Free trades, checked against only the five buyers who paid for them

MARKET STRUCTURE / BEST EXECUTION

FINRA · Letter of Acceptance, Waiver and Consent · Rule 5310

On 21 July 2026 a FINRA settlement against tastytrade, Inc. became effective: a censure and an $850,000 fine for failing to give its customers' stock orders the "best execution" the rules require. tastytrade, the self-directed retail broker formerly known as tastyworks, had built its routing around a familiar arrangement. Between January 2020 and January 2023 it sent every customer equity order, more than 8.8 million of them covering over 1.7 billion shares, to just five market makers, all of which paid the firm for that order flow. Payment for order flow is legal and widespread. The problem FINRA identified was not the payments. It was that the firm never checked whether those five were actually the best place for its customers' trades to go.

Best execution is one of the oldest duties a broker owes. FINRA Rule 5310 requires a firm to use reasonable diligence to find the best market for a customer's order so the price is as favourable as possible. Where a firm does not review every order individually, Rule 5310.09 requires "regular and rigorous" reviews, at least quarterly, that compare the execution quality a firm is getting against the quality it could get from competing markets. tastytrade did hold a best execution committee, and it did meet quarterly. But the committee only ever looked at the execution quality coming back from the same five market makers the firm already used. It never pulled data on what the venues it was not using might have delivered. The comparison that gives the rule its teeth simply never happened.

FINRA charged two violations from that single gap. The first is the best execution failure itself, under Rule 5310. The second is a supervisory failure under Rule 3110: the firm's systems and written procedures were not reasonably designed to achieve compliance, because they contained no method for comparing against competing markets or for reviewing orders by type. tastytrade consented without admitting or denying the findings, and it had already upgraded its supervisory systems in January 2023. The fine is modest and the conduct is common. What makes the case worth a close read is that the same fact pattern has now caught a long line of retail brokers, and the fix is entirely knowable in advance.

Fine

$850,000plus censure

Period

2020–23three years

Routed to

5 firmsall paid for flow

Charged under

Rules 5310, 3110and 2010

The first violation is of FINRA Rule 5310(a) and Rule 5310.09, the best execution obligation and its "regular and rigorous" review standard. FINRA found three specific deficiencies in how tastytrade ran those reviews. It failed to compare the execution quality of its routing arrangements against the quality available from competing markets it did not use. It failed to review all the relevant execution quality factors the rule lists, in particular price disimprovement, meaning the cases where a customer got a worse price than the best quote available when the order arrived. And it failed to review orders on a type-of-order basis, so it never separated, for example, marketable orders from non-marketable ones, even though those behave very differently and can receive very different execution quality.

The detail that makes the failure concrete is what the committee actually reviewed. It looked at a report containing genuine execution quality metrics, price improvement, fill rates, speed of execution, for the five market makers it used. But the data was aggregated: total shares routed to each of the five, with no breakdown by order type and no figure for price disimprovement, and crucially no comparison at all to venues outside the five. A review can look diligent, generating reports and quarterly minutes, while omitting the one comparison that would reveal whether customers were being underserved. That is what FINRA concluded happened here. A violation of Rule 5310 is also a violation of Rule 2010, the requirement to observe high standards of commercial honour, so the conduct breached that rule too.

The second violation is of FINRA Rule 3110, the supervision rule. tastytrade relied on the best execution committee's reviews as its entire supervisory system for the obligation, and those reviews were not reasonably designed for the reasons above. Separately, its written supervisory procedures contained no procedures for comparing execution quality at competing markets or for reviewing orders by type. In other words, the process gap and the documentation gap reinforced each other: nobody was told to make the competing-markets comparison, and no procedure existed that would have required it. FINRA noted that the firm improved both its supervisory systems and its written procedures in January 2023, which is where the relevant period ends.

The root cause is a subtle one, and it is easy to walk into. tastytrade was not ignoring execution quality. It had the committee, the cadence and the metrics that a compliant program is built from. What it lacked was the external reference point. Reviewing the quality of your five chosen venues tells you how those five compare to each other, but it cannot tell you whether all five are worse than a sixth you never tried. Best execution is inherently comparative, and a review that only ever examines the incumbents has quietly redefined the question from "are these the best venues available" to "are these venues performing consistently". Those sound similar and are completely different.

Payment for order flow is the incentive that makes this blind spot dangerous rather than merely incomplete. The five venues tastytrade used all paid it for the orders. That does not make the arrangement improper, and FINRA did not find that customers were actually harmed or that the payments drove bad routing. But it does create a structural conflict: the firm has an economic reason to keep routing to venues that pay, and the only thing standing between that incentive and the customer's interest is a rigorous, outward-looking review. When the review looks only at the paying venues, the conflict has nothing checking it. The control that is supposed to neutralise the incentive has been pointed away from the thing it exists to test.

The aggregation problem compounds it. By reviewing only total shares routed to each market maker, the committee smoothed away exactly the variation the rule wants surfaced. Price disimprovement events, which are individually small and collectively meaningful, disappear into an average. Differences between order types disappear into a blended number. A marketable order and a non-marketable limit order have different execution profiles, and a firm that does not separate them cannot see that one category is being served well and another poorly. Granularity is not a nicety in best execution review; it is the mechanism by which problems become visible. Aggregate reporting is comfortable precisely because it hides the things a rigorous review is meant to find.

There is also a documentation dimension that turned a practice gap into a second, separate charge. Even if individuals had wanted to run a competing-markets comparison, the firm's written supervisory procedures did not require it or describe how. Written procedures are where an obligation becomes a repeatable, auditable step rather than a matter of individual initiative. When the procedure is silent, the review depends on whoever happens to be in the room understanding that the comparison is required, and that dependency is fragile. FINRA treats the supervisory system and the underlying conduct as distinct failures for this reason: a firm can breach best execution through a one-off lapse, but a missing procedure is what lets the lapse persist for three years.

It is worth being precise about proportionality, because the framing matters. This was a design and review failure, not a finding of dishonesty or of proven customer loss. tastytrade consented without admitting or denying, remediated before the case concluded, and the $850,000 fine sits in the mid-range for this category. But the reason FINRA pursues these cases regardless of demonstrated harm is that best execution is a prophylactic duty. The obligation is to run the comparison that would catch harm, not merely to avoid harm you happen to notice. A firm that never compares has not met the duty even if, by luck, its customers were fine.

Because this fact pattern recurs so consistently, the controls that prevent it are well mapped. Drawing on the control design that RegLabs derives from this case, four gaps come up again and again, and each has a concrete, evidenced fix rather than a policy-drafting one.

A second pitfall is aggregation that hides the signal. Reviewing total shares per venue is comfortable and nearly useless. Execution quality has to be segmented by order type, because marketable and non-marketable orders behave differently, and the analysis has to include price disimprovement, not just price improvement, because the downside cases are where customer harm lives. A practical rule of thumb: if your best execution pack cannot show how limit orders fared differently from market orders, or cannot quantify disimprovement, it is not yet a Rule 5310.09 review.

Third is the payment-for-order-flow conflict left unmanaged. Routing to venues that pay for flow is permitted, but the economic incentive needs an explicit control wrapped around it. That means documenting and justifying each PFOF arrangement, demonstrating that the payments do not compromise execution quality, and re-validating that justification on a set cadence rather than treating it as settled once signed. The point is not to avoid PFOF; it is to be able to prove, on demand, that the routing decision was made on execution quality and the payment was incidental to it.

Fourth is the procedure that never names the step. tastytrade drew a second violation because its written supervisory procedures did not describe the competing-markets comparison or the order-type review at all. The fix is to write the specific, actionable steps into the WSPs: what is compared, against which venues, using which factors, by whom, how often, and how exceptions are escalated and resolved. A best execution committee is only as good as the charter that defines its scope; a documented mandate turns "we discussed routing" into a supervisory system a regulator can test.

Underlying all four is a discipline point that this case makes vivid. A best execution program can have every visible component, a committee, a quarterly cadence, real metrics, polished minutes, and still fail the rule if it never looks outward. The firms that stay clear of this are the ones that periodically ask the uncomfortable question at the centre of the tastytrade case: if a competing venue were consistently beating ours, would our current review actually show it? If the honest answer is no, the review is decorative, and increasingly regulators are saying so.

tastytrade is not an outlier. It is the newest entry in a remarkably consistent line of FINRA cases in which a retail broker routed customer orders to a small set of paying venues and then reviewed only those venues. The RegLabs record lines them up almost as a template. Robinhood was fined $1.25 million in 2019 for routing to four market makers that paid for flow without adequately assessing whether they were the best market and without systematically reviewing certain order types. Folio Investments was fined $1.3 million in February 2026 for reviewing execution quality only at its existing paying venues, limiting its analysis largely to price improvement, and failing to break results down by order type or size. The wording of these settlements is nearly interchangeable with tastytrade's, which tells you the deficiency is systemic to a business model, not particular to a firm.

Against FINRA's own history, best execution is a steady rather than a headline theme by volume, but its severity has just jumped. FINRA brought a handful of best execution actions a year through the early 2020s, with annual fine totals typically in the low single-digit millions. Then 2026 broke the pattern: just three catalogued actions so far, but totalling roughly $22.15 million, driven by the $20 million Canaccord settlement. In other words, the category is not getting more frequent so much as more expensive, as FINRA reaches firms whose failures sit alongside broader supervisory and AML breakdowns. A mid-range $850,000 fine for a clean, single-issue best execution case looks measured against that backdrop.

Globally, best execution is overwhelmingly a US enforcement theme, and specifically a FINRA and SEC one. Of roughly 777 catalogued best execution actions worldwide, FINRA accounts for 369 and the SEC for 174, together well over two-thirds of all cases, with the US exchanges and a scattering of overseas regulators making up the long tail. Two features stand out. FINRA enforces by far the most often but at a low average fine, around $266,000, because most of its cases are exactly this kind of review-and-supervision settlement rather than a headline penalty. And a single outlier regulator, the UK's FCA, carries an enormous average because a small number of its best execution cases were very large. The pattern tells retail brokers something specific: this is a duty that US regulators examine routinely and settle frequently, so the realistic risk is not a rare catastrophic fine but a near-certain finding if the outward-looking comparison is missing when an examiner looks.

This article is an independent editorial summary for information only and is not legal or compliance advice. tastytrade consented to FINRA's findings without admitting or denying them.