Why data-driven decision making will shape the future of FX

Automation and electronification have gone from nice-to-haves to essentials in FX. While firms can access several solutions providers that help them automate trade processes, many initiatives grind to a halt due to a fundamental issue: data. To be precise, the issues surrounding FX data are complex and pose significant hurdles for firms looking to ramp up automation. Bart Joris, Head of FX Sell-Side trading, Customer Proposition Data & Analytics,  LSEG, underscores data’s importance.

“Transparent access to data is the key to any electronification of the FX market,” he says. “The data is key to the lifecycle of any trading activity (discovery, creation, distribution, trading, post-trade, risk management, and TCA). The more the data becomes available, the more knowledge and derived activity (distribution, trading, TCA) it will generate.”

What are the benefits of data-driven decision-making in FX and how can firms overcome the common hurdles associated with these processes? Let’s take a look.

Transforming FX through strategic data use

The landscape of electronic FX trading continues to evolve dramatically, with data emerging as both catalyst and cornerstone of this transformation. Industry experts point to a virtuous cycle where increased electronification generates more data, which in turn enables further automation and market sophistication.

“Once electronification starts, it only accelerates the volume and accuracy of the data in the FX market,” notes Joris, highlighting how this self-reinforcing process is reshaping market structures.

This evolution extends far beyond simple automation. As Stephen Totten, Managing Director, Head of Institutional and Quantitative Products at oneZero, explains, the granularity of data application has reached unprecedented levels. “Trading desks are using data to optimise pricing and execution strategies in more and more innovative ways,” he observes. “Data enables clients to optimise risk management strategies across customer groups, instruments or entities, while also providing insights into trading patterns.”

The strategic value of high-quality data becomes particularly evident when examining execution outcomes. Paul Lambert, Chief Executive Officer at New Change FX, emphasises this critical connection.

“Without robust data, the ability to fully automate processes remains constrained,” Lambert states. “By leveraging historical transaction data, benchmark rates, and transparent execution costs, firms can embed best execution practices directly into their trading strategies.”

This integration creates tangible competitive advantages, allowing firms to automate with confidence. “Leading market participants increasingly integrate benchmark data into their execution frameworks,” Lambert continues, “allowing them to automate smaller trades with confidence, knowing they are operating at or near continuous benchmark rates.”

The resulting operational efficiencies enable strategic resource reallocation, with Lydia Solinski, Managing Director, Global Head of Liquidity, Data & Business Information, pointing to the market differentiation this creates.

“Data is revolutionizing electronic FX trading by boosting speed, efficiency and decision-making,” Solinski explains. “It differentiates FX market participants by enabling those with advanced analytics and real-time insights to execute faster, more informed trades.”

The technological foundation supporting these advancements continues to evolve as well. 

“Central to all of this is cutting-edge technology, like Artificial Intelligence, which is driving innovation in FX trading with data acting as the essential fuel,” Solinski adds, underscoring how data quality increasingly determines the effectiveness of automated workflows.

Building on the transformative role of data in electronic FX trading, firms adopting data-centric operational models gain advantages across the entire trading lifecycle, creating sustainable competitive edges in an increasingly automated marketplace.

“A data-centric approach can help you to find and optimize your trading outcomes,” explains Joris. “It is the learning cycle which gives you the competitive edge.” This continuous feedback loop transforms raw transaction data into actionable intelligence, enabling firms to “create a coherent price construction, determine market conditions and reassess this to build even better trading behaviours for the future.”

The evolution from basic metrics to sophisticated analysis has fundamentally changed how trading desks evaluate performance. Totten articulates this shift in management perspective. “Traditionally, desk heads relied on end-of-day PnL and traded volumes to gauge business success; however, with a sophisticated data platform, a desk head can analyse profitability at the level of individual clients, currency pairs, trading and execution styles,” he notes.

This granular insight allows firms to “assess performance across different market conditions, including high-volatility events and quieter periods.”

Independent data

As firms accelerate automation initiatives, Lambert emphasises the critical role of independent data in maintaining operational integrity. “Manual workflows are not only costly but also susceptible to errors, whereas automation enhances accuracy and streamlines execution. However, automation without proper oversight can lead to the repetition of suboptimal outcomes,” he cautions. “Independent benchmark data serves as a critical safeguard, providing the necessary checks and balances.”

Beyond risk management, proper data utilisation directly impacts financial performance through optimized capital allocation. “By ensuring that positions are marked to market correctly—particularly in the forward foreign exchange market—firms can minimize profit and loss volatility and optimize capital allocation,” Lambert explains. “A precise mark-to-market process reduces unnecessary capital reserves against positions, enhancing overall liquidity and financial stability.”

The client relationship dimension represents another strategic advantage, with Solinski highlighting how data transparency builds trust.

“A data-centric approach offers significant strategic business advantages for FX operations. First, it helps meet the growing requirements of best execution by enhancing transparency and trade outcomes,” she observes. “Data also drives transparency and integrity into the heart of FX relationships, and promotes a culture of truly treating clients fairly.”

Looking forward, Solinski points to emerging technologies that will further enhance data utilization. “We are now investing in ML and AI technologies to ensure that data drives Liquidity optimisation and ultimately insight into how to best place trades in the FX market,” she notes, underscoring the ongoing evolution of data’s strategic value.

Data-centric approaches create competitive advantages, particularly in execution quality and benchmarking performance. Lambert emphasises the fundamental role of independent reference points in this process. “Leading firms increasingly rely on independent benchmark data to ensure execution quality and optimize trading strategies rather than use data that provides a partial view of available pricing,” he explains. “By embedding independent data into their execution frameworks, firms can confidently automate smaller trades while focusing human expertise on larger, more complex transactions.”

This capability delivers tangible benefits beyond simple automation. “Access to independent, continuous pricing ensures greater visibility into spreads, liquidity, and execution costs,” Lambert adds. “This transparency reduces information asymmetry, enabling price takers to make more informed decisions and liquidity providers to offer fairer pricing.”

The multidimensional nature of execution quality extends well beyond simplistic price metrics, as Totten points out. “It’s very easy to just assume best execution equals best price, but in today’s market more sophisticated providers consider a much broader set of factors,” he notes. “To truly optimise pricing, traders must assess the potential market impact of their trades, taking into account variables such as time of day, liquidity, information leakage, and whether one of their LPs has an axe or interest.”

This nuanced perspective aligns with Joris’s observation about the varied interpretations of best execution across the industry. “Best execution has been a major buzzword in FX, but it does not mean the same to all,” Joris explains. “Though one point is clear, best execution can only happen based on the data used and what is available to compare it too.” The quality, depth, and accessibility of data therefore become critical factors in execution optimization.

This growing emphasis on independent market data is evident in the increased adoption of specialised datasets, as Lisa Danino-Lewis, Chief Growth Officer, CLS notes: “We have seen a growth in the take-up of our CLSMarketData sets. This is indicative of a wider market infrastructure trend from the front office, which is increasingly leveraging alternative sources of trade data to support activities across a broad range of functions. CLSMarketData is derived from the largest single source of FX executed data available to the market. Our data sets – FX Volumes, FX Flow, FX Outstanding and FX Pricing – provide quality insights both on a timely and historical basis. Within buy-side firms, FX Volume and FX Flow datasets are utilised by portfolio managers as a component into systematic trading models, and are being used to improve transparency in market activity.”

Despite progress in data utilization, significant challenges remain in certain market segments, as Solinski highlights. “Swaps and Forwards are the last bastion of opacity in FX, where quality streaming data is very rare,” she notes. “This will constrain the evolution of automation in the Swap and FWD space. The data must be there for automation and best execution to take place.”

Beyond availability, Yusuf Nurbhai, Head of BestX emphasises that data must be actionable and tailored to specific client requirements. “Data and transparency alone will not help the buy side in achieving best execution outcomes as each client’s trading style and best execution policy are unique,” he observes. “It is important to create a solution that allows clients to set their preferences for the analysis, with the ability to make those insights actionable at point of execution rather than providing only a post-trade evaluation of performance.”

This forward-looking approach represents an important evolution beyond traditional transaction cost analysis. “Historically, TCA has always been a backward-looking process,” Nurbhai adds. “Initiatives like our BestXecutor, harnesses historical and live data, delivering a forward-looking approach. Such tools allow the client to make more informed decisions right at the point of execution.”

Confronting access, volume, and legacy challenges

The evolution toward data-driven FX operations faces several significant hurdles that firms must overcome to fully realize the benefits of electronification. The issue of data availability varies significantly across FX market segments. While spot markets have achieved relatively high levels of transparency, other product areas remain less accessible, affecting pricing quality and execution efficiency.

“Data is not inherently democratic, as those with access to high-quality data often protect it within the solutions they offer,” explains Yusuf Nurbhai, Head of BestX. “Independent TCA venues like our BestX® have the potential to democratize data, making it accessible to everyone.”

This democratization process is gathering momentum through a combination of regulatory pressure and market-driven solutions, as Lambert points out. “Regulators are increasingly focused on making pricing data accessible, promoting transparency and levelling the playing field,” he notes. “Frameworks like MiFID II and the FX Global Code mandate fair data access, which helps reduce search costs and ensures more informed decision-making.”

The consolidation of pricing information represents another crucial step toward greater accessibility. “Aggregating pricing from multiple venues reduces the complexity of finding true market rates, allowing market participants to access accurate, up-to-date data in one place,” Lambert adds.

Significantly, this approach directly addresses cost concerns: “By integrating independent data directly into trading platforms, firms eliminate costly technical integrations. This enhances execution efficiency, reduces infrastructure costs, and allows participants to access real-time data seamlessly within their existing workflows.”

As data becomes more commoditized, economics will also shift. “As data becomes more commoditized, its price is expected to decrease, although the cost of Swap and FWD data will remain high due to its limited availability,” Solinski observes, highlighting the persistent premium on less accessible market segments.

As data volumes expand, firms face mounting challenges in storage, transport, and processing. These technical considerations have significant implications for operational efficiency and cost management.

“These days I see technology as an enabler rather than a restriction to the growth of data. Times where vast amounts of data were troublesome to handle are over,” observes Joris. However, he cautions that “it is more a discussion on what data is relevant for the use cases, and this will have greater impact on the success factor as more data does not always transgress into a better outcome. Be forensic on the data needs as data becomes one of the highest cost factors in your ecosystem.”

This selective approach to data becomes particularly important during periods of market volatility, as Totten highlights. “The most valuable insights tend to emerge during periods of market stress – precisely when liquidity providers generate the highest volume of data,” he explains. “Just as a desk head must ensure their trading desk operates well under stress, they must also ensure that their data platform can cope with large peaks, both ensuring that there are no gaps or anomalous data, and also that the data is processed rapidly for quants and traders to do any required investigations or optimisations.”

The technical solutions to these challenges continue to evolve, with Lambert noting several strategic approaches. “As data volumes increase, traditional storage solutions become costly and inefficient,” he explains. “Cloud storage offers scalability, allowing us to expand capacity as needed. By using tiered storage, we can differentiate between frequently accessed and archived data, optimizing storage costs.”

The transport of massive data volumes presents its own set of challenges. “The volume of data we consume could introduce latency. So we use data compression to improve transport efficiency, while ensuring compliance with data governance regulations,” Lambert adds. “By providing an integrated and diverse data set from various platforms we can reduce our customers’ data requirements and provide them the data using reliable electronic connections and using APIs to ensure the smooth integration across systems.”

Beyond storage considerations, data security and analysis capabilities have become increasingly critical. “At State Street, we are heavily focused on securitization of data, ensuring that the right data is delivered to the right hands,” Solinski emphasises. 

Perhaps the most persistent challenge in FX data utilization stems from the fragmentation between different market segments, where varying levels of electronification create distinct data ecosystems.

“The limitation of the available data is not directly related to the data distribution capabilities, as this is the same for any trading paradigm, though different trading paradigms will lead to different availability,” Joris explains. “Let us take swaps, on the interbank markets a lot more factors are more pronounced which influences the price creation in comparison to spot. This ranges from dates (balance sheet driven), credit (client), skew (positions), adjustment factors on terms and ccy (risk profile). So, it is not easy to determine a central tape of pricing data as out of trading data there is always a different angle.”

This complexity is compounded by legacy system limitations, as Lambert notes. “Many institutions still rely on legacy systems that are difficult to modify. These rigid structures make integrating diverse data sources challenging.”

However, solutions exist: “Upgrading legacy systems can be overwhelming, but working with third-party technology providers helps firms integrate external data sources efficiently. To overcome data silos, firms must adopt more agile, scalable data ecosystems that support diverse data sources and advanced analytics.”

Totten points to an encouraging trend toward greater electronification across previously opaque market segments. “While it’s true that certain products suffer from poor data quality, this is largely tied to their level of electronification – a trend that will continue to improve over time,” he observes. “Market understanding and expertise is key, as one needs to know when and when it’s not appropriate to use certain data sets, and the potential data quality issues that come with them.”

Banks occupy a unique position in this ecosystem, with potential advantages in certain market segments. “Banks have an added advantage in access to swap and FWD pricing,” Solinski explains. “While they do have the information, they are investing in their architecture to stream these quotes on an ESP basis. However, this investment has multiple benefits as it will ultimately underpin the future streaming environment that Swaps and FWDs will eventually evolve towards.”

This evolution toward more transparent and integrated data systems across all FX market segments represents the next frontier in market development. “Platforms like GlobalLINK are uniquely positioned as they host both buy- and sell-side data,” Nurbhai notes. “Due to their independence, TCA venues have the potential to serve as key vehicles for advancing data democratization for the buy side.”

The path forward requires both technological innovation and strategic partnerships. As Joris concludes, “The need for a clean neutral data curve without these influences is the key to price creation. This is where a lot of work is needed to create such datasets either from participants or from trading venues.”

Building the data-driven FX organisation

The journey toward becoming a data-centric institution involves more than just technology upgrades. It requires fundamental shifts in organisational culture, processes, and skills. As FX market participants navigate this transformation, they must balance strategic vision with practical implementation steps to maximize the value of their data assets.

“What will drive the use is traders adopting the mindset that data and electronification is there to help and can achieve better outcomes,” Joris explains. “The shift from execution to risk management, this will be the unlocking factor to a data-centric operation.”

This mindset transformation must extend throughout the organisation, affecting how different functions approach their roles. Totten highlights how this reshapes incentive structures and responsibilities. “Traders can look at market microstructure and execution strategies to work out how to better optimise their pricing and hedging. Salespeople, rather than being incentivised purely on volume, can be incentivized on PnL, and must thus understand what makes certain types of flow good or bad,” he observes.

This comprehensive shift “is not a simple transformation and will likely require an organisation to either hire specialised staff, or work closely with a business aligned vendor.”

The path toward data-centricity begins with strategic commitment, as Lambert notes. “Firms that successfully use data have in our experience taken a strategic decision that data will support every decision they make,” he states. “Some parts of foreign exchange have a long history of being data driven, such as quantitative trading strategies, while for others the use of data often remains clunky and disjointed.”

This organisational transformation requires structured governance and clear accountability, according to Solinski. “We have restructured our data to be centrally located, enabling us to leverage it in a strategic and powerful way,” she explains. “We have established protocols that are strictly followed to ensure we use only permitted data. This is important, as data comes with associated rights. Additionally, we have designated owners for data technology and functions, ensuring accountability in how and when data is utilized.”

The implementation of these strategic visions demands practical steps for normalising and structuring data flows. While many firms are turning to external partners for technical infrastructure, Totten cautions about potential tradeoffs. “Building and maintaining adapters to venues is something that more and more institutions are outsourcing to vendors that can deliver the critical low latency and infrastructure management requirements they have,” he notes. “However, normalisation can lead to loss of some critical venue specific information, such as if an ECN is tagging prices in a certain way to allow for more optimised execution.”

The fragmentation of data sources within organisations presents another significant challenge. “We still see disjointed structures within trading firms where the unfiltered exhaust of trading or broking activity is offered up as a source of market data,” Lambert observes. “Sometimes errors and omissions can only be seen when checked against another source.”

A centralised approach to data organisation emerges as a crucial element for maximizing value. “Trading firms should implement standardized tools to ensure their transaction and market data are clean, structured and optimized,” Solinski explains. “We have centralized this data, mapping values of a single trade through the whole trade cycle, and structured technology around this data to create efficiency at every point of the trade.”

Realistic timeframes for these transformations vary widely depending on organisational factors, but external partnerships can accelerate progress. “Managing and consuming data takes skill and commitment, and its complexities should not be underestimated,” Lambert notes. “Working with experienced and customer focused partners can certainly reduce the timeframe and costs of developing a successful data strategy.”

While the investment requirements are substantial, the market increasingly offers accessible paths to data transformation. “There are however third party data technology providers who democratize access to sophisticated use of data,” Solinski concludes. “We champion these as we share the same strategy- to deliver best trade outcomes for FX market participants.”

The value of specialised expertise and intuitive tools

The evolution of FX data management platforms has profoundly transformed how market participants interact with and extract value from their data. These next-generation solutions combine sophisticated analytical capabilities with increasingly intuitive user interfaces, making powerful data science accessible to non-technical users.

“There was an old belief that traders would evolve into ‘fincoders’, with the ability to write code and create their own models and data sets,” Joris observes. “This is rapidly changing, and we now have LLMs and AI enabling traders to skip the code and just use natural language to get to the required information. This is a fast-moving market and will be the future in making it more intuitive to the end consumers than direct programming skills.”

This shift toward natural language interfaces represents just one aspect of how solution providers are reimagining user experience. Totten highlights how deep workflow integration has become a central focus. “Our analytics are designed with this in mind, ensuring that users have seamless access to critical information,” he explains. “For example, ahead of a customer call, they can view all key data alongside the most relevant drill-downs to quickly address customer inquiries. Additionally, our platform highlights data points that may reveal opportunities to expand share of wallet, empowering users to make informed, strategic decisions with ease.”

The effectiveness of these front-end innovations remains intrinsically linked to robust back-end data management practices. “Our experience is that backend data management and frontend tools go hand in hand,” Lambert notes. “If your data is not stored, organised and consolidated efficiently, then it’s very hard to build good front-end tools that work well. But if you get both your data management and front-end tools right then the outcomes can be spectacular.”

When selecting data management partners, domain expertise emerges as a crucial differentiator. The unique complexities of FX markets demand specialised knowledge that goes beyond general data science capabilities.

“FX domain knowledge remains a key factor as understanding the market, FX data structure, impact and correctness is not just a numbers game,” Joris explains. “Real value lies in the firms which can make the leap between what is the deep domain knowledge of a system code and what is the finesse of the FX market.”

This domain expertise becomes particularly valuable when firms are navigating business transformation. “Many make the error where they try to implement a manual process into an electronic system, but this defeats the purpose and the value of electronification,” Joris adds. “Electronification means trading model changes on the client, sales and trading desks. Partnering with people who can help you to do this is key to its success.”

The advantages of specialised partners extend beyond implementation support to ongoing innovation and insight generation. “There are many reasons for working with sophisticated vendors in this space, including the setup and management of a robust data pipeline, the ongoing development of business focussed analytics, and our work on increasingly precise, actionable outcomes,” Totten notes.

This commitment to specialised expertise drives strategic investment decisions. “oneZero’s recent acquisition of Autochartist – a leader in market-data driven technical analysis and content is a demonstration of our belief and commitment to the real value that analytics can bring to businesses,” Totten adds.

The interpretation of data presents another area where domain expertise proves invaluable. “We see many firms offering reams of data, but with no relevant expertise in the underlying market, it is difficult for those firms to ensure they are handling the data correctly or to offer much-needed value-added services like advanced analytics using that data,” Lambert observes. 

“It is important therefore not just to get the raw data, but to have confidence that it provides the information you need and expect and to be able to query the context.”

Beyond technical capabilities and market knowledge, institutional stability and regulatory compliance represent additional considerations when selecting data partners. “When choosing a partner for your data and trading ambitions, it is important to collaborate with a trusted partner who can easily integrate with your existing ecosystem,” Solinski explains.

“Being part of a Globally Systemically Important Financial Institution (G-SIFI) institution, GlobalLINK adheres to the highest of data protocols in the industry. There is strict legal and regulatory framework in place to protect the use of all data that transmits across our technologies.” This foundation of trust becomes increasingly important as firms entrust sensitive information to external partners. “Our technologies are used by large investment corporations holding highly sensitive and valuable information,” Solinski concludes. “Our investment in technology and innovative approach will help our clients move with confidence as their data is safe and secured.”

The AI-powered future of FX data management

As data management capabilities continue to evolve in the FX market, artificial intelligence and machine learning are reshaping how firms extract value from their information assets.

Totten highlights this shift from specialised applications to mainstream adoption: “Where we’ve seen great steps forward recently is around the sophistication of language models.” This evolution enables more intuitive data interactions, where “a salesperson could have an LLM agent listening on a customer phone call or monitoring a chat to actively query and produce relevant reports on the fly, transforming how trading professionals access and utilise analytical insights.”

Despite technological advances, the fundamental principle of data quality remains paramount. “Having spent many years building alpha generation models in foreign exchange, I learnt the powerful lesson of ‘garbage in garbage out’,” Lambert emphasises.

This reality underscores that the sweeping changes that are brought forward by AI and ML will be most richly reaped by firms that have the best understanding of their own data” and who partner with organisations providing rich, accurate contextual information.

The practical implementation of these technologies is already delivering concrete benefits across the industry. As Lambert notes, “At NCFX our data is already feeding AI and ML tools both internally and for our clients who are garnering great understanding of the environment that they are operating in.”

As the market continues its data-driven transformation, organisations that successfully maintain high-quality data foundations while embracing new technological possibilities will capture the competitive advantages that AI-enhanced data management can deliver.