Most buy-side firms can say they have transaction cost analysis (TCA). Fewer can say their execution analytics consistently change day-to-day decisions on the trading desk.
The gap usually isn’t about table stakes access to data or the ability to generate quarterly PDFs. It’s about whether analytics are either treated as a cost/compliance artifact that arrives after the fact or a true decision support that protects performance, reduces friction, and improves repeatability in how the trading desk executes.
This article looks at:
Behaviors that trading desks should consider influencing to make the most of TCA.
Three necessary conditions for execution analytics among trading desks that seek to move those behaviors.
While execution analytics can inform a range of front-office functions, we focus on four high-priority areas where traders and heads of trading spend real time: broker reviews, trade planning, strategy selection and internal alignment with portfolio managers. A bit of context for each:
Broker selection (and the conversations that follow)
A broker review is ultimately a behavior-change exercise. For example, what flow should be sent, when, and under what instructions? The analytics need to support a conversation about why one execution outcome differs from another and whether the difference is repeatable. A static ranking of counterparties is not sufficient.
Timing and scheduling
Small shifts in timing—how aggressively to show size, when to step back, and whether to accelerate into liquidity—can result in outsized impacts. This is less about optimizing to a benchmark on paper and more about adapting to market conditions in near real time, based on any dimension of analysis.
Strategy and algo choice
In practice, strategy selection often comes down to a mix of established playbooks, intuition, and precedent with existing processes. Analytics help traders pressure-test those defaults across venues, regions, liquidity regimes, and other characteristics without requiring deep forensic work.
PM/trader communication
Many aspects of execution quality begin upstream: clarity on urgency, constraints, benchmark preferences, and risk tolerance, for example. Analytics can enable progress beyond an abstract post-trade score by revealing what worked, what didn’t, and what inputs mattered most.
Execution analytics must meet three basic condition requirements: confidence in the findings, context around the market regime, and delivery that fits how the desk works. Below is our perspective on how to foster those conditions.
Trading outcomes are variable, which is the first reason many TCA programs fail to change behavior: simple averages create false precision and overconfident conclusions.
A useful mental model is to think statistically, not arithmetically. Ask a group of people to flip the same coin five times each, and they are likely to reach different sets of outcomes even though nothing about the underlying process changed.
The same issue arises when comparing brokers or strategies across a small sample of orders. Given all the variables at play, two averages can differ numerically without being materially distinguishable.
To change behavior, TCA desks need to serve as a source of not just analytics, but evidence. That shows up in practical ways such as:
Separating signal from noise in rankings. Decisions on commission, routing, or algo usage shouldn’t come down to snap judgments based on tidy execution outcomes (e.g., “this number is 0.1 bps better”). Desks need to establish whether a result stemmed from a repeatable difference or random chance.
Working with distributions, not just summary stats. For most desks, there is not an average order. They trade a range of order types in a range of conditions. Looking at the shape of outcomes—where results cluster, where tails appear, how outliers behave—helps avoid overreactions to unusual trades.
Best execution questions help the TCA desk characterize the liquidity landscape it accesses:
Was the session liquid or thin?
Was volatility elevated?
In practice, that context becomes part of the quantitative underpinning for demonstrating and improving best execution, not just explaining outcomes after the fact.
As mentioned, the same score can imply very different execution quality depending on the regime. For that reason, many desks are moving toward relative or difficulty-adjusted benchmarking. That approach normalizes outcomes based on liquidity and market conditions rather than printing an absolute distance-to-benchmark number.
Context matters because markets evolve. Benchmarks that made sense up to 10 years ago may be incomplete as market structure, participant behavior, and venue composition have changed. In Europe, the rise of systematic internalizers provides a useful example. As more trading migrates from lit order books into bilateral/internalized channels, TCA desks must reassess what liquidity they’re accessing, how price formation is occurring, and whether routing decisions introduce different adverse-selection and transparency dynamics compared to traditional venue execution.
A context-aware TCA program typically has three core characteristics:
Multiple benchmarks for multiple dimensions of performance. Different benchmarks answer different questions on implementation shortfalls, arrival, VWAP, and close, for example. Treating one as universal tends to hide tradeoffs.
Benchmarks that evolve with the market. If dynamics such as 24-hour trading, retail participation, or venue behavior change the opportunity set, the benchmarking approach must be adaptable.
If the right analytics arrive late, require too much translation to act on, or land as unhelpful oversight, they are no longer useful. Firms seeking to help their traders make the most of TCA may want to consider the following principles:
Ergonomic presentation
Given traders adopt what they can interpret quickly, analytics should go beyond static tables to include visuals that make it easy to spot patterns, filter by the relevant slice, and drill down when needed. Rapid, decisive action is key. Users should be able to update views immediately and move from thousands of orders to one problematic order in a few clicks.
Transparency around methodology and data gaps
Analytics don’t have to be perfect to be useful. Clarity on what’s included, what’s excluded, where data is missing, and what that means for interpretation fosters trust. If analytics hide errors or silently drop data, traders ignore them.
Fit into real workflows
TCA discussions are seldom confined to one arena. Broker reviews, commission discussions, PM-trader standups, best execution, governance committee meetings, and compliance reporting are examples of the potentially relevant settings. Dashboards should be flexible enough to support any execution-quality conversation and export cleanly into static reports for downstream functions.
Interpretation at scale (AI/LLMs)
As analytics become richer and more multidimensional, AI and LLMs will increasingly have a role to play in translating results into clear initial takeaways. Summarizing key drivers, highlighting what changed, and surfacing what matters can make TCA more actionable with minimal delay.
A final point on the delivery requirement is that successful delivery increasingly includes integration. Execution decisions happen across OMS/EMS workflows, broker touchpoints, and beyond. Open architectures and APIs matter because they let firms place evidence where the organization consumes it across the portfolio lifecycle.
A common tendency is to frame these initiatives as modernizing TCA. However, the best outcomes typically result from narrowing the goal: Pick one business pain point and one decision you want to improve, then work backwards to the evidence required.
For consideration, a practical approach could look like this:
Name the decision. Illustrative examples include broker allocation for a specific strategy, timing guidance for open/close risk, default algo choice for a given order profile, or improving the quality of PM instructions.
Define confidence thresholds. Ask what sample size, variability, or level of distinguishability are required to enable the TCA desk to change behavior.
Define the context required. Thought starters include: What changes in the market are impacting the business, what liquidity measures or volatility states should be adjusted, and which benchmarks best fit the decision?
Design delivery around the workflow. Discern where will the decision be made (broker review, PM meeting, committee) and the most useful format (e.g., interactive views, summaries, or PDFs for memorialization).
Overall, in many organizations the effort is not a mathematical challenge so much as one of implementation and buy-in. In other words, proving value in a way that drives better decision-making.
Execution analytics work best when allowed to be what the TCA desk needs them to be: both a performance tool that improves day-to-day decisions and an audit trail that memorializes best execution for longer-term learning and governance.
The shift beyond TCA is not a shift away from measurement. Importantly, it’s a shift toward measurement that is confidence-aware, context-aware and delivered in a way humans will use. In that context, analytics become an evidence base for better decisions across evolving settings over time.
FactSet TCA delivers a powerful set of capabilities that bring greater context, confidence, and decision support to every trade. Explore the full offering on the FactSet TCA webpage.
This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.