Where Your Broker Actually Sends the Order

You click buy on your trading platform. The order fills instantly at a price close to what you expected. You assume the transaction went directly to the market, matched with another trader somewhere on the other side of the world. Most traders never ask where that order actually went or who handled it. The path between your click and your fill is invisible by design, and it varies dramatically by broker type and asset class. Understanding where your orders go isn’t paranoia. It’s basic due diligence, because the routing choice directly affects execution quality, cost, and whether your broker profits when you win or when you lose.

The Invisible Journey Between Click and Fill

When you click buy or sell, your order doesn’t teleport to some abstract market in the sky. It enters a routing system you never see, passing through intermediaries whose names you may not know and whose incentives don’t always align with yours. The path matters more than most traders realize.

Your broker has options when handling your trade. In equities, the order might go to a market maker like Citadel Securities, which executes nearly half of all US retail stock volume. In forex, the broker might send it to a liquidity provider—or simply take the other side of your trade themselves, a practice known as B-Book operation. Crypto exchanges route through internal matching engines, external liquidity pools, or split large orders across multiple venues using smart order routing. Each choice affects what you pay and how fast you get filled.

Execution quality isn’t binary. It’s not just whether your order fills, but at what price, with how much slippage, and after what delay. A market order might fill instantly but cost you three extra pips because your broker chose the venue that paid the highest kickback rather than the one offering the tightest spread. That kickback, called payment for order flow in equities, can run $0.001 to $0.003 per share. Robinhood collected $331 million from this practice in 2020, representing 81% of its revenue. The SEC later found that Robinhood’s routing choices cost customers $34.1 million in inferior execution over five years.

How different routing models affect who takes risk and who profits
Routing Model	Who Takes Market Risk	Revenue Source
A-Book (Agency)	External liquidity provider	Commission or markup
B-Book (Market Maker)	Your broker	Your losses (net)
Payment for Order Flow	Market maker (e.g., Citadel)	Rebate from market maker
Direct Market Access	Counterparty on exchange	Commission only

The table shows why this matters: when your broker profits from your order flow rather than from transparent commissions, their incentive shifts. In a B-Book model, common in retail forex, the broker literally wins when you lose. That doesn’t make every trade rigged, but it does create a structural conflict that pure agency models avoid.

Regulations like Regulation NMS require brokers to provide “best execution,” meaning the most favorable terms reasonably available. That standard sounds reassuring until you try to verify it. What does “reasonably available” mean when you don’t know which venues your broker can access? How do you compare execution quality across brokers when the data isn’t public? You can’t, which is why most traders never realize they’re getting suboptimal fills. The journey between your click and your fill remains invisible by design.

Payment for Order Flow: The Revenue Stream You Fund

When a broker advertises zero-commission trading, they’re not offering charity. They’ve simply moved the cost somewhere you can’t see it on the trade confirmation. The revenue comes from selling your order to the highest bidder before it ever reaches the market.

Who Pays and Why

Payment for order flow works like this: you click buy on 100 shares, and your broker immediately auctions that order to market makers like Citadel Securities or Virtu Financial. The winning bidder pays the broker a fraction of a cent per share—typically between $0.001 and $0.003—for the right to execute your trade. The market maker profits by capturing the spread between the bid and ask, and the broker collects millions in aggregate from routing millions of retail orders. Citadel alone handles roughly 47% of all US retail stock volume, which gives you a sense of the scale.

The conflict is structural. Your broker has a financial incentive to route your order to whichever market maker pays them the most, not necessarily the one that will give you the best price. Regulation NMS requires brokers to seek “best execution,” but that standard is vague enough to allow considerable latitude. Best execution doesn’t mean best price. It means the most favorable terms “reasonably available under the circumstances,” which leaves room for interpretation and, more importantly, for inferior fills.

The Cost of Free Trading

Robinhood’s numbers make the trade-off visible. In 2020, the company collected $331 million from payment for order flow, representing 81% of its total revenue. That same year, the SEC found that Robinhood’s order routing practices had cost customers $34.1 million in price execution harm between 2015 and 2020. The math is blunt: the broker earned nearly ten times more from selling order flow than customers lost in worse fills, but customers still paid.

The table below shows how execution quality can differ even when the nominal commission is identical.

Comparison of execution cost on a 100-share market order
Broker Type	Commission	Avg. Price Improvement per Share	Net Cost per 100 Shares
PFOF-funded (Robinhood-style)	$0.00	$0.008	$0.80
Direct routing (Interactive Brokers)	$0.50	$0.015	$0.00
Traditional broker with commission	$4.95	$0.012	$3.75

The zero-commission broker in this example gives you less price improvement, which means you’re buying at a slightly higher price or selling at a slightly lower one. The direct routing broker charges a small commission but delivers better fills, resulting in a lower total cost. The traditional broker with a flat commission still ends up more expensive, but not by as much as the headline numbers suggest.

For a single 100-share trade, the difference is trivial. But if you’re trading frequently, those fractions of a cent compound. A trader executing 500 trades per year at 100 shares each is quietly paying $400 in hidden costs through inferior execution, compared to $250 in visible commissions at a direct-routing broker. The free model costs more.

This matters less for long-term position trades where you’re holding for months or years. A few cents of slippage disappears into the noise of normal price movement. It matters significantly for active traders, scalpers, and anyone trading size. The incentive misalignment doesn’t vanish just because you can’t see it itemized.

A-Book, B-Book, and the Forex Shell Game

Your broker tells you they offer tight spreads, fast execution, and no conflict of interest. What they rarely mention is whether they’re betting against you.

In forex and crypto, the path your order takes determines who wins when you lose. Some brokers act as pure intermediaries, passing your trade to external liquidity providers like banks or electronic communication networks. They earn a commission or markup on the spread, and your success or failure doesn’t touch their bottom line. This is the A-Book model, named for how these orders appear on the broker’s accounting ledger.

Other brokers become your direct counterparty. When you buy EUR/USD, they sell it to you from their own inventory. When you close at a loss, that loss becomes their profit. This is the B-Book model, and it creates the starkest conflict of interest in retail trading: your broker makes more money when you fail.

The Market Maker Model

B-Book brokers aren’t necessarily fraudulent. They’re market makers in the technical sense, providing liquidity by taking the opposite side of your position. The math works because most retail traders lose. Studies consistently show that 70 to 90 percent of forex accounts end in the red within a year. If the broker can cover the winning minority with the losses from the losing majority, they pocket the difference without the operational cost of routing orders externally.

The problem isn’t the model itself. It’s the incentive structure. An A-Book broker benefits from high trading volume regardless of outcome. A B-Book broker benefits from client losses. One wants you to trade more; the other wants you to trade poorly.

Why Hybrid Models Dominate

Pure B-Book is risky for brokers when a skilled trader appears. Pure A-Book is less profitable when dealing with small, likely-to-lose accounts. So most brokers split the difference, routing trades dynamically based on client profitability.

How brokers route the same trade based on client profile
Client Profile	Routing Destination	Broker Revenue Source
New trader, small account	B-Book (internal)	Client losses
Consistently profitable trader	A-Book (external liquidity)	Markup or commission
High-frequency scalper	A-Book (to avoid risk)	Markup or commission
Average losing account	B-Book (internal)	Client losses

The hybrid model optimizes for broker profit, not client fairness. Algorithms monitor your win rate, position size, and trading style. Win too often and your orders get routed to real liquidity, where the broker earns less but carries no risk. Lose consistently and you stay in-house, feeding the B-Book.

This isn’t a conspiracy. It’s operational efficiency from the broker’s perspective. But it means that becoming profitable can trigger execution changes you never consented to and rarely notice. Your edge, if you build one, quietly shifts the playing field.

ECN and DMA: When You Actually Reach the Market

When a broker advertises “direct market access,” you’re supposed to be trading against the same liquidity pool as hedge funds and banks, not against the broker’s internal ledger. The distinction matters because it changes the fundamental relationship. ECN and DMA models remove the broker from the other side of your trade entirely.

An Electronic Communication Network connects you to multiple liquidity providers—banks, other traders, institutional desks—and your order competes with theirs on price and time priority. You see the depth of market, the actual bid and ask quotes sitting there, and your limit order joins that queue. If someone takes your price, you’re filled. The broker earns a fixed commission per lot or per million in notional value. They don’t care whether you win or lose. They care whether you trade.

Direct Market Access works similarly but typically refers to exchange-traded instruments where your order goes straight onto the order book of a centralized venue. In crypto, this means your limit order sits visible on Binance or Coinbase Pro’s public book. In forex futures, it’s CME’s order book. You’re not being quoted a spread by the broker. You’re seeing the spread that exists between all participants at that moment.

The cost structure flips. Dealing desk brokers offer tight fixed spreads with no commission because they’re betting against you. ECN and DMA brokers charge explicit commissions—often $3 to $7 per lot in forex, or 0.1% to 0.2% per trade in crypto—but the spread floats with actual market conditions. During London open, EUR/USD might trade at 0.4 pips. During thin Asian hours, it might widen to 1.2 pips. That’s not the broker manipulating anything. That’s the market.

Cost structure comparison: dealing desk vs. true market access
Model	Typical EUR/USD Spread	Commission per Lot	Broker’s Counterparty Role
Dealing Desk (B-Book)	1.0 pip (fixed)	$0	Takes opposite side
ECN / DMA (A-Book)	0.3–1.5 pips (variable)	$5–$7	None

The variable spread feels unstable at first, especially if you’re used to seeing the same number every time you open a chart. But fixed spreads are theatrical. Someone is absorbing that variability, and if it’s not you through visible commission, it’s you through hidden slippage, requotes, or the statistical edge of trading against a counterparty that knows your stops.

True market access costs more per trade in transparent fees. It costs less in conflicts of interest. For scalpers and high-frequency strategies, that tradeoff often doesn’t work—the commissions eat too much. For swing traders and position players, paying $6 to know your broker isn’t your opponent changes the entire probability equation. You’re playing the market, not the house.

How Crypto Exchanges Route Your Order

When you click buy on a centralized crypto exchange, your order doesn’t leave the building. Binance, Coinbase, Kraken and their peers run internal matching engines that pair your bid with someone else’s ask, settling the trade on their own ledger without touching a blockchain. The exchange is the venue, the market maker, and the record keeper all at once. No external routing, no third-party market maker taking the other side. Just an order book managed by the platform’s own software, applying price-time priority like any traditional exchange.

Some larger platforms use smart order routing to improve fills on big orders. If you’re buying a meaningful amount of ETH, the exchange might split your order across multiple internal liquidity pools or even route pieces to affiliated venues. This reduces slippage, but it also introduces opacity. Unlike regulated stock markets where Regulation NMS mandates best execution transparency, crypto exchanges operate with far less disclosure. You trust the platform to optimize your fill, but you don’t get a trade-by-trade breakdown of how they achieved it.

Centralized Exchange Mechanics

The matching engine ranks orders by price first, time second. If two traders offer to sell Bitcoin at $43,500, the one who placed the order earlier gets filled first. This is standard, predictable, and computationally fast. The exchange profits from trading fees, not from taking the other side of your trade, which aligns their interest with volume rather than your loss.

But that alignment isn’t guaranteed. Some exchanges operate hybrid models where they internalize certain trades, acting as the counterparty when liquidity is thin. Others lend out customer assets or route flow to affiliated market makers in ways that aren’t disclosed. The lack of regulatory oversight means execution quality varies more than most traders realize.

Decentralized Execution

Decentralized exchanges work from a completely different playbook. When you swap tokens on Uniswap or SushiSwap, there’s no order book and no matching engine. You’re trading against a liquidity pool, a smart contract holding reserves of two tokens. The price you get depends on the ratio of those reserves and the size of your trade relative to the pool. Larger trades move the price more, a mechanic called slippage that’s baked into the automated market maker model.

How your trade gets executed across exchange types
Exchange Type	Execution Mechanism	Transparency
Centralized (Binance, Coinbase)	Internal order book, price-time priority	Limited, proprietary algorithms
DEX (Uniswap, Curve)	Automated market maker, liquidity pool	High, all code on-chain
Hybrid (dYdX, some aggregators)	Off-chain matching, on-chain settlement	Moderate, depends on architecture

The table highlights a curious inversion: decentralized platforms offer more transparency about execution than their centralized counterparts, even though they’re less familiar to traditional traders. Every swap on a DEX is recorded on-chain. You can trace the exact reserves, calculate your slippage in advance, and verify that the smart contract did what it promised. Centralized exchanges give you speed and liquidity, but you’re trusting a black box.

That trust matters more than most retail traders assume. Without regulatory requirements for best execution or trade reporting, a centralized crypto exchange can quietly degrade fill quality, route your order to a worse price, or prioritize high-frequency traders who pay for preferential access. You won’t know unless you’re sophisticated enough to monitor execution quality across multiple venues, and most traders aren’t.

The lesson isn’t to avoid centralized exchanges or assume DEXs are always superior. It’s to recognize that crypto order routing operates with fewer guardrails and more variation in quality than traditional markets. Your fill quality depends not just on market conditions, but on the specific mechanics and incentives of the platform you chose.

Order routing is not a technical footnote buried in the fine print. It’s the infrastructure that determines whether your edge survives contact with reality. Every broker, every exchange, every execution model creates a different set of incentives, and those incentives ripple through to your fill quality, your costs, and ultimately your returns. Most traders never look under the hood. They assume the market is the market, that a buy order is a buy order, and that the only variable that matters is whether their analysis was correct.

But execution quality is part of your edge, or it’s part of your handicap. A trader with a 52% win rate and terrible fills loses money. A trader with a 48% win rate and consistently good execution might break even or better. The difference isn’t skill or discipline. It’s infrastructure.

Choosing a broker means choosing whose incentives you’re willing to trust. A B-Book broker profits when you lose. A payment-for-order-flow broker profits by selling your data and your fills to the highest bidder. An ECN broker profits when you trade, regardless of outcome. None of these models is inherently evil, but they’re not equivalent. The question isn’t which broker has the slickest app or the most aggressive marketing. It’s whose business model aligns with yours.

The next time you place a trade, ask where it’s going. Ask whether your broker routes all clients the same way or segments by profitability. Ask what the total cost of execution is, not just the headline commission. If you can’t get clear answers, that’s an answer in itself. Transparency isn’t a luxury. It’s the minimum standard for anyone treating trading as a game of skill rather than a leap of faith.

Where Your Broker Actually Sends the Order

The Invisible Journey Between Click and Fill