Broker And Venue Execution Scorecards

A fund routing orders through five brokers has, after a year of trading, thousands of individual TCA reports — one per trade, each comparing execution against a benchmark. On its own, that pile of reports doesn't answer the question that actually matters for the relationship: is Broker A consistently cheaper than Broker B once you control for what each was actually asked to trade? A scorecard is the aggregation layer that turns raw, trade-by-trade TCA data into a standing, comparable ranking — typically updated monthly or quarterly — that a fund can actually act on when deciding where to route the next order.

Why raw averages mislead, and what fixes it

The naive approach — average cost per broker across all trades — is badly misleading if brokers aren't handling comparable order flow. A broker who happens to get routed more small-cap, illiquid orders will show worse average cost purely because those trades are structurally harder, not because the broker executes worse. A proper scorecard controls for this by normalizing: comparing each broker's realized cost against a predicted cost from a pre-trade model calibrated for that order's specific size, liquidity, and volatility, so the scorecard reflects performance relative to a fair, order-specific expectation rather than a raw average that conflates broker skill with the difficulty of what they were handed.

Scorecards typically track several dimensions simultaneously, not just cost: fill rate, price improvement versus the quoted spread, speed of execution, and information leakage (does the market seem to move suspiciously in the wrong direction right after routing to this venue, suggesting the venue or a downstream party is trading on the order's presence). A venue that's cheap on average but leaks information systematically may be a worse choice overall than a slightly more expensive venue that doesn't, and a scorecard built on cost alone would miss that entirely.

Worked example: normalizing before ranking

Broker X handled orders averaging $2 million notional in liquid large-caps, realizing 12 bps of cost against a pre-trade model that predicted 10 bps for that mix — 2 bps worse than expected. Broker Y handled smaller, less liquid orders averaging $500,000 notional, realizing 22 bps of cost against a pre-trade model that predicted 25 bps for that harder flow — 3 bps better than expected. A naive comparison (12 bps vs 22 bps) would rank Broker X as clearly superior. The normalized comparison flips the story: Broker Y actually outperformed its own, harder benchmark by 3 bps while Broker X underperformed its easier benchmark by 2 bps — Broker Y is the stronger performer once the difficulty of each broker's actual flow is accounted for.

What this means in practice

Scorecards are the concrete mechanism by which TCA analysis translates into changed behavior — order flow shifting toward better-performing brokers and venues, underperformers being put on notice or dropped, and the whole execution process getting a continuous feedback loop instead of a one-off audit. They're also part of the regulatory record in many jurisdictions: best-execution rules generally require a fund to demonstrate it's actively monitoring execution quality, and a maintained scorecard is direct evidence of exactly that.