Most people who open a trading account spend their research time on the wrong things. They compare advertised spreads, hunt for the lowest commission, and read a dozen reviews about withdrawal speed.
What almost nobody examines is the machinery sitting between the click and the market — the routing logic, the pricing engine, and the liquidity aggregation software that decides, in a handful of milliseconds, what price an order actually receives. That layer is invisible by design, and over a few hundred trades it can move an investor’s results more than any published fee schedule.
Two very different things can happen after you press “buy”
When a retail order reaches a broker, it takes one of two broad paths. In the first, the broker passes the order out to external liquidity providers — banks, non-bank market makers, ECNs — and earns a markup or commission on the fill. This is usually described as A-book execution. In the second, the broker keeps the position on its own books and becomes the counterparty. That is B-book, and it is neither illegal nor automatically predatory; it is how a great many regulated firms handle small, uncorrelated retail flow.
The important point for an investor is not that one model is virtuous and the other is not. It is that the two models create different incentives, and that most brokers run both at once, sorting clients and orders between them according to rules the client never sees. A trader who is consistently profitable may be routed externally. A trader who is not may stay internal. Those decisions are made automatically, by software, according to parameters the broker configures.
The cost that never appears on a statement
Commission is easy to audit because it is printed. Execution quality is not. It shows up as slippage on entry, as a rejected order during a news release, as a fill a fraction of a point worse than the quote that was on screen. None of it is itemised anywhere, and none of it feels like a fee.
This is why the plumbing matters. A broker aggregating prices from several providers can usually construct a tighter, deeper book than one relying on a single feed, and can keep quoting when one source widens or drops out. A broker running a thin setup has fewer options when volatility arrives — which is precisely when retail orders tend to cluster. Two firms can advertise an identical spread and deliver materially different outcomes because of what sits behind the quote.
Why infrastructure has become a specialist industry
Very few brokerages build this stack themselves. The economics do not work: connecting to liquidity providers, normalising their feeds, managing failover, enforcing risk limits and staying compatible with each platform update is a permanent engineering commitment, not a one-off project.
Instead, an entire B2B sector has grown up to supply it. Vendors such as Takeprofit Tech build bridges, execution modules and risk tooling that brokers license and configure, in much the same way that a bank licenses a core banking system rather than writing one. For investors, the practical consequence is that the technology gap between a large brokerage and a small one has narrowed considerably. A modest firm running well-chosen infrastructure can execute perfectly well; a large firm running neglected infrastructure can execute badly. Brand size is a poor proxy for quality of fill.
Five questions worth asking before funding an account
None of this requires an investor to become a systems engineer. It does mean the standard due-diligence list is incomplete. Before committing capital, it is reasonable to ask:
Who regulates the entity that holds my money? Not the group, not the marketing brand — the specific legal entity on the account agreement, and the jurisdiction it answers to.
Is client money segregated, and where? Segregation at a well-capitalised bank in a strong jurisdiction is not the same thing as segregation in name only.
What is the execution model for my account? Many brokers will answer this directly if asked in writing. A refusal to answer is itself information.
What happens during high-impact news? Ask specifically whether leverage is reduced, whether orders may be rejected, and how stop orders are treated during gaps. These policies are usually documented, just not prominently.
Is there negative balance protection? In fast markets a leveraged account can theoretically go below zero. Whether the broker absorbs that is a written policy, not a courtesy.
Testing the answers cheaply
Documentation describes intent; live behaviour reveals practice. A small funded account traded at realistic size for a few weeks produces a usable sample. Log the quoted price at the moment of each order and compare it to the fill. Look at whether slippage runs symmetrically in both directions or consistently against the position. Note how the platform behaves in the minutes around a scheduled economic release.
A broker whose fills are broadly symmetrical and whose platform stays responsive under load is telling you something real about its infrastructure. One that slips consistently in one direction, or that becomes unavailable exactly when positions need managing, is telling you something equally real.
The takeaway
Execution quality is a genuine, recurring cost of investing, and it is one of the few that compounds silently. Investors are trained to scrutinise expense ratios on funds down to a basis point, then open a leveraged trading account on the strength of a headline spread. The technology behind the quote is not a technical curiosity — it is part of the product being bought, and it deserves the same scepticism as anything else on the fee page.