day
• Payout processing, on a schedule you can hold. Slow payouts are the fastest way to lose a trading community, and the queue does not care that it is Saturday
• KYC and onboarding review, at whatever speed marketing delivers buyers. A campaign that works creates a verification backlog the same day
• Dispute ownership, with one named person who can decide. Not a committee, and not the founder by default at one in the morning

Our recommendation, from watching both sides of this: cover risk monitoring and payouts first, because both fail loudly and expensively. Support and KYC can start thin and scale with volume, as long as you are honest with traders about response times. Disputes need an owner from day one even at near zero volume, because the first dispute handled badly is the one that gets screenshotted.

None of that is a headcount claim. It is a coverage requirement, and you meet it by hiring and rostering, or by buying it as a managed operational team for prop firms: trader support, withdrawal validation, behaviour review and compliance monitoring, run by agents trained on prop firm operations and billed per agent.

The hire that quietly breaks a firm

The most expensive mistake in this business is hiring the wrong people to run your firm. Not the wrong platform. The wrong people.

It is expensive because it does not fail visibly. A bad developer ships a bug and you see the bug. A bad risk or operations hire looks fine for weeks. The queue moves. Tickets close. Nothing is on fire. Then the damage arrives at once, in the form your traders notice before you do: payouts approved that should have been held, holds applied that should have been paid, rule explanations that contradict last month’s answer, and a copy trading group that ran for weeks because nobody read the pattern behind the flag.

By the time any of that reaches your dispute log, it has already happened to your traders. You can refund money. You cannot refund the belief that your firm is fair, and that belief is the product you are actually selling.

So there are three functions you should never staff on a guess. Withdrawal validation. Dispute handling. Fraud review. Each needs judgement, each is exercised on your traders in public, and none can be properly supervised by a founder who is also selling, hiring and building.

Which leads somewhere uncomfortable. Early on, buying a staffed operations team is often lower risk than hiring your own. Not automatically cheaper. Lower risk. It arrives trained on prop firm workflows, with a process somebody else has already debugged, and if it is not working you change the arrangement rather than manage a person out of a role only they understand. That is the case for buying the operations desk at the start, and the same case in reverse for hiring properly later, once you know your volumes and can write a job description from evidence instead of hope.

The first ninety days: demand and integrity

Survival in the first ninety days comes down to two engines running at once: strong marketing, and backend compliance protection strong enough to catch cheaters. One without the other fails, in opposite directions.

Demand is the first engine. Prop firm acquisition has a compliance dimension that most performance marketers meet for the first time on your account, and a banned ad account costs more than a bad week of return on spend, so paid campaigns built for prop firm economics is a different job from paid campaigns generally. Measure it on behaviour analytics that follow the funded trader rather than on clicks. Clicks are not the product. Neither are leads. A buyer who funds, trades within your rules and comes back for a second program is.

Integrity is the second, and it has to start on the same day rather than the first time you get burned. A firm that markets well and cannot detect coordinated abuse pays out to cheaters, and it does that fastest exactly when marketing is working, because volume is where patterns hide. Continuous account scoring with the evidence attached before a payout clears is what stops a good month from becoming an expensive one.

Three ways firms get this wrong, in order of how often we see it:

• Both engines built in month four, after the first bad payout and the first flat week. Common, and avoidable
• Marketing with no integrity layer. You pay organised groups, and your bank balance tells you before your dashboard does
• Integrity with no demand. Enforcement is perfect, there is nothing much to enforce, and the runway goes anyway

A short list to hold yourself to at ninety days. One tracked acquisition channel measured through to funded traders. Account scoring live before your first payout window opens. A payout schedule you have met every single time. A rulebook you have not quietly changed under pressure. One named owner for disputes, with a written escalation path behind them.

If you want this sequence priced against your own model, book a scoping call and bring your jurisdiction, your draft rulebook and your payment situation. Whatever you decide about a technology partner, start the entity and the merchant account application this week. Those are the two clocks you do not control, and the firms that launch when they said they would are the ones that started them first.