Begin with the End in Mind: How to Win SBIR/STTR Funding

At SBIR workshops, someone usually gets up and tells a great success story: a small company wins a Phase I, then a Phase II, and a few years later it’s selling a product the government actually buys. Everybody in the audience nods along. The question they’re all thinking, and the one I care about, is how you get from where you’re sitting to where that company ended up.

My answer, which I gave at the Catalyst SBIR-STTR Summit in 2020 and again at the Catalyst SBIR Success Workshop in 2026, comes from Stephen Covey’s The 7 Habits of Highly Effective People (paid link): begin with the end in mind. Decide what Phase III looks like before you write the first word of your Phase I proposal, and then write a proposal that leads the evaluator straight there.

I’ve learned this from both sides of the page. I was the principal investigator on a Missile Defense Agency SBIR, MDA11-031, through Phases I and II. I led an Army SBIR, A214-028 (the Risk Assessment Modeling Tool, or RAMoT), through Phases I and II and into a Phase III Other Transaction Agreement – about $19.2M in Phase III and more than $21M in aggregate. I also set up the industry–university team for a Navy STTR, N17B-T032, Phase I with the UAH Information Technology & Systems Center. Along the way I’ve read more solicitations than any reasonable person should, enough that I wrote a dissertation about getting machines to help.

This post is the hub for a series. It covers the big picture; later posts go deeper on each piece.

Views here are my own, not those of any employer or client. As an Amazon Associate I earn from qualifying purchases.

What SBIR and STTR are

The Small Business Innovation Research (SBIR) program goes back to Roland Tibbetts at the National Science Foundation, which started a three-phase small-business research program in 1977. Congress made it government-wide with the Small Business Innovation Development Act of 1982. The Small Business Technology Transfer (STTR) program followed in 1992. Its main difference is that the small business has to team with a university or nonprofit research institution. That’s how my Navy STTR came together with UAH.

According to the Congressional Research Service, 11 federal agencies run SBIR programs and six of them also run STTR. Agencies with extramural R&D budgets of at least $100 million set aside 3.2% for SBIR, and those at $1 billion or more set aside another 0.45% for STTR. SBIR.gov, which the Small Business Administration runs, bills the two programs together as “America’s Seed Fund.” That’s fair. They’re the biggest pool of early-stage, high-risk research money open to small businesses.

The government isn’t doing this to be nice. The programs exist to stimulate technical innovation, to get more small businesses working on federal R&D needs, to widen participation (including by woman-owned, veteran-owned, service-disabled veteran-owned, and economically disadvantaged firms), and to turn federally funded research into products. For the government, that means innovative, high-risk research from small, nimble companies more affordably than it often can elsewhere, plus government use rights in what it funds.

What SBIR is not

Here’s the part people outside the community tend to miss. An SBIR award is not a loan, so you take on no debt. It is not an investment, so you don’t hand over any equity. And it doesn’t take your intellectual property: you keep your IP and your data rights. (The government gets rights to use what it paid for, so read the data-rights clauses in your contract. That’s a topic for another post.)

It’s also not free money. You’re doing contract R&D for a customer with a specific problem. Treat it like a grant to chase your own curiosity and you’ll write weak proposals and lose.

Three phases, one story

Both programs run in three phases:

Phase I: feasibility. A short, small award to show your idea can work. As of October 16, 2026, SBIR.gov says agencies may award up to $335,937 for Phase I without needing SBA approval.
Phase II: prototype. The main R&D effort, which in DoD usually runs about two years and is expected to produce a prototype. The same page puts the Phase II threshold at $2,239,578.
Phase III: commercialization. Taking the technology to market, whether that means sales to the government, sales to commercial customers, or both. Phase III money comes from outside the SBIR/STTR set-aside.

Those dollar thresholds get adjusted periodically, and the rules just changed (more on that below), so check SBIR.gov and your solicitation before you plan around them.

Most people read that list from top to bottom. I’d like you to read it from the bottom up.

Start at Phase III

Phase III is where the payoff is, for you and for the government. For the small business, the program can do a lot. It can bootstrap a new company with seed funding that brings more funding behind it. It can build intellectual property you own. It can teach you business development, introduce you to new customers in other service branches and agencies, and give you experience as a prime contractor. Getting to Phase II pushes you to set up an approved accounting system, which opens doors of its own. And because SBIR awards are competed, winning one can open doors to follow-on work.

None of that happens by accident. So before you write anything, answer the dull question: how will you make money? You might sell a product on the open market. You might sell it to the government. You might put people on time-and-materials contracts to support what you built. Which customer would buy it? Would you sell through a prime contractor? Through which contract vehicle? Does this build on an award you already have? In my book, every additional dollar of revenue counts as success. That’s the bar.

Then turn the Phase III answer into a product. Productize the concept on paper before it exists in hardware or code. Who uses it, what does their day look like with it, and what does it replace? In one Phase I draft I worked on, the first section opened with a day in the life of the end user after the technology had been delivered. That framing makes the reader picture the end state before the technical detail starts.

With that picture in hand, Phase II becomes the prototype that proves it out, and Phase I becomes the feasibility work that clears the biggest risks first. Your job is to walk the evaluator along that whole path. Don’t just claim there’s a route to Phase III. Show it to them.

Gate 0: Pick your battles

Your proposal team has limited time and limited people. A solicitation doesn’t care. One DoD cycle I studied, SBIR 23.3, ran 360 pages, and the companion STTR 23.C added another 133. You can’t propose to everything, and you shouldn’t try.

When the pre-solicitation announcement drops, the first down-select is yours. Look for topics in technical areas where you have real expertise and a solid track record, and where the customer is one you already work with or is close by. Being picky is the point.

I treat this as a gate with exit criteria. To advance a topic, I want a named principal investigator, an innovative approach, a first cut at the research questions, any subcontractors identified, and a high-level commercialization strategy. That last item deserves the exclamation point I put on the slide. If you can’t sketch the Phase III story at Gate 0, a weekend of writing won’t fix it.

(How teams do this down-select in practice, and why it leaks, is a whole post of its own. So is whether a language model can help. My research says it can, with caveats.)

Talk to the TPOC

Each topic lists one or two Technical Points of Contact (TPOCs). A TPOC is a trusted technical expert inside the government organization, and often the person who wrote the topic. In the DoD cycles I’ve worked, there’s a pre-release window, historically about thirty days, when you can call or email TPOCs directly. Once the solicitation formally opens, that direct contact ends, and questions go through the Defense SBIR/STTR Innovation Portal Q&A, where every competitor sees every question and answer. So call early. Ask for a thirty-minute phone call before the open date, and check your solicitation for the current rules.

The point of the call is to understand their problem. Every TPOC brings one particular problem, and your job is to solve that one. Not some general problem in the area, and not the problem you wish they had. Introduce your team briefly, give a short preview of your approach, and then ask open-ended questions:

What’s your history with this topic? Did you write it?
What has your organization tried before, and what frustrated you about it?
Why did you cite the papers listed under References?
What program would benefit from this, and who else might partner?

Then there are the business questions, which people are oddly shy about:

How many awards do you plan to make in Phase I?
Has anyone done related work, past or present? Who?
Are there related topics we should look at?
Could we help shape a topic in a later round?
Should I spend my time proposing on this?

That last one feels rude. Ask it anyway. The most useful skill on these calls is getting the TPOC excited enough to talk at length, then staying quiet. Write down the exact terms and concepts they use. Afterward, ask yourself whether you can now restate the problem more sharply than the topic description. If you can, you have the opening of your proposal.

Gate 1: Go or no-go

After the calls, decide honestly. Bidding without a TPOC conversation is a cold call. Sometimes it pays off, but usually it doesn’t. Writing or rewriting a proposal is a lot of work in a short time. So ask: Was the TPOC interested? Did the topic sound wired for someone else? Does your approach really fit the requirement? Does the team have the bandwidth? Will the result be compelling, not just compliant? If the answers are mushy, walk away and save the effort for a topic that deserves it.

Make it a sales document

This is the central idea, and the 2020 talk outline ended on it for good reason: your proposal’s job is to sell. It isn’t a technical white paper.

Here’s the failure I see most often. The proposal spends page after page on the technology, waves its hands at the work plan, and then adds one thin paragraph on commercialization at the end. Engineers write it that way because the technology is the fun part. Evaluators don’t fund fun. They fund a credible plan that solves their problem.

So flip the proportions. Make it easy for the evaluator: say plainly what you’ll do, how, in what order, and how you’ll know it worked. On commercialization, be specific. Name the market need you’re addressing, estimate how big the market is, and lay out a schedule of what you expect to achieve. Show that you’ve thought it through. Quantify, quantify, quantify. If the solicitation suggests one page for commercialization, I’d write two, as long as the page limits allow it.

“Innovative” is required, and it means there’s some real chance the research fails. Keep it practical anyway. Evaluators want to see risk you’ve thought about and planned for. They don’t want science fiction.

Read the solicitation

I put this one in capital letters on a slide, and I’ll repeat it here: read the solicitation. All of it, including your component’s instructions, which modify the general requirements.

Every solicitation spells out how you’ll be scored. DoD’s SBIR/STTR guide lists three factors in descending order of importance: how sound, technically meritorious, and innovative your approach is; the qualifications of your PI and team, including their ability to commercialize; and the potential for commercial application, government or private. The same guide warns that reviewers judge only what’s in the proposal. They can’t be assumed to know your company or your people. That’s one more reason to hold the evaluator’s hand: anything you leave out, they won’t give you credit for.

Compliance is the price of entry. Build a compliance matrix that maps each “shall” in the solicitation to the section where you answer it, and have someone other than the author review the draft against that matrix, then against the evaluation criteria, then once more. A proposal that loses on a formatting rule or a missing certification never gets its technical ideas read.

A few practical tips

Start early. The TPOC window closes before the solicitation opens, and good questions take time to develop.
Use the TPOC’s words. If they said “sensor fusion,” don’t write “data integration.”
Restate the problem before solving it. Your first page should show you understand their need better than the topic description does.
Write the work plan to be checked. Tasks, schedule, deliverables, and how you’ll show feasibility. Evaluators can spot handwaving.
Make commercialization concrete. Name a customer, a path to market, and a dollar figure you can defend.
Build on what you’ve already won. Earlier SBIR or STTR work is evidence you can deliver.
Know when to walk away. Choosing well beats proposing a lot, and soon it may also be required.

A word on the reauthorization

That last tip is new. SBIR and STTR authority lapsed for about six months until the Small Business Innovation and Economic Security Act (P.L. 119-83) was signed on April 13, 2026, reauthorizing both programs through September 30, 2031. Among other changes, it creates Strategic Breakthrough Awards of up to $30M for companies with a prior Phase II and matching funds, adds Phase III training for contracting officers, strengthens due diligence against foreign risk, and, starting in FY2027, requires agencies to cap how many proposals one company can submit. When proposal counts are capped, picking your battles stops being advice and becomes a rule. I’ll cover the details in a later post. Until then, treat any number in this post as something to check against the current solicitation.

In this series

Planned follow-ups, which I’ll link here as they’re published:

Where SBIR came from
What SBIR is not
Make it a sales document
Anatomy of a DoD solicitation
Down-select by committee
Can an LLM pick your SBIR topics?
Trust but verify: measuring LLM topic recommendations
What changed in the reauthorization