# Prop firm churn and trader retention statistics (2026)

> Discentra's core numbers on prop firm churn and trader retention, each labelled by where it comes from: regulator-published, published research, industry-cited, or Discentra primary data and disclosed estimates.

Last updated: 20 October 2026. 11 statistics.

**What percentage of traders quit within 90 days?** ~75% of retail traders quit within their first 90 days. That is the industry-cited working figure, and we have not been able to trace it to a named primary source. No regulator publishes a churn rate: regulators publish loss rates, a related but different metric. ESMA published national regulators' 2018 analyses putting the loss rate at 74-89% of retail CFD accounts, a baseline from before its own restrictions. ASIC recorded 68% of Australian retail CFD investors losing money in FY2024. Every statistic below carries a label, and the numbers that circulate without a verifiable source are flagged at the end, not recycled.

## 1. ~75% of retail traders quit within 90 days (traders-quit-90-days)

The lead number in trader retention, repeated across retail brokerages, prop trading firms and crypto exchanges. If it holds for your firm, most of your churn sits in the first 90 days, and a retention effort that starts at month six arrives after it. Your own cohort data will show whether it does.

Honesty note: no single canonical 90-day figure exists, and the estimates in circulation vary with segment and with how quitting is defined. ~75% is the working number we use: industry-cited, not regulator-published. The structural forces behind it are unpacked at [why traders quit](https://discentra.ai/why-traders-quit).

**Source:** Industry-cited working figure. We have not been able to trace it to a named primary source, so treat it as a widely-repeated working number rather than a measured one. No regulator publishes a churn rate; see statistic 2 for what regulators do publish.

## 2. 74-89% of retail CFD accounts lost money before the 2018 EU limits (esma-loss-rate)

In its 2018 product intervention decision, ESMA published EU national regulators' analyses showing that "74-89% of retail accounts typically lose money on their investments, with average losses per client ranging from €1,600 to €29,000."

The distinction matters: this is a loss rate, not churn, and a 2018 baseline from before ESMA's own leverage caps and negative-balance protection, not a current reading. It measures money lost, not people leaving, and the two get conflated across aggregator sites. The churn side of the equation, what it costs a broker and what changes it, is the subject of [behavioural retention for brokers](https://discentra.ai/brokers).

**Source:** [ESMA, product intervention analysis, press release 27 March 2018](https://www.esma.europa.eu/press-news/esma-news/esma-agrees-prohibit-binary-options-and-restrict-cfds-protect-retail-investors). Regulator-published loss rate, not a churn rate.

## 3. 68% of Australian retail CFD clients lost money in FY2023-24 (asic-cfd-losses)

ASIC's Report 828 reviewed 52 licensed Australian CFD issuers between October 2024 and December 2025. It returned nearly $40 million to more than 38,000 retail investors after finding more than half the sector had breached the product intervention order. In the Australian 2023-24 financial year, 133,674 of 195,386 retail clients, or 68%, realised a net loss after fees, totalling $458 million including $73 million in fees. ASIC notes 5% would have made a net profit but for fees, so part of that rate is cost structure rather than trading.

The line worth reading twice: in response to ASIC's intervention, 42 issuers put in new processes to monitor client trading outcomes and behaviours, or made significant improvements to existing ones. Any lesson in that for prop firms is our reading, not ASIC's. Our full read of the report: [retention you can defend](https://discentra.ai/blog/retention-you-can-defend). Like ESMA's figure, this is a loss rate, not churn.

**Source:** ASIC Report 828, January 2026. Regulator-published loss rate. Figures in Australian dollars; the refund total is about US$27 million at the Reserve Bank of Australia's rate on the day of the release (A$1 = US$0.6738, 20 January 2026).

## 4. 1 of 251 negative reviews: the writer described their own behaviour driving the loss (reviews-strict-count)

In October 2026 Discentra read the most recent one- and two-star Trustpilot reviews of 10 prop firms, 251 in all, and counted a review as behavioural only when its writer described their own behaviour driving the loss or the breach. Under that rule, 1 of the 251 counted. Almost every other review was about the firm: disputed breaches, payouts, identity checks, platform problems.

Read it as a count of reviews, not a share of churn. Reviews come from the traders who choose to write one, and a looser counting rule finds more. The behavioural side of a firm's churn would show in its own trade data. What a read of reviews can and cannot tell a firm is set out in [what reviews can't show you](https://discentra.ai/churn-report).

**Source:** Discentra primary data: 251 one- and two-star Trustpilot reviews across 10 prop firms, read in October 2026 under a strict rule (behavioural only where the writer describes their own behaviour driving the loss or the breach). All 251 re-read, and every behavioural count confirmed, in a second, separate read under the same rule; methodology below.

## 5. 10-15% estimated monthly churn at prop firms (monthly-churn-rate)

We know of no regulator or industry body that publishes an audited prop firm churn rate, so any monthly figure you see is an estimate. This one is ours, a disclosed assumption in our cost models, including the worked model in [the real cost of trader churn](https://discentra.ai/blog/the-real-cost-of-trader-churn). A firm's true churn rate comes from its own cohort data, which is the number worth knowing.

**Source:** Discentra working estimate, used as a disclosed assumption in its churn cost models. Not an audited industry figure; we know of none.

## 6. $200 to $2,000 to acquire one trader (trader-acquisition-cost)

On our working range, firms spend $200 to $2,000 acquiring each trader, depending on the segment and channel: paid ads, affiliate commissions, influencer deals, SEO content, and the sales team's time. When a trader churns, that spend is gone. Not reduced. Gone.

The full unit economics, including onboarding, compliance, and departure costs, are worked through in [the real cost of trader churn](https://discentra.ai/blog/the-real-cost-of-trader-churn).

**Source:** Discentra working range from our own market research. We have not been able to name a published source for it, so it is our assumption rather than an industry finding. Our churn models use $200 to $500 as the disclosed working assumption.

## 7. $1,200 to $3,500 a year: our working assumption for trader value (trader-lifetime-value)

On our working assumption, a trader generates $1,200 to $3,500 a year, provided they stay long enough to trade meaningful volume, which early churn cuts short. In our own worked example (statistic 10), lost remaining lifetime value is the largest single cost.

**Source:** Discentra working assumption across retail brokerage models, not a published industry figure. Note the unit: our worked churn example uses $1,200 as an ANNUAL value at the conservative end, so read this range as annual revenue per trader rather than whole-relationship lifetime value.

## 8. 250+ prop firms in operation (prop-firm-count)

Our count of firms offering paid evaluations and funded accounts runs past 250. In our view, when that many offers share similar rules, fee structures and platforms, retention becomes one of the few edges left. Our playbook for competing on it without re-pricing the offer: [the prop firm trader retention playbook](https://discentra.ai/blog/prop-firm-trader-retention-playbook).

**Source:** Discentra market analysis: our own count. We do not publish the counting rule or a dated source, so treat it as an estimate.

## 9. A 5% cut in defections raised profits by 25-85% in the businesses Reichheld studied (retention-profit-effect)

Across Reichheld and Sasser's 1990 paper and Reichheld's later work at Bain, cutting customer defections by 5% raised profits by 25% to 85%. The spread is industry-dependent: the 85% comes from a single bank's branch system, with 50% in an insurance brokerage and 30% in an auto-service chain. The Bain work puts the financial-services figure at more than 25%. Even the floor of that range makes a case for a retention budget.

A widely circulated version gives the range as 25% to 95%. We do not use it. The 95% ceiling does not appear in Reichheld and Sasser's paper or in the Bain material it is usually sourced to, and the 2014 Harvard Business Review article it is often attributed to was written by Amy Gallo, summarising Reichheld's earlier work rather than reporting a new finding.

**Source:** Reichheld, F. F. and Sasser, W. E., [Zero Defections: Quality Comes to Services, Harvard Business Review, September-October 1990](https://hbr.org/1990/09/zero-defections-quality-comes-to-services). Service-industry research finding.

## 10. $18.5 million: the annual churn cost worked example (annual-churn-cost-worked-example)

A composite mid-size prop firm: 10,000 active traders, 10% monthly churn, $500 average CAC, $1,200 annual LTV. That firm loses 12,000 traders a year, and the total annual churn cost lands at $18,540,000 across lost remaining lifetime value ($11.4M), replacement acquisition ($6M), onboarding the replacements, departure support and review damage. It leaves out what each leaver cost to acquire, which was spent whether they stayed or not.

Every assumption is disclosed so you can run it on your own numbers. [Calculate your own churn cost](https://discentra.ai/calculator).

**Source:** Discentra worked example with disclosed composite assumptions, not a measured industry figure. Full arithmetic in [the real cost of trader churn](https://discentra.ai/blog/the-real-cost-of-trader-churn).

## 11. 4x lifetime value past the 90-day mark (ltv-multiplier-90-days)

On our modelling assumption, a trader who makes it past 90 days is worth about 4x one who does not. If the 90-day figure in statistic 1 holds for a firm, its early churn is the expensive kind. On our working assumption of $1,200 to $3,500 a year, a trader who leaves in month two takes most of a year's value with them.

**Source:** Discentra modelling assumption used in our churn cost analysis. We have not been able to trace it to a published source, so it is not an industry-measured multiplier.

## Numbers that circulate without a source (unsourced-numbers)

Two figures appear across affiliate aggregator sites, social posts, and pitch decks with no verifiable primary source behind either of them:

- "95% of traders fail" (no verifiable primary source)
- "98% of traders quit within 6 months" (no verifiable primary source)

We could not trace either of these to a regulator, a peer-reviewed study, or a disclosed dataset. The pages we found citing them tend to cite each other, and repetition is not a source. Some may have started as a reasonable guess; circulation turned them into "statistics."

Discentra does not use these figures, and this page does not present them as facts. If a number cannot carry a source label, it does not appear in the list above.

## Methodology (methodology)

The primary data on this page is a review-by-review read of public Trustpilot reviews. In October 2026 Discentra read the most recent one- and two-star reviews of 10 prop firms published in the previous 12 months, newest first and up to 70 per firm: 251 reviews in all. Each review was given one code. It counted as behavioural only where its writer described their own behaviour driving the loss or the breach; a loss on its own did not count. All 251 were then re-read, and every behavioural count confirmed, in a second, separate read under the same rule.

Financial estimates use disclosed working assumptions (10-15% monthly churn, $200 to $500 prop firm acquisition cost), not proprietary benchmarks. Four limitations worth naming:

- It is 10 firms, not the industry, and the firms are not named.
- How far back each read reaches differs by firm, from a full year to a single day.
- Reviews come from the traders who choose to write one. The count describes reviews, not the traders who left, so it is never a share of churn.
- The rule is strict by design. A looser rule, one that also counts a trader's reaction to the firm's rules or a repeat purchase after a failure, finds more.

Until October 2026 this page carried a behavioural share of one prop firm's negative reviews, counted in April 2026. We retired it: a second count of that firm's reviews 5 days later, on a fresh sample under a written rule, gave a different share and a different most common pattern.

What a read of public reviews can and cannot show is set out in [what reviews can't show you](https://discentra.ai/churn-report). Discentra is a B2B AI voice coaching platform for prop firms, brokers, and crypto exchanges. Coaching, not financial advice.

## How to cite this page (how-to-cite)

Discentra Ltd (2026). Prop firm churn and trader retention statistics. https://discentra.ai/prop-firm-churn-statistics. Last updated 20 October 2026.

Every statistic has its own anchor on the live page, so you can deep-link a single number, for example `https://discentra.ai/prop-firm-churn-statistics#traders-quit-90-days` for the 90-day figure. Reproduce any statistic with its source label intact: the label is part of the claim.

## Start with your own firm's reviews

We read your most recent one- and two-star reviews for free and send one page: what they show, and what they can't. The behavioural side would show in your own trade data. Ask for it on [what reviews can't show you](https://discentra.ai/churn-report#review-read). Discentra is a behavioural coaching platform for financial institutions. Coaching, not financial advice.

## Frequently asked questions

### What percentage of traders quit within 90 days?

~75% of retail traders quit within their first 90 days. This is an industry-cited working figure, which we have not been able to trace to a named primary source, and not a regulator-published rate. No regulator publishes a churn rate. Estimates in circulation vary with segment and definition; ~75% is the working number we use, and this page labels it as industry-cited wherever it appears.

### What is the average prop firm churn rate?

Discentra's working estimate for monthly churn at prop firms is 10-15%, used as a disclosed assumption in its churn cost models. We know of no regulator or industry body that publishes an audited prop firm churn rate, so any monthly figure you see is an estimate. A firm's true churn rate comes from its own cohort data. The related industry-cited figure is that ~75% of retail traders quit within 90 days.

### How much of prop firm churn is behavioural?

We know of no published figure that measures it, and public reviews cannot size it. In October 2026 Discentra read 251 one- and two-star Trustpilot reviews across 10 prop firms under a strict rule, counting a review as behavioural only when its writer described their own behaviour driving the loss: 1 of 251 did. Almost every other review was about the firm, such as disputed breaches, payouts, identity checks and platform problems. That is a count of reviews, not a share of churn. The behavioural share of a firm's churn would show in its own trade data, where each order is recorded.

### Do regulators publish trader churn rates?

No. Regulators publish loss rates, not churn rates. ESMA's 2018 product intervention decision published national regulators' analyses showing 74-89% of retail CFD accounts losing money, a baseline from before its own restrictions. ASIC's Report 828 recorded that 68% of Australian retail CFD clients realised a net loss after fees in the Australian 2023-24 financial year. A loss rate measures money lost; churn measures people leaving. The ~75% quit-within-90-days figure is industry-cited, not regulator-published, and this page labels it that way.

### Where do these statistics come from?

Four source classes, each labelled on this page: regulator-published figures (ESMA, ASIC), published research (Reichheld and Sasser's retention findings, Harvard Business Review, 1990), one industry-cited working number (the ~75% 90-day figure), and Discentra's own primary data with disclosed working estimates (a strict read of 251 one- and two-star Trustpilot reviews across 10 prop firms, October 2026). Figures that circulate without a verifiable primary source, like "95% of traders fail", are flagged on this page and not used. Discentra is a B2B AI voice coaching platform for prop firms, brokers, and crypto exchanges. Coaching, not financial advice.

---

This is a Markdown mirror of [https://discentra.ai/prop-firm-churn-statistics](https://discentra.ai/prop-firm-churn-statistics). Provided for LLM citation. © Discentra Ltd. Coaching, not financial advice.