Zulic v CMC Markets Asia Pacific Pty Ltd (No 3)

This judgment sits in the middle of a larger class action. The applicants are suing CMC Markets Asia Pacific Pty Ltd over alleged practices connected with online trading in CFDs and binaries. The Court did not decide the truth of those allegations here. Instead, it dealt with a practical but important question: what more did CMC have to hand over so the case could be prepared properly?

That matters because discovery disputes often shape the evidence that will later be available at trial. If a party gets access to customer-level data, internal committee papers or witness-specific records, that can materially affect how the case is run. If the Court refuses those requests, the case may proceed on a narrower evidentiary base.

Jackman J noted that there had already been an earlier discovery dispute in 2023 and that CMC had already produced substantial material, around 26,000 unique documents from a review pool of 1.7 million documents. Even so, the applicants said they still needed more targeted categories of data and documents. Some categories were agreed or modified, but six remained in dispute.

The Court approached the application by reference to the Federal Court Rules and the overarching purpose in the Federal Court of Australia Act. In practical terms, that meant asking whether each category would genuinely help resolve the proceeding justly, quickly, inexpensively and efficiently.

The applicants wanted several categories of further discovery. The clearest categories in the available reasons were customer trading and loss data, suitability data, and hedging-related documents. The formal orders also show that some additional custodian, committee pack and group member witness categories were granted, although the visible extract does not include the full reasoning for all of them.

Category 1(a) sought a copy of data in the format of a spreadsheet called the Raw Data Sheet for the period from 1 April 2016 to 30 April 2021. That spreadsheet had already been created by CMC to respond to an ASIC notice. The judgment says it contained annual summary information for each account and customer ID, including whether the customer was retail, partner or institutional, the number of CFD and binary trades, turnover, losses, fees, rebates, and profitability with and without fees.

The applicants said this mattered because their case was not limited to broad averages. They wanted to rely on many individual instances to support an inference that CMC engaged in a broader system of conduct or pattern of behaviour. The Raw Data Sheet would let them look at customer outcomes over time and across categories of customers.

Category 2 sought suitability data. The fields requested included application and customer identifiers, appropriateness status and reason, whether the customer agreed to proceed, acceptance of assessment, application status, and withdrawal or rejection reason. The applicants said this would help them examine how CMC actually dealt with customers who did not pass internal suitability or appropriateness checks.

Category 3 sought documents about hedging, including the use of a C Book and criteria for allocating customers to risk book categories such as TOXIC, FLOW and RN1 to RN5. The applicants wanted to understand what policies existed and how customers were classified for hedging purposes.

The legal issue was whether the Court should order further discovery, and if so, for which categories. This was governed by the Federal Court Rules 2011, especially the rule that a party must not apply for discovery unless the order sought will facilitate the just resolution of the proceeding as quickly, inexpensively and efficiently as possible. Jackman J said that rule reflects the overarching purpose in section 37M of the Federal Court of Australia Act 1976.

Because the applicants were seeking discovery by categories rather than standard discovery, the Court also considered the standard discovery criteria referred to in the Rules. In broad terms, the Court looked at whether the material was directly relevant to the issues raised by the pleadings or affidavits, whether it was within the respondent's control, and whether the burden of compliance was justified by the likely probative value.

That framework is important for businesses. Discovery is not a free-for-all. A party asking for more documents must show practical utility, not just curiosity. But once a category is shown to be directly relevant and proportionate, the Court may order production even where extraction takes time and specialist staff effort.

The Court ordered further discovery in part. CMC was ordered, by 29 August 2025, to give verified discovery of the data described in categories 1(a) and 2, and documents responsive to categories 6(b), 7(a), 7(b) and 8(a) to 8(f). Discovery was to be produced in accordance with the parties' Electronic Exchange Protocol. The amended interlocutory application was otherwise dismissed, and costs were ordered to be costs in the cause.

For category 1(a), Jackman J accepted that the Raw Data Sheet information was directly relevant. The applicants alleged a system of conduct or pattern of behaviour designed to induce customers to commence and continue trading, increase trading despite losses, continue paying fees, and return after inactivity. The Court accepted that those broader allegations could be proved by inference from a sufficient number of individual instances. Aggregated weekly reports were not enough because they did not provide the same individualised trading performance over time. The Court also accepted that, although only one person in Australia could extract the data and the task would take three or four weeks, that burden was not excessive given the potential probative value.

For category 2, the Court accepted that suitability data was relevant to allegations that CMC failed to adequately assess customers' suitability for trading and continued trading in CFDs or binaries. The Court accepted that policy documents alone did not show CMC's actual responses to customers who did not pass suitability assessments. The applicants were entitled to seek data showing whether customers were nevertheless allowed to open accounts after acknowledging risk warnings, whether some customers applied multiple times after failing, and what happened to those applications. The Court also accepted that linking category 2 data with category 1(a) data could be a meaningful step in the applicants' case, even if the suitability data would not by itself be conclusive. The respondent's evidence was that producing category 2 would take a combined total of eight days, which the Court did not regard as overly burdensome.

For category 3, the Court refused discovery. Although the Court accepted that the concern about customer allocation to the C Book or later equivalents was relevant, it considered that the overwhelming likelihood was that any truly relevant documents were within the control of CMC's UK parent, which managed risk at group level, rather than within the possession, custody or power of the Australian respondent. The Court also considered the prospect of finding useful material in the nominated Australian custodians' very large mailboxes too speculative and remote to justify the burden.

The available reasons then move into category 4, concerning instant messaging records, but the visible text cuts off before the Court's full reasoning on that and later categories. The formal orders nevertheless make clear which categories were ultimately granted.

For a business owner, the most useful way to read this case is as a reminder that courts often want to see how systems operate in practice, not just what the policy manual says. If your business has onboarding checks, appropriateness assessments, risk warnings, override pathways or customer segmentation, those elements may later be examined together.

One of the strongest points in the judgment is the Court's acceptance that actual customer-level outcomes can matter. A business may think it is protected because it has a written process requiring warnings or acknowledgements. But if the data shows that customers who failed an internal assessment were still allowed to proceed, and later suffered losses, that combination may become highly relevant in litigation.

The decision also shows the importance of data architecture. CMC had already created a Raw Data Sheet for an ASIC notice, and that existing format became central to the discovery application. If your business can already extract a dataset for a regulator, a court may be more willing to order production of similar data in litigation.

Another practical point is corporate group structure. Businesses often assume that because a related entity overseas manages a function, the local company can simply say the documents are elsewhere. Sometimes that will be right, but the issue turns on control, not just practical involvement. Businesses should understand in advance which entity owns, stores and can access key records.

Finally, this case is a reminder that discovery disputes are often won or lost on specificity. The applicants succeeded where they could point to concrete datasets, fields and use cases. They failed where the request was too speculative or where the likely documents were outside the respondent's control.

The judgment is a good example of how different records can combine into one litigation narrative. The applicants were not just asking for policy documents. They wanted customer trading data, suitability data, committee materials, custodian documents and witness-specific records. That reflects how modern disputes are often proved: by connecting operational data, internal governance and customer interactions.

For businesses, that means recordkeeping should be thought about across systems, not in silos. Marketing, onboarding, risk, customer service and governance records may all become relevant to the same allegation. If those systems do not align, the inconsistency itself can become a problem.

The case also highlights the difference between aggregate reporting and customer-level evidence. Summary dashboards may be useful internally, but they may not answer the questions a court asks. If the issue is how individual customers were treated over time, more granular data may be required.

There is also a procedural lesson. The Court expressly tied its analysis to the Federal Court Rules and the overarching purpose provisions. That means parties should expect discovery arguments to be framed around efficiency and proportionality, not just broad assertions that more documents might be interesting. Businesses involved in litigation should therefore be ready to explain both the relevance of requested material and the real burden of producing it.

The judgment was delivered on 9 May 2025 by Jackman J in the Federal Court of Australia. The hearing took place on 30 April 2025. The orders required CMC to give verified discovery of the granted categories by 29 August 2025. The costs of the interlocutory application were made costs in the cause, meaning they will follow the outcome of the broader proceeding unless the Court later orders otherwise.

This page should be read as an explanation of a procedural step in ongoing litigation. It does not describe the final result of the representative proceeding.

This explainer is based on the Federal Court judgment in Zulic v CMC Markets Asia Pacific Pty Ltd (No 3) [2025] FCA 464. The available text clearly sets out the formal orders, the governing discovery principles, and detailed reasoning for categories 1(a), 2 and 3. The visible text then becomes truncated during the discussion of category 4, so the reasoning for some later categories is not fully available here.

Because of that limitation, this page focuses on what can be stated confidently from the judgment text that is visible, especially the procedural setting, the categories expressly ordered, and the detailed reasons that are fully available.