Your SBIR got rejected. Now you have to decide: resubmit to the same agency, switch to a different one, or stop and rethink the whole grant strategy.
Most founders make that call on emotion. They resubmit because giving up feels like quitting, or they walk away because rejection stings. Both are guesses.
A good SBIR resubmission strategy is built on two numbers almost no one pulls: where your application scored relative to the funding line, and your agency's Phase I-to-Phase II conversion rate. Get those two numbers and the decision usually makes itself.
This guide gives you the exact method we use to make that call, including the USAspending analysis that produces the conversion data nobody publishes.
Should I resubmit my SBIR?
Resubmit to the same agency only if your application scored close to the funding line and the reviewer comments point to fixable gaps. If you were triaged or "not discussed," that is a signal of a deeper fit problem, and switching agencies or pivoting your approach usually beats resubmitting the same idea to the same reviewers.
That is the short version. The rest of this explains how to find your real position, how to read your agency's rules, and how to run the math.
Your SBIR resubmission strategy has four options, not two
Most resubmission advice frames the choice as "fix it and try again" or "give up." That framing costs founders months.
The real decision has four outcomes:
- Resubmit to the same agency. Your score was close. The gaps are specific and fixable.
- Switch agencies. Your technology fits more than one agency, and a different one has better odds for your situation.
- Pivot the strategy. Same technology, different framing, different program type, or a state grant to build a track record first.
- Stop. The grant path is wrong for this technology at this stage. Redeploy the 80 hours.
Naming all four matters because the data points to different answers depending on why you lost. A founder who only sees "resubmit or quit" misses the option that the numbers actually favor.
Why founders get this decision wrong
The pattern we see most is the serial resubmitter: a founder who keeps sending the same application back into the same review queue because they read the rejection as "almost," when the data says "not close."
Here is an illustrative case. A founder builds a diagnostics device and applies to the same NIH institute, and the first application is not discussed.
They make edits, resubmit, and get not discussed again. They do it a third time. Each cycle burns roughly 60 to 80 hours and three to four months of calendar time.
The problem was never the writing. A "not discussed" outcome means the application landed in the bottom half of the pile and never reached panel discussion. That is a fit-and-fundamentals signal, not a polish signal. Three rounds of editing the same document does not move a fit problem.
Rejection is data. Most founders never decode it, so they optimize the wrong variable.
Step 1: Diagnose the rejection, not your feelings
Before you decide anything, get your real position. The number you need depends on the agency.
At NIH, you get an impact score and, for many applications, a percentile. Impact scores run from 10 to 90, and lower is better.
You also get a summary statement with reviewer critiques. The single most important distinction:
- Scored but unfunded. Your application was discussed and scored, but landed above the institute's payline. This is a near-miss. Resubmission odds are real.
- Not Discussed (ND), also called triaged. Roughly the bottom half of applications are not discussed at all. This is a fundamental signal. Editing rarely fixes it.
At DoD, SBIR is topic-driven and you can request a debrief. The signal is whether you were "selectable" (technically acceptable, lost on ranking or budget) versus rejected on the merits. Selectable-but-not-selected is the DoD near-miss.
At NSF, you get panel reviews and ratings. A proposal rated competitively that missed the cutoff is a near-miss. One rated low across reviewers is not.
One honest caveat: paylines move. NIH paylines vary by institute and by fiscal year, and a score that funds at one institute will not fund at another. So "near the line" is always "near this institute's line this year," not an absolute.
If you only do one thing before deciding, do this: write down your exact score or ranking and whether you were discussed. That single fact reframes the entire decision.
Step 2: Know your agency's SBIR resubmission rules
"Resubmission" means different things at different agencies. The NIH model is not the DoD model, and applying NIH logic to a DoD topic will waste your time.
| Agency | Resubmission model | Phase II access | Practical takeaway |
|---|---|---|---|
| NIH | One formal resubmission (the "A1"). After an unsuccessful A1, you may submit the same idea as a new (A0) application. The A2 was eliminated in 2009. | Competitive: you apply for Phase II | You get one clean resubmission shot at the A1. Make it count. |
| NSF | Declined proposals can be revised and resubmitted to a future window. A project pitch gate comes first. | Competitive: Phase I awardees apply for Phase II | Resubmission is allowed but treated as a fresh review. Address every panel concern. |
| DoD | Topic-driven. There is no NIH-style A1. You reapply by responding to new topics in future solicitations. | Phase II by application, Phase I awardees only | "Resubmit" really means "find the next relevant topic." Topic fit dominates. |
| DOE | Reapply in future cycles. | Phase II by invitation to Phase I awardees only | Phase II is gated behind a Phase I win, so the Phase I conversion rate matters most. |
| ARPA-H / NASA | Program- and solicitation-specific. Newer or mission-driven structures. Read the specific notice. | Varies by program | Do not assume NIH rules. Confirm in the solicitation. |
The takeaway: your agency's rules set the menu. The NIH A1 is a genuine "improve and resubmit" mechanism. The DoD model rewards finding the right topic far more than rewriting the same proposal.
Step 3: Pull the Phase I to Phase II conversion data nobody publishes
This is the number that should drive your decision and the one almost no founder ever sees.
A Phase I award is not the prize. Phase II is where the real money sits, often five to ten times the Phase I amount. So the question is not just "what are my odds of a Phase I award at this agency," it is "what are my odds of reaching Phase II money through this agency."
That requires the Phase I-to-Phase II conversion rate: of the companies that win a Phase I at this agency, what share go on to win a Phase II.
Agencies rarely publish this in a founder-friendly format. But you can compute it yourself from public award data. Here is the exact method we use.
How to compute Phase I to Phase II conversion rate: Pull all Phase I awards for an agency and technology area from USAspending or SBIR.gov over a multi-year window, then pull all Phase II awards for the same agency and window. Match recipients by unique entity identifier or company name. The share of Phase I recipients that also appear as Phase II recipients is your conversion rate.
For NIH, Phase I awards carry the R43 activity code and Phase II awards carry R44, which makes the matching clean. For DoD and other agencies, match on Phase designation in the award record. Use a window that gives Phase I winners enough time to have applied for Phase II, typically a one to three year lag.
What you find varies a lot by agency, and you should run it for your specific agency and technology area rather than trusting a headline number. As a rough orientation only:
| Agency | Phase I-to-Phase II conversion (commonly reported range) | Notes |
|---|---|---|
| NIH | Often cited around 40-50% | Competitive Phase II; varies widely by institute |
| DoD | Generally lower and component-dependent | Transition depends heavily on a customer/program of record |
| DOE | Phase II is invite-only to Phase I winners | High conditional rate, but gated entirely behind the Phase I win |
Treat those as order-of-magnitude ranges, not precise figures. The point of the method is that you can replace them with the real number for your agency and field in an afternoon of data work.
Step 4: Run the expected-value math
Once you have two numbers, the resubmit-versus-switch decision becomes arithmetic.
The compound-probability rule: Your probability of reaching Phase II money through an agency is the Phase I award rate multiplied by the Phase I-to-Phase II conversion rate. A high Phase I hit rate with a low conversion rate can be worse than a lower hit rate with strong conversion. Always compare agencies on the compound number, not on Phase I odds alone.
Here is an illustrative comparison with made-up numbers to show the mechanics:
- Agency A: 18% Phase I award rate, 25% Phase I-to-Phase II conversion. Compound odds of reaching Phase II: 0.18 x 0.25 = 4.5%.
- Agency B: 12% Phase I award rate, 50% Phase I-to-Phase II conversion. Compound odds of reaching Phase II: 0.12 x 0.50 = 6.0%.
Agency A looks easier because the Phase I hit rate is higher. But Agency B gets you to real money more often. If your technology genuinely fits both, the math says switch to Agency B, even though resubmitting to Agency A feels safer.
Plug in your real diagnosis. If you were a near-miss at an agency with strong conversion, resubmission is usually the highest-value move. If you were triaged at an agency with weak conversion, you are optimizing the worst cell in the table.
The decision framework: resubmit, switch, pivot, or stop
Put the diagnosis and the math together and the call falls out of a small set of rules.
| Your situation | What it means | The data-driven move |
|---|---|---|
| Scored near the payline, fixable critiques | Genuine near-miss | Resubmit to the same agency. Address every critique specifically. |
| Scored unfunded but agency has weak conversion | Near-miss at a low-yield agency | Compare compound odds. Often switch to a higher-conversion agency that fits. |
| Not discussed / triaged once | Early fit signal | Get an outside read on fit. Consider a different agency or program before resubmitting. |
| Not discussed / triaged twice | Confirmed fit problem | Switch agency or pivot the framing. Do not resubmit the same idea a third time. |
| Strong fit at multiple agencies, low conversion where you applied | Wrong-agency pattern | Switch to the agency with the best compound odds for your technology. |
| No agency clears a reasonable compound threshold | Grant path is weak right now | Pivot to a state grant or non-dilutive alternative to build track record, or stop. |
This table is the public mirror of the decision logic we encode in our assessment tooling. It is deliberately mechanical, because the whole point is to take the emotion out of a decision founders usually make emotionally.
Three scenarios and what the numbers say
Scenario A: the near-miss. A founder's NIH application is scored and lands just above the institute payline, with critiques about preliminary data. This is a textbook A1.
The work is close, the gaps are specific, and NIH gives one clean resubmission. Resubmit, and spend the cycle generating the missing preliminary data, not rewording.
Scenario B: triaged twice. A founder is not discussed on two consecutive submissions at the same agency. Two triages is the data telling you the panel does not see the fit.
A third attempt at the same agency is the lowest-value option on the board. Switch agencies if the technology fits another, or pivot the framing.
Scenario C: good Phase I odds, dead-end conversion. A founder keeps winning interest at an agency with a high Phase I hit rate but discovers, after running the conversion analysis, that almost no one in their field converts to Phase II there. The Phase I award felt like progress but was a dead end. The move is to switch to an agency where the compound odds of reaching Phase II are higher, even if Phase I there is harder to win.
All three are illustrative. The pattern is the same: the diagnosis plus the conversion data points to a clear move, and it is often not the obvious one.
When the math says stop
Sometimes the honest answer is that the grant path is wrong for your technology right now. No agency clears a reasonable compound threshold, the fit signals are consistently weak, and another application means another 60 to 80 hours against long odds.
That is not failure. That is the analysis doing its job. The cost of a bad SBIR strategy is not one rejection, it is six rejections and 18 months you could have spent on a state grant, a strategic partnership, or revenue.
We would rather tell a founder "the data says stop here, build a track record first, and come back stronger" than watch them resubmit into the same wall four times.
Get a data-driven read on your rejection
If you have one or more SBIR rejections and you are not sure whether to resubmit, switch agencies, or pivot, that is exactly the analysis we run.
We pull your diagnostic data, compute the conversion rates for the agencies that fit your technology, and build a recovery roadmap that tells you where your next application has the best compound odds of reaching Phase II money. Cada has written hundreds of proposals across 30+ agencies, so we know where the conversion data tends to break by field.
It is a 15-minute call to start. No pitch, no obligation, just a straight read on whether your next move is resubmit, switch, or something else. A good SBIR resubmission strategy starts with the data, and if you have already invested 80 hours and real money into applications that are not landing, the worst thing you can do is guess again.