What Each SBIR Phase Actually Proves
In the SBIR program, the most useful question is not whether an award appears in a company’s history, but which phase that award paid for. This chapter traces the evidence that Phase I, Phase II, sequential Phase II, and Phase III are meant to produce, so readers can interpret a phase label as a bounded claim about work and funding rather than as a certificate of market success. The distinction is practical: it tells us what to look for next when an award announcement says that a technology is moving toward commercialization.
Phase I establishes a feasibility question
Phase I is the program’s initial investigation. A small business uses the award to determine whether an idea has technical merit and is feasible enough to justify further development. The idea may have commercial potential, but potential is part of the question being investigated, not the answer already established.
That difference is easy to lose in ordinary language. “The agency funded an innovative technology” sounds like a judgment about the technology’s future. The narrower and more accurate reading is that the agency chose to fund work that could resolve an important uncertainty. The uncertainty might concern whether a material can perform as proposed, whether a device can work under relevant conditions, or whether a software approach can meet a mission need. The award makes the investigation possible; it does not make the proposed result true.
The agency’s reason for caring also matters. SBIR was designed both to stimulate small-business innovation and to help federal agencies meet research and development needs. A project can therefore be worth investigating because it addresses a public mission even if the eventual commercial market is small. The agency may be assessing a capability it needs, a technical option that could improve a public program, or a line of research that could produce wider benefits. Commercial potential is relevant, but it is not the only lens through which the award decision is made.
Phase I evidence should consequently be read at the level of feasibility. It may show that an experiment worked, that a design constraint is manageable, or that a proposed approach deserves a larger development effort. It may also show that the idea needs a different design or application. Both results are useful forms of R&D learning. Neither establishes that a product can be manufactured, integrated into a customer’s operation, purchased at a sustainable price, or supported over time.
A useful way to read a Phase I result is to ask which uncertainty became smaller. If the work demonstrates that a technical approach is possible, the next uncertainty may be repeatability. If it reveals a promising performance characteristic, the next question may be whether that characteristic matters to a real user. If it shows that an idea is not feasible in its original form, the project may need to change direction. The award does not promise that every uncertainty will shrink in the desired order; it finances a structured attempt to learn.
A Phase I award also does not guarantee a Phase II award. The next decision depends on what the initial work demonstrates and on the agency’s continuing needs and funding choices. Even when Phase II follows, the transition means that the project has passed one technical decision point; it does not mean that the market has accepted the technology. The distinction is similar to the difference between proving that a bridge design can bear a test load and proving that a transportation system can build, operate, maintain, and pay for the bridge. The analogy is only a way to clarify the separate questions. The SBIR evidence itself remains about a specific R&D project and its phase.
This is why the phrase “commercial potential” deserves careful handling. It is a reason to investigate, not a report of realized demand. A Phase I description may identify an intended market or government use, but an intended user is not necessarily a committed customer. The proposal may explain how the technology could be sold or deployed, but a plan is not a transaction. To establish more than feasibility, later evidence must show what happened when the technology encountered actual performance requirements, buyers, funders, or procuring agencies.
There is also a difference between an agency’s technical selection and a market’s willingness to pay. The agency can decide that learning about a technology is worthwhile even when the commercial case is uncertain. Public R&D support is partly justified by knowledge that may help a mission, create an option, or enable later work. A company announcement that accurately reports the award can still create an inaccurate impression if the reader supplies the missing market judgment. The phase label is not misleading; the inference drawn from it is.
Phase II continues development; Phase III changes the funding environment
Phase II extends the research and development begun in Phase I. According to the program descriptions reviewed by GAO and NIST, Phase II can include continued technical work and prototyping. That makes the project more concrete, but it does not make every unresolved question disappear. A prototype is evidence that a design has been embodied and tested in some way. It is not automatically evidence that the design can be produced reliably, adapted to a customer’s operating environment, or sold at a price that supports the business.
The difference between a prototype and a product is not a matter of prestige. It is a difference in obligations. A research team can learn from a small number of carefully made units. A customer may need repeatable performance, documented support, predictable delivery, compatibility with existing systems, and a clear answer about who bears the risk when something fails. The evidence needed for those decisions may arrive after the technical work described in a Phase II award. The phase can make those later tests more plausible without carrying them out in full.
Phase II often changes the type of uncertainty rather than eliminating uncertainty altogether. Phase I asks whether the idea can work. Phase II may ask whether it can work consistently, under more demanding conditions, or in a form that others can evaluate. A prototype can expose new constraints: a process may be difficult to repeat, a component may be hard to source, or an application may require adaptation to a customer’s environment. These are not signs that the earlier award was a mistake. They are the kinds of constraints that become visible only when a concept is made more real.
The distinction between technical and commercial evidence is especially important for technologies that must be integrated into a larger system. A device may perform well on its own but require changes to a facility, workflow, software environment, or procurement specification. A material may achieve a target property but still need a production process that can deliver consistent batches. A scientific instrument may work for researchers while serving a small and specialized market. Phase II can support the work of discovering and addressing these issues, but its existence does not show that the surrounding system has accepted the change.
Phase II can also reveal a project’s transition burden. A small business may need additional engineering before it can scale production. It may need to test an application with a particular customer rather than merely demonstrate performance in a laboratory. It may need private capital or a non-SBIR government contract to pay for work that is necessary but not eligible for the original award. A Phase II award can therefore be both a sign of progress and a reminder that the most expensive transition questions remain ahead.
Some projects receive a sequential Phase II award, meaning that the agency supports another continuation of the same development effort. This is an important step in the program’s logic because it can keep technically promising work alive beyond one Phase II period. It is still a continuation of R&D, however. The additional award says that the project remains worth developing under the program’s rules; it does not say that customers have adopted the technology or that external funding is already secured. Indeed, the requirement for a prior and sequential Phase II later becomes part of the eligibility chain for the Commercialization Assistance Pilot Program, which exists precisely because further transition work may still be needed.
A sequential award can preserve something valuable: continuity. The technical team may retain knowledge, maintain equipment, and continue a line of experiments instead of stopping at the moment when the next question becomes visible. But continuity is not the same as completion. Additional public R&D can reduce technical uncertainty while leaving demand, pricing, manufacturing scale, and procurement unresolved. It can also make a project more attractive to outside parties without replacing their eventual decision.
Phase III marks a different relationship to funding. The program describes it as the stage for commercializing technologies developed under Phases I and II, but it is not funded by additional SBIR Phase I or Phase II money. The next source may be a non-SBIR agency program, a government contract, private investment, licensing, or sales. That change matters because the decision makers, incentives, and proof requirements can change with the funding source.
A federal research manager may ask whether the project addresses an agency need and whether the proposed work is technically credible. A private investor may ask how much capital is required before revenue, what risks remain, and whether the market is large enough. A government acquisition office may ask whether the technology is ready to integrate into an operational program and whether a procurement route exists. An industrial customer may ask whether the supplier can deliver consistent quality, meet schedules, and provide support. Phase III is not one universal test; it is a transition into these outside decisions.
This change in decision maker is one reason that phase progression should not be mistaken for a single continuous score. The project carries technical knowledge from one phase to the next, but it does not carry automatic approval from every party whose participation is required later. A government research program may be willing to fund a promising option before an acquisition office is ready to buy it. An investor may support additional development without guaranteeing demand. A customer may test a technology without committing to repeat purchases. Each event resolves a different uncertainty.
The phrase “Phase III commercialization” should therefore not be read as “commercialized.” It identifies the phase in which the technology is pursued through non-SBIR sources. A Phase III contract can be strong evidence of government procurement, but it is not necessarily evidence of broad private adoption. A license can show that another organization has taken rights to use or develop the technology, but it may precede production or sales. A first sale establishes market contact, but not necessarily repeat demand or durable economics. The meaning depends on the specific outcome being claimed.
The sequence also permits pauses and changes of path. A project can finish one technical stage and wait for a funding commitment. It can be adapted to a different market or government mission. It can produce expertise or an invention that supports a different product from the one named in the original award. A firm can make technical progress while failing to secure the production capacity or customer commitment required for the next step. None of these possibilities makes the earlier award meaningless. They show why phase progression should be treated as a sequence of decisions, not an automatic conveyor belt.
The most useful reading practice is to attach one unanswered question to every phase. For Phase I, ask whether the proposed approach is technically feasible. For Phase II, ask what development or prototype evidence has been produced and what performance or production uncertainties remain. For a sequential Phase II, ask what additional R&D problem justifies continuation. For Phase III, ask which outside source is carrying the technology toward use and what that source has actually committed to do. For commercialization, ask whether the evidence shows a customer, procurement, repeat use, revenue, or another explicitly defined outcome.
A phase label is valuable because it tells us where to place the evidence. It is not valuable because it lets us skip the evidence that comes next. Phase I can make an idea worth developing; Phase II can make a technology more capable and concrete; a sequential Phase II can preserve the runway for further work; and Phase III can connect the technology to external funding, procurement, licensing, or sales. The later outcome still depends on what happens at those connections.
That is the boundary the program’s structure quietly insists on. SBIR can finance the creation of technical possibilities, but commercialization requires a handoff to people and institutions that will fund, buy, integrate, produce, or use the result. The next question is why that handoff can fail even when the technical work is promising.