Why we publish our own bid/no-bid rules
A capture manager sends an email at 4:40 on a Thursday. The solicitation closes in nine days. The message says, in full: "Do you do AI? Can you help?" That email cannot be answered. Not because we are precious about process, but because there is nothing in it to decide on. So a call gets scheduled for Tuesday, the call produces a follow-up call, and the answer that finally lands is a soft maybe with four days left on the clock. Everybody loses that exchange.
We would rather remove the guessing entirely. What follows is the exact input set our team uses to run a bid/no-bid on a subcontract role, what each input decides, and what happens when it is absent. Send those five things and you get a yes or a no inside one business day. If it is a yes, you also get a one-page scope written in language you can drop straight into your volume outline, with a named engineer and a rough labor basis attached.
None of this is a screening test. It is the shortest known path to a real answer. Every item on the list is there because leaving it out has, at some point, cost somebody a week they did not have.
What actually moves a subcontract bid/no-bid call
Editorial weighting of how much each input changes the answer. Illustrative, not a measured statistic.
The five inputs, and what each one decides
Five lines in an email. That is the whole intake. Each line answers a different question, and no line substitutes for another.
| Input | What it decides | If it is missing |
|---|---|---|
| 1. The solicitation or program | Whether the scope is inside what our engineers build, and what the evaluators will actually score. | We read a paraphrase and answer the wrong question. |
| 2. The close date | Whether the schedule is buildable, and where our internal reviews land. | We reserve bench time we cannot honor, or decline work we could have done. |
| 3. The work share | Team structure, statutory performance floors, and whether the effort is a real project. | Scope creeps to fill an undefined budget and the price conversation restarts at the deadline. |
| 4. Named key personnel or not | Which of our people we commit, what resumes and letters we produce, and the conflict check we run. | A resume arrives in the wrong prescribed format on the last day. |
| 5. The data and access path | Whether the technical approach is real work or a promise, and which safeguarding clauses flow down. | Week one of performance is spent waiting for an account that was never scheduled. |
Input one: the solicitation itself, not a description of it
Send the identifier. For DoD SBIR and STTR that is the component code and the topic number. For a full and open procurement it is the SAM.gov notice ID. For a state or local bid it is the RFP number and the issuing department. A link is even better than a number.
Our team reads the source document every time, in full. A prime's summary is written in good faith and it is still where scope drift begins, because the summary carries the technical ask and drops the machinery around it. The evaluation criteria and their order, the page caps per volume, the mandated section headings, the deliverable list, and the small-business participation language all live in the source. Any one of them can flip an answer. A 15-page technical volume with a prescribed heading structure is a different job from a 20-page volume where we control the outline, and the difference is roughly a week of writing.
Reading the source also catches the items that decide team shape rather than technical fit. Organizational conflict of interest language under FAR Part 9.5 sometimes rules out a firm before any technical discussion matters. A systems-engineering support role on one program can bar a firm from competing on the hardware it evaluates, and the cleanest time to find that out is hour two, not week two.
Input two: the close date, and the date you actually need us
Two dates, not one. The portal close sets the outer bound. The internal date, when you need our technical text, our resume, our letter of commitment, and our cost inputs, is the date that governs everything. Most primes want subcontractor material seven to ten working days before submission so the volume can be integrated, red-teamed, and page-checked.
The honest arithmetic on our side: for a defined technical-volume role we want ten working days from yes to final text. For a narrow, well-bounded scope, three or four sections and no new research, we have turned it in five. For an STTR where a research institution also has to sign, add a week for the university's sponsored programs office, which controls the budget approval and moves on its own calendar regardless of how urgent the bid feels.
Short fuses are not automatic declines. A 72-hour ask for a two-page approach section and a named engineer is a yes on many days. A 72-hour ask for a full technical volume on a problem class nobody has scoped is a no, and telling you that on day one is the most useful thing we can do.
Input three: the work share, in dollars or in percent
Give a number or a band. A $40,000 scope and a $180,000 scope are different projects with different staffing, different deliverables, and different answers. "We will figure out the split later" reliably means the split gets figured out at 11 p.m. the night before the portal closes, which is the worst possible negotiating environment for both parties.
The work share is also where statute constrains the deal, and the ceilings are not negotiable between us. Under the SBIR and STTR Policy Directive, an SBIR Phase I small business must perform at least two-thirds of the research and analytical effort itself, and at least one-half in Phase II. That leaves a hard ceiling of one-third for all subcontractors and consultants combined in Phase I. STTR runs on a different split: at least 40 percent to the small business and at least 30 percent to the single partnering research institution, with the remainder available. Eligibility mechanics for the small business sit at 13 CFR 121.702. If you are the SBIR prime, tell us your ceiling and we will size to it rather than pushing you past it.
Outside SBIR, the relevant limits are different. On small business set-asides, FAR 52.219-14 caps what can flow to firms that are not similarly situated at 50 percent of the amount paid for services personnel. Large primes carrying a subcontracting plan under FAR 52.219-9, required when a contract exceeds $750,000 with subcontracting opportunities, have their own small-business goals to report against, and a small-business AI scope helps that column. Tell us which regime applies and the sizing conversation takes one exchange instead of five.
Input four: whether our people are named
Two very different asks hide behind the phrase "we need your team." The first is named key personnel: a specific human appears in the proposal by name, with a resume, a percentage of effort, and often a signed letter of commitment. The second is labor categories: you are buying hours against a category and we staff against it.
We name people. Our standing bench includes engineers, licensed professional engineers, and domain specialists across defense, health, energy, transportation, and public-sector data, and every one of them consents to be named before appearing anywhere. Naming carries checks we run inside the 24 hours: availability across the stated period of performance, whether the same person already appears on a competing team for the same requirement, and whether the solicitation restricts an individual to one proposal per cycle.
Resume format matters more than most primes expect. Some agencies prescribe the fields and the order, and a resume in the wrong structure is a compliance finding rather than a style note. DARPA solicitations, for example, routinely prescribe a school, degree, and year format for every listed individual. Tell us the format at intake and the resume arrives correct the first time. Also tell us your key-personnel substitution language, because a clause requiring contracting-officer approval for any change affects who we are willing to commit for a 24-month period of performance.
Input five: the data and the access path
This is the input primes skip most often and the one that predicts delivery risk best. The question is not whether data exists. It is when our engineers can touch it, and under which instrument.
Customer-owned or public data available at award. The clean case. Name the source and we can put a real evaluation plan in the proposal instead of a placeholder.
Government furnished data promised in the solicitation. Schedule risk that belongs in the proposal risk table. Ask the contracting officer for the delivery date during the question period and hold the answer in writing.
Controlled unclassified information. DFARS 252.204-7012 flows down to subcontractors and brings the NIST SP 800-171 control set with it, and DFARS 252.204-7020 requires a current assessment posted to the Supplier Performance Risk System. Say the word CUI in your first email and we answer the compliance question in the same reply as the technical one.
Export-controlled technical data. ITAR sits at 22 CFR Parts 120 through 130. We hold a current DD Form 2345 under the Joint Certification Program, CAGE 1AYQ0, so militarily critical technical data can move to us without waiting on a fresh certification cycle. That alone has saved primes several weeks of schedule.
Anything requiring a facility clearance. We say so in the first reply and propose the team structure that covers it, before you build a schedule on an assumption.
What gets a fast yes
The pattern is consistent enough to write down.
- The scope is an AI, ML, data, or cloud engineering problem with a defined output someone can inspect.
- Real data exists, or the path to it has a date attached.
- The work share is stated as a number or a band.
- There are ten or more working days to the internal deadline, or the scope is narrow enough to fit fewer.
- The access instrument is identified, even if the answer is "CUI, and here is the flowdown."
- The prime owns the customer relationship and the compliance volumes.
When those hold, our answer is usually a yes within a few hours rather than a full day, because the only remaining work is scoping. We build production systems for federal customers, and the technical read on a well-described problem is fast. Our people have twenty years of building this class of system across five consulting firms, three of them federal, and the pattern recognition on "is this buildable in the stated period of performance" is the part we are fastest at.
What gets a fast no
Equally consistent, and worth publishing so nobody spends effort on a shape we will decline.
- Staffing rather than engineering. Requests for bodies against a rate card, with the technical direction held elsewhere, are a different business than ours.
- An unresolvable conflict. FAR Part 9.5 issues we cannot cure with a mitigation plan, or an existing commitment to another team on the same requirement in the same cycle.
- Open-ended exclusivity. A contractor team arrangement under FAR 9.601 does not require locking a partner out of an entire agency for an unlimited term, and we do not sign that.
- No scope with a full-volume ask. Writing an entire technical volume at risk, with no defined work share and no stated deliverable, is not a partnership offer.
- A data path that is "we will sort it after award." That sentence has a predictable ending and it is a Phase I with nothing to show at the review.
- The technical premise cannot be evaluated. If nobody can state what result would count as success, the proposal will not survive an evaluator either.
Why a fast no is worth more than a slow maybe
What a capture manager is really buying from a teammate is certainty about a schedule. Enthusiasm costs nothing to give and settles nothing. A no on day one leaves eight working days to find another partner, which is enough. A maybe held until day seven leaves two, which is not.
There is a cost side too. Bid and proposal money is real money, and hours spent chasing a partner who was never going to close are hours not spent on the win themes that decide the award. A firm that tells you no quickly is protecting your B&P budget, not just its own.
The reputational math runs the same direction. A prime remembers who wasted a cycle far longer than it remembers who declined politely. Every no we send carries the reason, in one or two sentences, so the next email from that capture manager arrives better aimed. Several of the partnerships we value most started with a declined bid and a clear explanation of what would have made it a yes.
What the 24 hours actually looks like
The clock starts when the email lands, not when the call gets scheduled. There is no call in the loop unless the scope needs one.
From your email to our answer
What arrives with a yes
A yes arrives as a working document rather than a paragraph of agreement, and it fits on one page: a scope statement written so you can lift it into your technical volume, the deliverable list with acceptance language, the named engineer with a percentage of effort, a labor basis you can price against, the assumptions and dependencies we are carrying, and the compliance items we can own inside our own scope.
That last category is worth naming because primes often forget it is available. Data rights markings under DFARS 252.227-7018 for SBIR-developed software, where the protection period runs twenty years from award, are ours to get right on the components we build. Software bill of materials artifacts, security documentation mapped to the control set that applies, model documentation and evaluation reports: these are deliverables our engineers produce as part of the work rather than a separate line item you have to chase.
We also send a mutual NDA if one is not already in place, and we sign yours instead if you have a standard form. That exchange is not a negotiation for us in the normal case, and it should not consume a day of your schedule.
The rules we hold on every teaming deal
Five commitments, stated plainly, that apply to every bid regardless of size.
One team per requirement per cycle. If we are on your team for a requirement, we are not on another team for the same requirement. We check this before we answer, not after.
No open-ended exclusivity. Requirement-specific and cycle-specific exclusivity is normal and we sign it. Agency-wide and indefinite is not.
We read the source document. Every time, before answering. It is why the answer is worth something.
Nobody is named without consent. Every person who appears in a proposal we contribute to has agreed to appear, knows the scope, and knows the period of performance.
Every no comes with a reason. In writing, in the first reply, short enough to read on a phone.
Common questions on how we team
Can you cite our past performance, or can we cite yours?
Under FAR 15.305(a)(2)(ii), a source selection may consider the past performance of subcontractors that will perform major or critical aspects of the requirement. That runs in one direction: a prime may cite a subcontractor's record on the portion that subcontractor will perform. The reverse does not work. On DoD SBIR Phase I evaluations, past performance is generally not a scored criterion at all, which changes the calculus for early-stage teams.
Do you need a teaming agreement before you will write anything?
No. A mutual NDA is enough to start scoping. A contractor team arrangement under FAR 9.601 is the right instrument before substantial writing begins, and we can work from your paper. What we do want before writing is the five inputs, because a teaming agreement over an undefined scope protects nobody.
What if the work share is smaller than you would normally take?
Small scopes get the same 24-hour answer. A tightly defined $30,000 role with a clear deliverable is often more useful to both sides than a vague $200,000 role, and it is how a number of longer relationships start. What we cannot do well is an undefined role of any size.
Will you tell us the real risk in our technical approach?
Yes, in the first reply, whether the answer is a bid or a no-bid. If the stated result is not reachable with the data available in the period of performance, that is worth knowing before you write the volume rather than after the debrief.
Frequently asked questions
The solicitation identifier, the close date, the work share, whether key personnel must be named, and the data or access path. Those five inputs decide technical fit, schedule feasibility, team structure, personnel commitment, and delivery risk. Anything else can be worked out after the answer.
Under the SBIR and STTR Policy Directive, the small business must perform at least two-thirds of the research and analytical effort in Phase I and at least one-half in Phase II, so subcontractors and consultants share a ceiling of one-third and one-half respectively. STTR splits differently: at least 40 percent to the small business and at least 30 percent to the partnering research institution.
Because the data path predicts whether the technical approach is buildable in the period of performance. Government furnished data with no delivery date, or a promise to sort access after award, is the most common reason a technically sound project produces nothing to show at the first review.
The opposite. A no on day one leaves time to find another partner. A maybe held for a week leaves two days and forces a worse choice. A partner who declines quickly and explains why is protecting your bid and proposal budget along with their own.
For controlled unclassified information, DFARS 252.204-7012 and the NIST SP 800-171 control set flow down, with DFARS 252.204-7020 requiring a current assessment in the Supplier Performance Risk System. Export-controlled technical data brings ITAR at 22 CFR Parts 120 through 130 and, for militarily critical technical data, a current DD Form 2345.