When a company arranges the delivery of a good or service provided by a third party, ASC 606 requires it to determine whether it is a principal, recognizing revenue at the gross amount the customer pays, or an agent, recognizing only its fee or commission for arranging the transaction (ASC 606-10-55-36). The determining factor is control: does the entity control the specified good or service before it is transferred to the customer? If yes, the entity is a principal. If no, it is an agent. Three indicators support the control assessment but do not replace it.
Key References – Principal vs. agent framework: ASC 606-10-55-36 through 55-40 · Control indicators: ASC 606-10-55-37A through 55-37B
The gross versus net question has real consequences, not just for the revenue line, but for margins, ratios, and how investors read the business. A marketplace that recognized revenue gross and should have recognized net has materially overstated its top line. A software company that recognized net on a bundled third-party service may have understated the scope of what it delivered. The stakes are high enough that the analysis deserves more than a quick judgment at deal signature.
The standard updated this guidance under ASC 606 replaces the older “risks and rewards” test with a clearer control-based framework. Companies that relied on inventory risk or risk-of-loss as their primary argument should revisit the analysis using the current standard.
The Control Test Is the Primary Analysis
Everything in the ASC 606 Principal vs. Agent determination flows from one question: does the entity control the specified good or service before transferring it to the customer? Control means the ability to direct the use of, and obtain substantially all of the remaining benefits from, the asset (ASC 606-10-25-25). For principal vs. agent, it means the entity has obtained the good or service from the third party and can redirect it, use it, or benefit from it before the customer receives it.
If the entity never actually takes possession of the good or never controls the service, if it simply connects the customer with the third-party provider, it is an agent. The fact that the entity is involved in the arrangement, takes on some risk, or sets pricing doesn’t make it a principal. The question is whether control transferred to the entity first.
For goods, this is relatively concrete: did the entity take title, bear inventory risk, and have the ability to direct where the goods go? For services, it’s more judgment-intensive: is the entity primarily responsible for delivering the service and for the outcome, or is it arranging for a third party to deliver it directly?
The Three Indicators — Supporting Evidence, Not a Checklist
ASC 606-10-55-37A lists three indicators that support a finding of control:
- The entity is responsible to the customer for the acceptability of the specified good or service. If the customer holds the entity accountable for delivery and quality — not the third-party provider — this points toward principal. Primary responsibility for fulfilling the promise.
- The entity has inventory risk before the good or service is transferred to the customer, or after the customer has the right to return. This is more clearly applicable to goods than services. Inventory risk.
- The entity has latitude to set the price for the specified good or service. An agent typically receives a fixed fee or percentage; a principal can price the product however it chooses. Discretion in establishing price.
These indicators are evidence, not a three-part test. Meeting two of three doesn’t automatically make the entity a principal. Failing two of three doesn’t make it an agent. The indicators inform the control assessment; they don’t replace it. A company that has full pricing discretion but never actually takes control of the goods is still an agent.
The updated guidance matters for companies using pre-606 analysis
Under prior guidance (ASC 605-45), the indicators carried more independent weight and “risks and rewards” was explicitly part of the framework. ASC 606 removed that framing and made control the primary test. Companies that built their gross vs. net conclusion on risk-of-loss or insurance arguments, without working through the control analysis, should revisit. The conclusion may be the same; the documentation needs to reflect the current framework.
The Analysis Runs Per Performance Obligation, Not Per Contract
A single contract can have multiple performance obligations, and the entity can be a principal for some and an agent for others. This matters for bundled arrangements where the entity delivers its own product alongside third-party services.
Let’s say a SaaS company licenses its software (a PO it controls or principal) and includes a professional services engagement staffed by a third-party implementation partner (a PO it arranges but doesn’t control or agent). If the company prices the bundle as a single arrangement, it needs to separate the transaction price: gross recognition for the software license, net recognition for the implementation of services. Getting this wrong in either direction misrepresents the revenue composition of the deal.
The same logic applies to marketplace platforms that sell both their own products and third-party products. Platform fees on third-party sales are agent revenue; the company’s own product sales are principal revenue. The allocation of the contract price and the presentation in the income statement follows the PO-level determination.
Common Fact Patterns of the ASC 606 Principal vs. Agent
Four arrangements generate most of the principal vs. agent questions in practice:
- The platform connects buyers and sellers. If the platform controls inventory before sale — like Amazon Retail vs. Amazon Marketplace — it is a principal. If it facilitates the transaction without taking control, it is an agent recognizing commission revenue. Marketplaces and platforms.
- If the reseller takes title, bears inventory risk, and has pricing discretion, it is typically a principal. A reseller that simply passes orders to a manufacturer and receives a margin without ever controlling the goods may be an agent. Resellers and distributors.
- A software vendor that includes third-party data feeds, compliance services, or professional services in its contract needs to determine control for each third-party PO. The fact that the vendor is contractually responsible to the customer doesn’t automatically mean it controls the third-party service. Software companies with embedded third-party services.
- Classic agent arrangements: the entity sells on behalf of airlines, hotels, or venues and retains a booking fee. The entity doesn’t control the airline seat or the hotel room before the customer receives it. Travel, events, and ticketing.
Häufig gestellte Fragen
What determines whether a company is a principal or agent under ASC 606?
Control of the specified good or service before it transfers to the customer (ASC 606-10-55-36). Three indicators — primary responsibility for fulfillment, inventory risk, and pricing discretion — support the analysis but don’t replace the control test.
What is the difference between gross and net revenue recognition?
A principal recognizes the full amount the customer pays as revenue. An agent recognizes only its fee or commission for arranging the transaction. The difference directly affects reported revenue and gross margin, even when operating cash flow is identical.
Can a company be a principal for some performance obligations and an agent for others?
Yes. The analysis is per performance obligation, not per contract. A SaaS vendor that licenses its own software (principal) and arranges third-party implementation services (agent) must separate the transaction price accordingly and present each portion appropriately.
Does contractual responsibility to the customer make a company a principal?
Not automatically. Being primarily responsible for fulfilling the promise is one indicator of control — but the control test is the primary analysis. A company can be contractually responsible to the customer while an underlying third party controls the actual delivery of the good or service.
How does the ASC 606 principal vs. agent guidance differ from prior GAAP?
ASC 606 replaced the “risks and rewards” framework with a control-based test. The prior guidance (ASC 605-45) gave more independent weight to individual indicators like inventory risk. Under ASC 606, those indicators support the control assessment — they don’t drive it independently.
Practical implication for complex arrangements
Multi-element contracts that combine proprietary products with third-party services require PO-level principal vs. agent determinations before the allocation step. Getting to the right allocation — and the right revenue presentation — requires a system that can hold the principal/agent classification at the PO level and apply gross or net recognition accordingly. For the full revenue recognition framework, see the BillingPlatform Revenue Recognition guide.
