The four cases where the honest answer is no
Nearly everything written about federal teaming is written to sell teaming. Add a specialist, widen the coverage, satisfy a participation goal, win more. Sometimes that is right. Often it is not, and the capture manager who adds a sub anyway spends three weeks of calendar and a stack of contracts hours buying a line item that made the proposal weaker. We work both chairs. Our team takes the sub seat on federal bids, and we run subs beneath us when the scope calls for it. From both sides the same four patterns appear: the scope is too small to split, the interface cost is larger than the benefit, an organizational conflict is sitting under FAR Subpart 9.5, or the prime's own bench is already deeper than what the sub would bring. When one of those is true, we say so in writing, on the first call.
Saying no costs a firm nothing and buys it something durable. A prime who hears an honest no on a bid that did not fit calls back on the bid that does. The rest of this is mechanics.
Case one: the scope is too small to split
Start with arithmetic, because arithmetic settles most of these before judgment enters. A DoD SBIR Phase I is commonly funded between $100,000 and $300,000, and the SBA statutory guideline for a Phase I award is adjusted for inflation each year. Suppose a prime wants to hand off twelve percent of a $250,000 effort. That is a $30,000 subcontract. Against it sit a teaming agreement, a mutual NDA, a subcontract with flow-downs, a price-analysis memo, monthly invoice review across an organizational boundary, and closeout. Forty hours of contracts and program-management time, blended, eats a quarter of the award before an engineer writes a line of code.
Our working floor is roughly fifteen percent of the effort or about $40,000, whichever is larger. Below that there is rarely enough continuous work to keep a second organization coherent. The sub's engineer never reaches steady state. They ramp, they context-switch to their own firm's higher-value work, they return cold. The prime pays a full coordination tax for a fractional person.
There is a related failure that looks like generosity: splitting a single role across two firms. One engineer at ten percent inside the prime and ten percent inside the sub is worse than one engineer at twenty percent inside either. Two timekeeping systems, two invoices, two sets of security expectations, one deliverable. If the work cannot be described as a clean seam with its own inputs, outputs, and acceptance test, it should not cross a contract line.
What a subcontract actually costs to stand up
The interface is not free, and most capture plans price it at zero. A first-tier subaward of $30,000 or more pulls the prime into subaward reporting under FAR 52.204-10, which means the sub needs a UEI and the prime files each month. If covered defense information touches the sub, DFARS 252.204-7012 flows down: the sub carries a NIST SP 800-171 posture, a score posted in SPRS, and a 72-hour incident-reporting obligation. Section 889 representations flow down under FAR 52.204-24 and 52.204-25. On a cost-type bid the prime owns a price-analysis record for the sub's proposal under FAR 15.404-3, and certified cost or pricing data enters at the $2 million threshold in FAR 15.403-4.
Intellectual property is where teams get surprised late. Under DFARS 252.227-7017 the offeror identifies and asserts restrictions on technical data and computer software before award, and a sub with meaningful background IP has its own assertions. Those tables have to be merged, reconciled, and defended as one document. A sub that shows up in week three of a four-week bid with a list of proprietary components changes what rights the government receives under DFARS 252.227-7013 and 252.227-7014.
| Interface item | What triggers it | Who carries the work |
|---|---|---|
| Subaward reporting FAR 52.204-10 | First-tier subaward of $30,000 or more | Prime files monthly; sub must have a UEI and current data |
| Safeguarding covered defense information DFARS 252.204-7012 | CDI reaches the sub's systems | Sub: 800-171 posture, SPRS score, 72-hour incident reporting |
| Data rights assertions DFARS 252.227-7017 | Sub brings background IP into the deliverable | Prime merges and defends one assertions table before award |
| Subcontract price analysis FAR 15.404-3 | Any priced subcontract in the cost volume | Prime documents fair and reasonable pricing |
| Certified cost or pricing data FAR 15.403-4 | Subcontract above the $2 million threshold | Sub prepares; prime reviews and forwards |
| Limitations on subcontracting FAR 52.219-14; 13 CFR 125.6 | Small-business set-aside award | Prime tracks the 50 percent split against similarly situated entities |
Then there is the human interface, which is the one that actually breaks schedules. Communication paths grow as n(n-1)/2. A four-person team runs six paths. Add three people at a second firm and the team runs twenty-one, twelve of which cross a contract boundary where questions become emails, emails become change requests, and nobody debugs anything at 9 p.m. without asking whose hours it lands on.
Case two: the interface cost exceeds the benefit
The test we use is short. Name the evaluation factor the sub moves. If a capture manager can point at a stated factor or subfactor and say "this sub raises our score there, and here is the sentence the evaluator will read," the sub earns its overhead. If the honest answer is that the team looks broader, the answer is no. Evaluators do not score breadth. They score the criteria in front of them.
Where adding a sub moves an evaluated score
Editorial weighting from public rules and practitioner reading; illustrative, not a measured statistic.
Participation credit deserves its own note, because it is the most common reason a large prime adds a firm it does not need. Subcontracting plans are required under FAR 19.702 for contracts expected to exceed $750,000 awarded to other-than-small businesses, and FAR 52.219-8 puts utilization language into the file. Those goals matter to a prime's standing across a portfolio. They are still not a reason to bolt a token scope onto a bid where no seam exists. The better play is one bid a year where a small firm carries a whole subsystem.
Case three: an organizational conflict under FAR Subpart 9.5
This is the case that turns an inconvenience into a protest. FAR Subpart 9.5 covers organizational and consequential conflicts of interest, and FAR 9.504 puts the duty on the contracting officer to identify and evaluate potential conflicts as early in the acquisition process as possible. FAR 9.505 states the two underlying principles: prevent conflicting roles that might bias a contractor's judgment, and prevent an unfair competitive advantage. The specific situations follow at FAR 9.505-1 through 9.505-4, covering systems engineering and technical direction, preparation of specifications or work statements, provision of evaluation services, and access to proprietary information.
Protest practice, running back to the Aetna Government Health Plans line of GAO decisions, sorts these into three buckets that are easier to use on a Tuesday afternoon. Unequal access to information means the firm holds nonpublic government or competitor data and could use it in its own proposal. Biased ground rules means the firm helped write the requirement it now bids. Impaired objectivity means the firm would be asked to assess something it has a financial interest in, including its own product.
Where this bites a teaming decision. A prospective sub holds an advisory or technical-support seat with the same program office. A prospective sub wrote part of the performance work statement two years ago under a different contract. A prospective sub sells the tool the new effort is supposed to evaluate. Any of those puts the prime's own award at risk, because the conflict travels with the team, not with the entity that happens to sign the contract. On major defense acquisition programs, DFARS Subpart 209.571 adds its own coverage.
Mitigation exists. Firewalls, personnel screens, a written mitigation plan submitted with the proposal, and in narrow cases a waiver under FAR 9.503 signed at the agency level. All of it is priceable, and that is the point. A firewall costs legal review, a screened staffing plan, and a paragraph in the proposal that a hostile competitor will read looking for a hole. If a sub carries an $80,000 scope and the mitigation package costs half of that in time and risk, the sub is not worth adding. We have walked away from seats for exactly this reason, and we say why.
Ask three questions before the teaming agreement
Does this firm hold information the rest of the field cannot get? Did this firm shape the requirement it now wants to bid? Would this firm be judging its own work, or a product it sells? A yes to any of the three means the conflict analysis happens before the technical discussion, not after. FAR 9.504 puts the evaluation on the contracting officer, but the record the CO reads is the one the prime builds.
Case four: the prime's own bench is already stronger
Personnel sections are read closely and compared line by line. A sub added for coverage brings a named human whose resume sits beside the prime's named humans in the same table. If that resume is thinner, the sub lowered the average, and no amount of narrative recovers it. This is the quietest of the four failures because nobody writes it down. The debrief says the team was adequate, and adequate is what a diluted personnel section produces.
The SBIR case is sharper still. DoD SBIR Phase I evaluation turns on technical merit, the qualifications of the principal investigator and supporting staff, and commercialization potential. Past performance is not among the stated factors, so adding a firm for its contract history buys nothing in that lane. Where past performance is evaluated, FAR 15.305(a)(2)(iii) says the agency should consider the past performance of subcontractors that will perform major or critical aspects of the requirement, which is a specific and narrow permission. A sub holding four percent of the work does not qualify as a major or critical aspect, and stapling its record to the volume invites the evaluator to notice the mismatch.
The reverse test is the useful one. Put the sub's proposed engineer next to the strongest person the prime would otherwise assign, and ask which one you would rather defend in an oral presentation. If the in-house answer wins, drop the sub.
The SBIR and STTR version of all four
Under the SBIR program, performance-of-work requirements at 13 CFR 121.702 hold the small business to at least two-thirds of the research and analytical effort in Phase I and at least one-half in Phase II. STTR runs a different split: the small business performs not less than forty percent and the single partnering research institution not less than thirty percent. The principal investigator's primary employment must sit with the small business during the project. Those rules cap the sub before any judgment enters, and they make case one arrive faster than most teams expect.
Work the numbers on a Phase I. Two-thirds is reserved. The PI, the deliverable engineering, the data work, and the reporting are all inside that reservation on a well-built plan. What is left for outside parties is often ten to fifteen percent, which on a $200,000 award is $20,000 to $30,000. That is the range where a subcontract is a rounding error with a compliance obligation attached. Consultants at an hourly rate, structured as direct labor with no separate accounting interface, are frequently the right structure at that size, and they take four days to arrange rather than three weeks.
What a fast no is worth
Time is the asset a bid cannot buy back. A teaming agreement and mutual NDA with a responsive counterparty run three to seven business days. A sub cost volume with reps and certifications, a rate build-up, and a price-analysis memo run another week. Against a thirty-day solicitation window that is most of the runway, spent on paperwork rather than on the technical volume that decides the award.
A no delivered on day two hands that runway back. A no delivered on day eighteen, after the sub has been briefed, priced, and written into the management plan, is the expensive one. Which is why the answer belongs at the front of the first call rather than at the end of the third.
The ten-minute test before you send a teaming agreement
When the answer is yes
Four conditions have to hold together, and when they do the teaming decision is easy. The scope has a hard seam that can be handed off with its own acceptance test. The capability is genuinely absent from the prime's bench and named somewhere in the evaluation criteria. The workshare is large enough to be continuous rather than fractional. And no FAR 9.5 exposure exists that mitigation cannot price cleanly.
Those are the bids where our team belongs on a prime's page. We run production AI, ML, data, and cloud work, led by a former professor in technology who ranks in the Kaggle Top 200 of more than 200,000 competitors and holds seven cloud certifications, with twenty years building production systems for federal agencies across five consulting firms, three of them federal. Behind that sits a standing bench of named engineers, licensed professional engineers, and domain specialists across defense, health, energy, transportation, and public-sector data. SAM.gov active, CAGE 1AYQ0, JCP and DD-2345 certified. When a workshare is a real subsystem, we take it, staff it with named people, and hold the schedule.
What this changes about the first call
The most useful thing a subcontractor can give a capture manager in March is a clean answer in twenty-four hours. Not a capability deck. Not a discovery process. A yes with a scope and a number, or a no with a reason. Both outcomes let the prime move, and only one of them costs anybody anything.
We keep a written record of the no-cases we call, because they compound. A prime who was told in plain language that a $22,000 workshare was not worth the flow-downs remembers that the next time a $400,000 AI subsystem needs an owner. That is the entire business case for honesty in this lane, and it is a better one than the alternative.
Frequently asked questions
As a working rule, below roughly fifteen percent of the effort or about $40,000 the coordination cost swallows the value. A first-tier subaward of $30,000 or more already pulls the prime into subaward reporting under FAR 52.204-10, and the flow-downs, price analysis, and monthly invoice review do not scale down with the dollar value. At that size a consultant arrangement is usually the cleaner structure.
FAR Subpart 9.5 covers it, with the contracting officer's duty at FAR 9.504 and the specific situations at FAR 9.505-1 through 9.505-4. Protest practice groups them as unequal access to nonpublic information, biased ground rules where a firm shaped the requirement it now bids, and impaired objectivity where a firm would assess its own work or product. The conflict travels with the team, so a sub's exposure becomes the prime's problem.
Where past performance is an evaluation factor, FAR 15.305(a)(2)(iii) says the agency should consider the record of subcontractors that will perform major or critical aspects of the requirement. That permission is narrow. A sub holding a token share does not meet it, and DoD SBIR Phase I does not evaluate past performance at all, so adding a firm for its record buys nothing there.
Performance-of-work requirements at 13 CFR 121.702 hold the small business to at least two-thirds of the research effort in Phase I and at least one-half in Phase II. STTR requires the small business to perform not less than forty percent and the research institution not less than thirty percent. Those floors cap the outside share before any business judgment enters.
It can. Subcontracting plans are required under FAR 19.702 for contracts expected to exceed $750,000 awarded to other-than-small businesses, and those goals carry weight across a portfolio. The credit is still worth more when one bid a year gives a small firm a real subsystem than when several bids each carry a token line item that nobody can describe in a debrief.