Gross deposits tell you that money arrived. They do not tell you whether the brokerage made money.
A client can deposit, collect a bonus, trade briefly, withdraw, create support work, and never return. Another client may start with a smaller deposit, fund again three times, trade consistently, and create little operational risk. Treating both as one funded account is how a healthy-looking dashboard hides weak economics.
The useful question is not: how much did this month’s clients deposit? It is: what value did this group of clients keep after the costs and risks required to acquire, serve, process, and retain them?
That group is a cohort. Once a brokerage starts looking at cohorts, it stops arguing about vanity metrics and starts seeing where the business actually earns or leaks money.
Quick Summary
- Deposits are client inflows. They are not a clean proxy for broker revenue or profit.
- A useful cohort groups clients by when and how they entered: month, source, GEO, payment method, campaign, or product.
- Track retained net value after acquisition, payments, bonuses, withdrawals, support, compliance, partner payouts, and execution or risk costs.
- Do not wait for a perfect 12-month LTV model. A 30, 60, and 90-day contribution view is enough to make better early decisions.
- Segment the numbers. Blended averages hide bad affiliates, weak payment routes, bonus abuse, and risky client behavior.
Deposit Number Is Useful, But It Is Not The Answer
Deposits matter. No brokerage grows without clients funding accounts. But deposits are movement of client funds through the business, not proof of margin.
In a regulated setting, that separation is more than an accounting preference. Client money is subject to specific protections and handling requirements. The FCA’s client money rules, for example, make clear why firms must distinguish client assets from their own money. Exact obligations depend on the entity and jurisdiction, but the operating lesson travels well: do not treat money held for clients as available business income.
For an operator, the same discipline applies to the dashboard. A $500,000 deposit month can be a strong month. It can also be the start of a costly cohort if most of the volume came from a high-CPA source, paid through a fragile route, used a bonus, or left within two weeks.
That is the gap between a deposit report and unit economics.
What A Cohort Should Answer
A cohort is simply a group of clients who share a meaningful starting condition. The most common grouping is the month of first deposit. That is a good start, but it is rarely enough.
In brokerage, a cohort should help answer questions such as:
- Which affiliate brings clients who remain active after the first withdrawal?
- Which GEO converts registrations into clean funded accounts?
- Which payment method approves well but later produces more disputes?
- Which bonus campaign creates repeat deposits rather than one-time volume?
- Which client segment creates a risk or liquidity cost that marketing cannot see?
- How long does it take for an acquired client to repay their acquisition cost?
These are operating questions. They lead to decisions: raise a cap, pause a partner, add a payment route, change a bonus rule, reduce a campaign, or review an execution setting.
A weekly total deposit figure cannot answer any of them.
Start With A Cohort Definition That Matches A Decision
There is no prize for the most complicated dashboard. A cohort is useful only if the team will act on it.
| Cohort Cut | Best Used For | What It Can Reveal |
|---|---|---|
| First-deposit month | Retention and payback over time | Whether newer client groups are improving or becoming less valuable. |
| Affiliate or campaign | Acquisition decisions and payout terms | Sources that create fast deposits but poor retained value. |
| GEO and payment method | Payment routing and compliance review | Approval problems, disputes, withdrawals, or local-method gaps. |
| Bonus or promotion | Incentive governance | Whether a promotion produces repeat activity or only subsidised first deposits. |
| Product or client behavior | Risk and execution review | Concentrated trading behavior, news sensitivity, or unusual cost to serve. |
At launch, use three cuts: first-deposit month, source, and GEO. Add payment method once volume is high enough to avoid making decisions from five transactions. Add behavior and product segments when risk or dealing data is reliable.
Trying to slice every account by twelve dimensions from day one usually creates a report nobody trusts. Start with the decision you need to make next week.
Core Metric: Retained Net Value
Different firms account for revenue differently. A brokerage’s execution model, jurisdiction, products, and agreements all matter. So there is no universal single formula.
But the operating view can be simple:
Retained net value = revenue contribution from the cohort – direct acquisition cost – payment costs – incentives – partner payouts – expected disputes and fraud – direct support and compliance cost – execution and risk cost.
Use the formula before fixed company overhead. That gives you contribution margin: the amount a cohort contributes toward the business after the costs that exist because that cohort exists. You can layer payroll, technology, legal, and management costs on top later.
The order matters. A cohort that is negative before fixed costs is not being saved by scale.
| Measure | Why It Belongs In The Cohort View | Common Mistake |
|---|---|---|
| Acquisition cost | Shows what it cost to obtain the funded client. | Using lead cost instead of cost per approved, funded client. |
| Payment cost | Captures fees, retries, failed-payment handling, and route-specific drag. | Using one blended PSP fee for every country and source. |
| Bonus and credit cost | Shows whether an incentive bought retention or just activity. | Recording it only as a marketing expense without linking it to behavior. |
| Withdrawal and dispute behavior | Shows liquidity pressure, chargeback exposure, and trust problems. | Looking at withdrawals only as a finance metric. |
| Support and compliance load | Measures cost to serve and early signs of friction. | Assuming every funded client costs the same to support. |
| Execution and risk cost | Connects client volume to the economics of the actual trading flow. | Reviewing trading revenue without the cost or exposure that created it. |
Deposits look big. What did the cohort actually keep?
Enter a simple 60-day cohort view. The calculator separates client money movement from contribution after acquisition, bonus, payment, dispute, support, and risk costs.
An Illustrative Cohort P&L
The figures below are illustrative. They are not industry benchmarks and should not be copied into a business plan. Their purpose is to show how a good-looking cohort can become weak once the full path is visible.
Imagine 200 first-time depositors from one affiliate in one GEO, assessed after 60 days.
| Item | Illustrative Amount | What The Deposit Dashboard Misses |
|---|---|---|
| Gross deposits | $140,000 | This is client money movement, not a margin line. |
| Revenue contribution from client activity | $31,000 | It is lower than deposits and depends on the business model. |
| Affiliate payout and paid-media cost | -$20,000 | Paid before long-term client quality is proven. |
| Bonuses and credits | -$5,500 | May increase FTDs while lowering client quality. |
| Payment fees, failures, and retry handling | -$3,000 | Route performance is rarely uniform. |
| Expected disputes and fraud loss | -$2,800 | Often appears after initial partner commissions are accrued. |
| Support, compliance, and operational allocation | -$2,600 | High-friction cohorts use more people and time. |
| Execution and risk reserve | -$2,200 | Client behavior can carry a cost beyond the visible spread or commission. |
| 60-day contribution before fixed overhead | -$5,100 | Deposit growth did not produce a profitable cohort. |
Nothing in this example means affiliates, bonuses, or PSP fees are bad. They are normal parts of brokerage operations. The point is that the broker has to see them together before increasing volume.
This is the same blind spot behind deposit growth that masks weakening economics. The amount deposited can rise while the value left after costs falls.
Why 30, 60, And 90 Days Are More Useful Than A Single LTV Number
Lifetime value is helpful in theory. Early-stage brokerages often misuse it because they forecast years of value from a few weeks of data.
A better early rhythm is to review cohorts at set points:
| Review Window | What You Can Reliably See | Decision It Supports |
|---|---|---|
| 7 days | Registration-to-KYC flow, first-deposit approval, early payment friction, obvious abuse. | Fix the funnel or pause a route before more spend lands. |
| 30 days | Activation, first withdrawal signals, support tickets, bonus behavior, basic source quality. | Adjust caps, promotion rules, and source-level acquisition terms. |
| 60 days | Repeat funding, retained activity, disputes, cost to serve, preliminary contribution. | Scale proven cohorts or stop subsidising weak ones. |
| 90 days | More stable retention and payback patterns, subject to settlement and reporting lag. | Set longer-term terms with partners and decide where to invest. |
Do not call a cohort profitable on day seven because it deposited quickly. Do not call it unprofitable on day seven either. Some costs and some client value arrive later. The aim is to learn fast without pretending the first week is the whole story.
What should the team decide at 7, 30, 60, and 90 days?
Choose the review window. The block shows what is reliable enough to inspect, what decision it supports, and what mistake to avoid at that point.
Payment And Withdrawal Data Belong In The Same View
Marketing teams often see drop-off after registration and assume the landing page is weak. Sometimes it is. But the client may have passed KYC, tried to deposit, been declined twice, and left because the relevant local method was missing.
That is why payment performance needs a cohort view. Compare approval, retry, withdrawal, and dispute behavior by source, country, payment method, device, and KYC status. A blended rate is polite. It is rarely useful.
A payment setup can make the same traffic either affordable or unworkable. Payment conversion in brokerage shapes the part of the funnel where a registration becomes a real client relationship.
Withdrawals should sit beside deposits, not in a separate finance spreadsheet. A clean, predictable withdrawal can support repeat funding and trust. A delayed or confusing withdrawal can turn a previously good cohort into a support and reputation problem.
Execution Risk Is A Cohort Cost, Not Just A Dealing Desk Report
Client groups do not behave the same way. One cohort may trade small, diversified positions during normal hours. Another may cluster around a single instrument before high-impact news. The second group can have a different liquidity, hedge, and market-risk cost even if the deposit totals match.
That does not mean a broker should label every profitable trader as bad. It means the economic view needs to match the actual client flow and the firm’s execution model. This is one reason retail OTC leveraged products draw regulatory attention. IOSCO’s report on retail OTC leveraged products discusses the investor-protection issues that arise in this area.
At an operating level, add a risk or execution reserve by cohort where appropriate. Review concentrated behavior, event-window activity, instrument mix, exposure, hedging costs, and exceptions. The goal is not to punish clients for trading well. The goal is to understand the cost of serving a client group before marketing scales it.
As a brokerage grows, disconnected reports become a real risk. Brokerage risk management at scale often breaks first in the gaps between CRM, payments, trading, support, finance, and partner data.
Expert Insight: The Best Cohort Is Not Always The Largest One
A source that brings 1,000 first-time depositors may look more valuable than one that brings 250. But size is not quality. The smaller source may have a better approval rate, a higher second-deposit rate, fewer disputes, lower support demand, and a more predictable payout cycle.
In that case, the smaller source is a better business until the larger one proves it can improve. Do not let scale outrun evidence just because the top-line dashboard looks busy.
The biggest source is not always the source to scale
Compare three acquisition sources. The score rewards retained value and clean behavior, then discounts disputes, support drag, bonus dependency, and payment friction.
Affiliate A
Community B
Paid Search C
Affiliate Payouts Need A Cohort Payback Rule
CPA and revenue-share terms are commercial tools, not automatic growth levers. The difficulty is timing. The brokerage often pays or accrues the partner reward before it knows whether the client will stay, withdraw quickly, create a dispute, or generate enough contribution to repay the acquisition.
A practical early approach is staged rather than generous by default:
- Cap volume for a new source until the first cohort has been reviewed.
- Write down what a valid funded client means before traffic begins.
- Review payment, KYC, chargeback, and withdrawal behavior by source.
- Increase the cap or CPA only after the 30- or 60-day numbers support it.
- Use a hybrid payout only when reporting and client-quality data are strong enough to settle it fairly.
This should not become an excuse to underpay good partners. The opposite is true. When a partner produces clean, retained, net-positive cohorts, the data gives the brokerage a reason to pay more with confidence.
Expert Insight: Reconciliation Is Part Of Unit Economics
It is easy to build a cohort model in a spreadsheet and forget that every number has to come from somewhere. If CRM says 500 FTDs, payments says 470 settled deposits, the affiliate system says 530 conversions, and finance has a different number again, the model is not ready for a scaling decision.
Pick a data owner for each event. Agree on the definition of registration, KYC approval, first deposit, settled deposit, withdrawal, chargeback, and payable conversion. Then reconcile the systems on a fixed schedule.
A useful brokerage CRM is valuable here because it can connect commercial activity with payments, client status, support, and retention. It still needs clear ownership. Software cannot resolve a business definition that the team has never agreed on.
What To Measure Every Week
Do not turn the weekly meeting into a dashboard tour. Use a short scorecard that forces a decision.
| Metric | Review By | Question To Ask |
|---|---|---|
| Cost per approved funded client | Source and GEO | Are we paying more for a cohort that is worth less? |
| Deposit approval and retry rate | PSP, payment method, device, and GEO | Is payment friction making good traffic look bad? |
| Second-deposit rate | Source, campaign, and payment method | Did the first experience create trust? |
| Withdrawal time and failure rate | Method, GEO, and client segment | Where could trust or operational capacity break next? |
| Bonus cost as a share of contribution | Promotion and cohort | Is the incentive creating retained value? |
| Disputes and chargebacks | Source, method, and GEO | Which segment needs a cap or a review? |
| Net contribution at 30, 60, and 90 days | First-deposit cohort and source | Should we scale, fix, or stop this acquisition? |
Common Errors That Make Cohort Reports Useless
Using Only A Blended Average
Blended CAC, deposit approval, and retention can all look fine while one affiliate or one country is quietly losing money. Segment first, then roll up.
Counting A Client Before The Cash Has Settled
A tracked deposit, an approved deposit, and a settled deposit are not always the same event. Make the definition clear, especially before paying a partner or declaring a campaign successful.
Ignoring Small Cohorts
Small data sets can be noisy, but they can still expose a serious issue: a payment route failing, repeated KYC friction, or a source that produces obvious abuse. Use them as a signal to investigate, not as proof of a permanent trend.
Calling All Withdrawals A Problem
Legitimate withdrawals are part of a trusted brokerage relationship. The useful question is not whether clients withdraw. It is whether the pattern, timing, and cost differ by cohort in a way that explains weak retained value.
Building A Model That Nobody Can Operate
A perfect model that takes three weeks to close is less useful than a simple one the team can review every Friday. Start with a stable definition and improve it as data quality improves.
How To Build The First Cohort View
- Choose one cohort definition: first-deposit month plus acquisition source.
- Set the event definitions and confirm which system owns each one.
- Pull 30-, 60-, and 90-day data for deposits, activity, direct revenue contribution, acquisition, payments, bonuses, payouts, disputes, support, and risk costs.
- Reconcile the totals with finance before using the model for partner or budget decisions.
- Find the cohort with the weakest contribution and identify the first leak, not every possible leak.
- Choose one action: cap a source, fix a payment route, change a promotion, adjust validation, or investigate execution.
- Review the same cohort next week to see whether the action changed the numbers.
For a new business, a connected operating stack can make that work less manual. A white label brokerage setup may include shared platform, CRM, billing, reporting, and risk tools. It can reduce the number of data handoffs. It does not remove the brokerage’s responsibility to define the metrics, review the economics, or act on a weak cohort.
Bottom Line
Deposits are a starting signal. Cohort economics is the operating truth.
The broker that measures only registrations, FTDs, and gross deposits will eventually spend more on what looks good. The broker that measures retained net value by cohort can see which clients, sources, and payment routes deserve more capital.
You do not need a perfect model on day one. You need a consistent model that makes weak economics visible before they become expensive.
