The three models defined
Every broker has to answer one structural question before it prices a single trade: when a client clicks buy, who ends up on the other side, and where does the price come from? The three classic answers — STP, ECN and market maker — describe different plumbing, different economics and, crucially, different alignment of interest between the broker and its clients.
- STP (Straight-Through Processing). The broker forwards client orders directly to one or more external liquidity providers for execution, without a dealing desk deciding trade by trade. Fills come back from the provider that priced them. The broker is a routing layer that earns from a markup or commission on flow, not from client losses.
- ECN (Electronic Communication Network). The broker connects clients to a shared venue where many participants — banks, non-bank market makers and, in some venues, other clients — submit orders that match anonymously at the best available bid and offer. Pricing is transparent and multilateral; the broker typically charges a flat commission per side and does not set the price.
- Market maker (dealing desk). The broker quotes its own two-way prices and takes the opposite side of client trades on its own book, internalising the risk. It may hedge externally, but the client's counterparty is the broker itself. Revenue comes from the spread and from the net trading result of the internalised book.
These are not marketing labels so much as descriptions of where risk and price formation sit. Understanding them is the foundation for any conversation about execution quality, cost and disclosure.
STP vs ECN vs market maker, compared
The three models diverge on who the counterparty is, how the price is formed, how the broker earns, and whether the arrangement carries a conflict of interest. The table sets them side by side.
| Dimension | STP | ECN | Market maker |
|---|---|---|---|
| Counterparty to the client | External liquidity provider(s) | Other venue participants (multilateral) | The broker itself |
| Price formation | Streamed provider quotes, often aggregated | Anonymous order matching at best bid/offer | Broker-quoted, may reference external feeds |
| Broker revenue | Markup on spread or commission on flow | Flat commission per side; raw spread | Spread plus net result of the internal book |
| Conflict of interest | Low — broker does not profit from client losses | Low — broker is not the counterparty | Structural — broker gains when clients lose |
| Typical book type | A-Book | A-Book | B-Book |
| Operational complexity | Moderate | Higher (venue, matching, reporting) | Lower to run, higher to risk-manage |
| Best-fit broker | Growing brokers wanting neutral execution | Volume-driven, latency-sensitive clients | Firms with strong risk management and disclosure |
Structural comparison. Real-world setups often blend elements — an ECN feed inside an STP model, or an internalised book hedged externally — so treat the rows as tendencies, not absolutes.
How each relates to A-Book vs B-Book
Execution model and book type are frequently conflated, but they answer different questions. A-Book vs B-Book describes what a broker does with the risk: pass it to an external counterparty (A-Book) or retain it internally (B-Book). STP, ECN and market maker describe the mechanism and venue through which orders are executed. The two dimensions line up, but they are not the same axis.
In practice, STP and ECN are the delivery mechanisms for an A-Book: the risk leaves the broker and lands with a liquidity provider or venue counterparty. Pure market-making is the mechanism for a B-Book: the risk stays on the broker's own book. This is why an operator can honestly describe itself as "A-Book" while using either an STP bridge or an ECN venue underneath — the label refers to risk, the plumbing can differ. For the full treatment, see A-Book vs B-Book: how brokers manage risk.
Where liquidity comes from in each model
In an STP model the broker sources streamed prices from external providers and routes to them; the depth and reliability of those providers determine the client's experience. In an ECN model liquidity is contributed by the venue's participants and consolidated into a single order book, so the venue's membership defines its depth. In a market-maker model the broker is the immediate source of liquidity, quoting from its own book and hedging residual exposure in the market as it sees fit.
For A-Book operators, the practical question is how to reach deep, diversified liquidity without holding a dozen bilateral bank relationships. That is precisely what a Prime of Prime provides: aggregated tier 1 bank and non-bank feeds delivered through one connection and one credit line, suitable for either an STP or an ECN-style setup. The mechanics of combining those feeds are covered in how FX liquidity aggregation works.
Conflict of interest and disclosure
The sharpest distinction between the models is alignment. In STP and ECN the broker does not win when the client loses, because the broker is not the client's counterparty — its revenue is the markup or commission on activity, regardless of the client's outcome. In a pure market-making model the broker holds the opposite position, so a client's loss is, mechanically, the broker's gain. That is not inherently improper, and well-capitalised firms run internalised books responsibly, but it is a structural conflict that must be understood, risk-managed and disclosed to clients and, where applicable, regulators.
Related execution practices bear on trust as well. Firm versus market-maker pricing, last-look rejection and the transparency of markup all shape how a given model actually treats client orders, independent of the label on the tin.
Cost to the end client: spread vs commission
The models also present cost differently. Market makers and many STP brokers embed their revenue in the spread — the client sees a slightly wider bid-offer and pays no separate fee. ECN brokers, and STP brokers running a raw-spread account, typically pass through a tight or raw spread and charge an explicit commission per lot or per million traded. Neither is automatically cheaper: a tight raw spread plus commission can beat a wide all-in spread, or not, depending on volume, instrument and the underlying liquidity. What matters to a client is the total cost per trade and the consistency of fills, not the pricing label.
Which model suits which broker profile
There is no universally correct answer. An emerging broker that wants neutral execution and a clean conflict profile often starts with STP against aggregated liquidity. A firm serving active, latency-sensitive or higher-volume clients may favour an ECN for its transparent, multilateral pricing. A well-capitalised operator with strong risk management may internalise flow as a market maker to capture spread and net trading results, accepting the conflict and disclosure obligations that come with it. Most established brokers do not choose one exclusively — they run a hybrid book, routing some flow externally and internalising the rest. Whatever the mix, the A-Book portion still needs deep, reliable liquidity, which is where a Prime of Prime relationship fits. The glossary defines the supporting terms used here.