Spreads are the least of it. Depth, fill behavior, failover, and commercial terms — how to evaluate an LP like an operator, not a shopper.
BridgingFX editorial · 2026-09-01
Every liquidity provider's sales deck shows the same thing: tight spreads on EURUSD during London hours. Every broker's live experience shows something else: what happens to those spreads at rollover, during NFP, and when your biggest client hits the book at 3am. Choosing an LP on headline spreads is like choosing a car on its top speed. Here's how operators actually evaluate liquidity.
You're not buying prices — you're buying fill behavior. The questions that matter: what's the average slippage on market orders during normal and stressed conditions? What percentage of orders get requoted or rejected? How does the feed behave in the five minutes around major news? An LP with 0.1-pip spreads and 15% rejection rates is more expensive than an LP with 0.3-pip spreads and 99% fills.
Ask for real execution data, not marketing spreads. Serious LPs will share slippage distributions and fill-rate statistics. The ones who won't — or who only show London-session snapshots — have told you everything you need to know.
Top-of-book spread is a vanity metric for anyone doing real volume. What matters is depth: how much can your clients trade before the price moves against them? A thin book with a tight top looks great on a comparison table and bleeds your clients through slippage on every meaningful order.
Evaluate the full depth ladder across your most-traded symbols, and do it across sessions — Asian hours depth is where retail brokers discover whether their LP is real. If your target market trades during hours your LP's book is thin, you'll pay for it in client complaints and churn.
Volume commitments, tiered pricing, setup fees, monthly minimums, termination clauses. The LP business model runs on commitments, and the standard trap is signing volume tiers you can't fill in your first year — then paying penalty rates that quietly erase your markup.
Negotiate like a startup, not like the enterprise you hope to become: short initial terms, realistic tiers, clean exit clauses. The best LP relationship is one you can leave — paradoxically, that's when LPs work hardest to keep you.
Single-feed brokers learn about redundancy the hard way, always during the one event they can't afford. Your liquidity architecture needs at least two independent feeds with automatic failover, and you need to have tested that failover before you need it — not during.
Ask prospective LPs how their feed behaves when their own upstream fails. Ask your technology partner how fast failover switches and what clients experience during the switch. Then run a drill. The brokers who survive flash events aren't the ones with the best spreads; they're the ones whose second feed was already warm.
Run a structured trial: same symbols, same sessions, same order types across two or three candidate LPs, for at least a month including one major news cycle. Measure slippage, fill rates, rejections, and effective spread after markup — per session, per symbol.
Then talk to their existing broker clients — not the logos on the website, but operators at your size. Ask about support responsiveness at 2am, about how disputes are handled, about what changed after the contract was signed. Liquidity is a relationship business wearing a technology costume, and the costume comes off in the first crisis.
Finally, re-evaluate annually even when things work. Markets evolve, your flow profile changes as you grow, and the LP that was perfect at $50M monthly volume may be wrong at $500M. The brokers with the best long-term execution costs treat LP selection as a recurring discipline — a yearly trial against the incumbent — not a one-time decision made at launch.
Evaluate LPs on fill behavior and depth, not headline spreads. Negotiate commercial terms you can actually grow into, build failover before you need it, and trial with real data across real sessions. The cheapest liquidity is the kind your clients never have to think about.
Let's scope it — honestly, and in writing.
