Rule 606: separate routing payments from execution results
A routing payment can identify an economic incentive without measuring what a customer lost. Rule 606 research becomes useful when the reader separates the order population, routing destination, payment base and contractual conditions. A percentage of routed orders is not automatically a percentage of executed shares.
This guide concerns U.S. Regulation NMS disclosures. It uses the SEC Trading and Markets staff’s Rule 606 FAQ, including its background and questions 11.01–12.02, read on October 26, 2026. The staff page is interpretive guidance, not a new Commission rule. Every broker, venue, order and monetary amount in the examples is fictional.
Identify which report answers the question
The FAQ background distinguishes public quarterly Rule 606(a) reports from customer-specific not-held-order reports available on request under Rule 606(b)(3), subject to applicable exceptions. They are different evidence objects.
For the public report, start with broker identity, quarter, month within the report, security grouping and order category. Preserve the distinction between S&P 500 stocks and other NMS stocks and between market orders, marketable limits, non-marketable limits and other orders where applicable. Do not combine stock and options units.
Write down whether a percentage concerns non-directed orders and exactly which category supplies its denominator. A venue share within one category cannot be compared directly with another broker’s all-category total. A report may contain several legitimate percentages for the same venue because the underlying questions differ.
Retrieve the original report and its relationship narrative together. A screenshot of payment rates without the notes can omit minimum volumes, profit sharing, rebates or conditions needed to interpret the economics. Keep the retrieval date, file identity and any corrected version.
Work a conditional payment calculation
Assume a fictional broker’s stated routing percentages are 60% to Venue A and 40% to Venue B for one compatible category. The example adds an explicit assumption: those percentages also equal the allocation of 100,000 eligible executed shares. This extra share assumption is not established by the public order percentages.
Assume the rates are payments received by the broker, with no additional fees or adjustments: A pays 0.20 currency units per 100 eligible executed shares, and B pays 0.05 per 100. All shares meet the relevant payment terms.
| Venue | Assumed eligible executed shares | Payment per 100 shares | Conditional payment |
|---|---|---|---|
| A | 60,000 | 0.20 | 120 |
| B | 40,000 | 0.05 | 20 |
| Total | 100,000 | — | 140 |
The arithmetic is 60,000 / 100 × 0.20 + 40,000 / 100 × 0.05 = 140. The blended rate is 140 / 100,000 × 100 = 0.14 per 100 eligible shares.
Neither the 140 nor the blended rate is an observed broker revenue total. They are conditional outputs from the additional execution-share assumptions. If the actual eligible quantities, fee offsets or rate conditions are missing, retain an unknown rather than presenting the calculation as measured income.
Also preserve direction. A payment received, an exchange fee paid and a net rebate are not interchangeable positive amounts. Confirm the report’s sign convention before aggregating, and retain gross components when available so a net total does not conceal offsetting charges.
Show why an order percentage does not identify shares
Consider a separate counterexample with ten routed orders: six to A contain 100 shares each, while four to B contain 1,000 each. A receives 60% of orders but only 600 / 4,600, approximately 13.04%, of routed shares.
Assume all of these shares execute and qualify for the earlier rates. A’s payment would be 1.20 and B’s 2, totaling 3.20. Multiplying 4,600 shares by the previously calculated blended rate of 0.14 per 100 would instead produce 6.44 and be wrong for this mix.
Partial fills create another gap: routed shares and executed shares may differ. An order sent but not executed can affect a routing population while contributing no execution-based payment. Obtain the eligible payment quantity directly when possible; the order-share conversion cannot be repaired by assuming every order has the same size.
Reconstruct the tier before interpreting the marginal incentive
FAQ 12.02 addresses disclosure of material tiered-pricing terms. A heading such as “volume incentive” does not reveal whether a higher rate applies only to extra shares or retrospectively to the whole eligible base.
Assume a purely fictional incremental schedule: A pays 0.20 per 100 on the first 60,000 eligible shares and 0.30 per 100 on the next 10,000. At 70,000 shares, payment is 120 + 30 = 150. The extra 10,000 generate 30.
Under a different hypothetical all-volume schedule, crossing the threshold makes 0.30 apply to all 70,000, producing 210. Relative to 120 at 60,000, the increase is 90. These are distinct contracts, not two ways to calculate the same disclosed rate.
Record the threshold, measurement window, eligible order types, retrospective treatment and fee offsets. Missing terms prevent a reliable marginal-payment calculation. Neither schedule establishes that routing actually changed because of the incentive; it identifies a mechanism that needs further evidence.
Follow an intermediary to its actual role
FAQ 12.01 distinguishes a broker that executes orders from a routing broker that does not execute them. In the latter scenario, the downstream execution venues are relevant, and relationship disclosure and incorporation by reference can matter.
Build a small role map: customer-facing broker, routing service, executing venue and economic payment link. Do not call an intermediary the execution venue merely because the customer-facing broker sends an order there. Conversely, a broker can itself be an executing destination in the circumstances described by the FAQ.
Avoid adding a customer-facing broker’s payment from its routing provider to the provider’s separate venue payment as though both were one recipient’s revenue. They concern different legal relationships. A brand match or identical monetary figure is not a supported consolidation rule.
Execute a routing-economics worksheet
| Check | Input to preserve | Stop condition |
|---|---|---|
| Report object | Public quarterly or requested customer report | Different objects mixed |
| Population | Stock group, month, order type and direction | Percentage denominator unclear |
| Role | Customer broker, router or executing venue | Destination role unsupported |
| Payment | Received/paid sign, amount and unit | Net and gross terms confused |
| Quantity | Orders, routed shares or eligible executed shares | Share conversion unavailable |
| Tier | Threshold, window and marginal/all-volume rule | Terms too incomplete to calculate |
| Outcome | Separately matched execution evidence | Incentive mistaken for customer loss |
The deliverable should keep observed fields, disclosed contract terms and analyst assumptions in separate columns. If sources conflict, first compare category, month, venue role, signs and corrections. A narrative annual relationship description and a monthly reported amount may address different periods rather than contradict each other.
Execution quality requires separately matched evidence, such as the appropriate Rule 605 population or customer order records. A payment of 140 does not establish an execution loss of 140, a spread cost of 140 or that another routing choice would have improved results. Public aggregates generally do not identify that counterfactual.
Update when a corrected report appears, the payment schedule changes, an intermediary’s role is clarified or compatible execution quantities become available. Preserve the earlier assumptions and recompute the conditional amounts. The conclusion can identify disclosed incentives and unresolved outcome questions without manufacturing measured customer harm.
Preserve the routed child order’s classification
The FAQ’s questions 11.01 and 11.02 discuss assessing marketability for the order routed to a venue and, when a parent is split, the routed child rather than simply the original parent. A parent-level label therefore cannot automatically classify every destination row.
Assume a fictional parent becomes two routed children at different times. The first is marketable against the applicable quote and the second is not. Putting both into one category because the parent originally appeared marketable would distort the category’s routing share and any payment comparison attached to it.
Save the available classification basis and reference time. If the public report does not expose the child-level records, do not claim to have independently rerun that classification. The appropriate result is a check of the reported categories and their documented definitions, with the unavailable underlying inputs identified.
FAQ
Can 60% of routed orders be treated as 60% of executed shares?
Only with additional support. Order sizes, partial fills and eligibility can differ.
Does payment for order flow equal the customer’s execution loss?
No. The payment and the customer’s outcome are separate quantities requiring different evidence.