The quiet line in ASIC's report
In January 2026, ASIC published the results of a sector-wide review of contracts for difference. The headline was the refund: nearly $40 million returned to more than 38,000 retail investors after the regulator found more than half the sector had broken its product intervention order. The figure is in Australian dollars.
The headline is not the part prop firm operators should read twice.
The part to read twice comes after the refunds. Alongside the issuers that changed their target markets and improved their onboarding questionnaires, the regulator recorded that 42 issuers "implemented new processes for ongoing monitoring of client trading outcomes and behaviours, or made significant improvements to existing processes." The report treats that monitoring as part of an obligation issuers already had, not an extra: issuers "should not rely on client questionnaires as the sole or key method" of keeping distribution inside the target market, and ASIC found most "did little to no ongoing monitoring of client trading outcomes and behaviours." The question a high-risk-product firm answers is no longer only "how did you sell this?" It is now "what did you do once you saw the client struggling?"
This piece is for operators who run a challenge-and-funding model and have read the coverage as a broker problem. It is not only a broker problem. In our reading, the question the review asks of CFD issuers applies just as well to that model.
What did ASIC's CFD review find?
ASIC's Report 828, Risky business: Driving change in CFD issuers' distribution practices, reviewed 52 licensed CFD issuers between October 2024 and December 2025. The findings, in the regulator's own numbers:
- Nearly $40 million in refunds to more than 38,000 retail investors.
- More than half the sector breached the product intervention order through a margin-discount workaround.
- In the Australian 2023-24 financial year, 68% of Australian retail CFD clients realised a net loss after fees (133,674 of 195,386), 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.
- In response to ASIC's intervention: 39 issuers made changes to their target markets, 46 improved website content, 44 improved their onboarding questionnaires, and 42 put in new processes to monitor client trading outcomes and behaviours or made significant improvements to existing ones.
- The CFD product intervention order expires on 23 May 2027 unless it is remade, which puts the sector under a live clock.
ASIC Commissioner Simone Constant put it in one sentence: "Each year, thousands of Australians lose money trading CFDs and through our review we have helped put $40 million back in the pockets of more than 38,000 investors."
Why this reaches prop firms, not just brokers
ASIC reviewed CFD issuers, not prop firms, and its January 2026 report and release do not mention prop trading. But in an August 2024 interview reported by Finance Magnates, Dr Rhys Bollen, then ASIC's Senior Executive Leader of the Digital Assets and Markets Group, said the commission "is monitoring the emergence of 'prop trading' firms or services relating to CFD trading" and planned, for 2024/25, "detailed surveillance of new and emerging distribution methods across the CFD industry."
Read that as a category being watched, not a verdict being passed. The comment does not say how ASIC would classify a paid-evaluation model, and that question is not settled.
The direction is not Australia-specific. The UK's Consumer Duty already requires FCA-regulated firms to monitor the outcomes their retail customers receive, and the FCA expects them to be able to evidence those outcomes rather than assume them. Read together, these point at what happens to the client after the sale, not only at how the sale was made. A prop firm can put the same question to itself, whether or not any of these frameworks reaches it.
The reframe: a better question
For a decade, the implicit question a prop firm answered was "how many evaluations can we sell?" A better question to ask now is "what happens to the clients who lose?" Those are different businesses wearing the same logo.
The distinction between the two numbers matters here. ASIC's 68% is a loss-rate: the share of Australian retail CFD clients who realised a net loss after fees in one Australian financial year. Regulators publish that figure. The ~75%-quit-within-90-days number prop firms cite is churn, and it is industry-cited rather than regulator-published. The full sourcing on the churn figure is here. One number measures money lost. The other measures people leaving.
A firm cannot answer "what happens to the clients who lose?" with a discount code for the next attempt. It answers with evidence. Evidence that it saw the trader tilting, reached them before the breach, and tried to change the outcome rather than bank the next fee. That evidence does not exist unless the firm built the layer that produces it.
What "retention you can defend" looks like
Defensible retention has three properties.
It monitors behaviour, not just trades. A transaction log records what happened. It does not flag the revenge entry, the rapid-fire sequence, or the oversized position. The six behavioural triggers we watch for can show in the data before a breach, to an engine watching in real time. Here that monitoring is the product, not something added later under pressure.
It acts inside the window. Monitoring that produces a report tomorrow is not action. We design for the minutes between a trigger and the next trade, a working window of about four minutes: our own operational construct, not a measured figure. A text nudge at the warning threshold. A coaching call at active tilt. The intervention has to fire while the trader is still at the desk, not in the morning email.
It coaches, it does not advise. A coach asks the trader what their own plan says. It reflects the trader's rules back to them. It does not recommend trades, suggest position sizes, or predict prices. That boundary is not a disclaimer bolted on at the end. It is the difference between a layer that helps a client and a layer that creates new liability. Coaching, not financial advice.
Set the two models side by side and the difference is hard to miss.
| Extract-and-churn | Coaching-led retention | |
|---|---|---|
| Optimises for | Evaluation fee volume | Funded-trader longevity |
| When a client struggles | Sell the next challenge | Intervene before the breach |
| What the firm can show | Sales to traders who keep losing | A record of what it did when a client struggled |
| How it scales | More ad spend | One engine, every funded trader |
Does this make Discentra a compliance product?
No. Discentra is a performance-coaching layer, not a compliance or risk tool. It does not block trades, file regulatory reports, or make a firm compliant. A firm's regulatory classification is the firm's own responsibility. Discentra monitors behaviour and is built to call the trader within seconds of a trigger, with coaching rather than advice. Any compliance value is indirect: a firm that monitors client outcomes and acts on them has a record to show if anyone asks.
The window is the opportunity
REP 828 reads two ways. As a cost, it is a year of retrofitting onboarding questionnaires and rewriting website disclosures to satisfy a checklist.
As a signal, it is something better. It counts what happened to clients after onboarding, not only how they were sold to. Knowing what happens to the client after onboarding, and acting on it, is also the work aimed at moving the retention number: the trader who blows up on a Tuesday afternoon is the one a firm loses. Our retention playbook and a firm's answer on client outcomes are not two projects. They are one layer.
A firm that builds it can say, to a trader and to a regulator in the same breath: we saw it coming, and we called.
Coaching, not financial advice. The plan stays the trader's. The rulebook stays the trader's. The intervention is a coach reading that rulebook back to the trader in the minutes after a trigger, the window we design for. That is retention you can defend.
