Solicitation triage and capture, scored against your NAICS and set-asides

Federal opportunities are published as notices on SAM.gov. In a list, every one of them looks the same: an agency, a title, a NAICS code, a response date. Whether the work is something you can win is not in the list. It is in the statement of work, the evaluation criteria and the attachments — several clicks and a careful read further down.

That is the whole problem. Reading a notice properly costs real time, and most of that time is spent on notices you were never going to bid. A small business does not have someone whose job is to absorb that cost every morning, so it gets absorbed by whoever has a light week.

The failure mode is predictable. You find the right opportunity two days before the response is due. The sources sought notice that shaped the requirement closed a month ago, and you never saw it, so the requirement was written around somebody else’s capability. By the time the solicitation is public, the outcome is largely decided.

The system scans government solicitations daily and scores each one against your firm’s set-asides, codes and capabilities. A score is a sort order — it tells you what to read first, in what sequence, on a morning when you have forty minutes.

NAICS match

Does the code on the notice sit inside the codes you are registered and credible under? An exact match on your primary code and a distant match on a secondary one are not the same signal, and a scoring model that treats them the same is useless.

Set-aside eligibility

This is a gate, not a weighting. If the work is set aside for a socioeconomic category you do not hold, no amount of capability fit makes it winnable as a prime — but it can still be a teaming target, which is a different row in the pipeline.

Capability fit

Measured against what you have actually delivered, not against keywords in the notice. This is the part that decides whether a notice is a real opportunity or a plausible-looking distraction, and it is the part that gets better the more of your own history the system has been given.

Incumbency signals

A recompete with a settled incumbent is a different proposition from a new requirement, and the notice often says so if you read it. Signals like that belong in front of the person deciding whether to chase the work. We do not bury them inside a number, and we do not claim to predict who wins.

What a score is not: a probability of award. Nobody has that, and a vendor who implies otherwise is selling you a number they cannot defend.

Scoring on its own just produces a better-sorted list. The value shows up when a notice moves — when the thing you read on Tuesday is still tracked in October, with the teaming conversation and the draft attached to it. That is four stages, and one system of record.

1. Stage 01

   Triage

   Scored notices come in ranked, and most get closed quickly. The point is that a pass becomes a decision somebody made, on a record you can go back to — instead of a notice nobody ever opened.

2. Stage 02

   Capture

   What survives triage becomes a pursuit with a shape: the agency, the requirement, the dates that matter, and what you would have to prove to win it.

3. Stage 03

   Teaming

   Plenty of good-fit work is out of reach alone — wrong size standard, wrong set-aside, a past-performance requirement you cannot meet yet. Those become partner conversations instead of dead ends.

4. Stage 04

   Drafted response

   The pursuit produces a draft: assembled from the requirement and from what you have already written about yourself. It is a starting document for a person, and it is never more than that.

Underneath it is a 32-table pipeline. The number matters only because of what it implies: notices, pursuits, partners, contacts, documents and decisions are separate records with real relationships between them, not rows in a spreadsheet that lose their history the moment somebody re-sorts it.

Nothing the system drafts goes anywhere until a person approves it. That is a design decision we would not trade away, and it is worth being precise about why.

Federal responses are representations about your business. Size status, socioeconomic status, past performance, key personnel, certifications — these are assertions you are accountable for, and getting one wrong is not a formatting problem. Generated text is confidently wrong in exactly the places that matter most, because the wrong version reads as fluently as the right one.

So the gate is the product, not a limitation on it. The system does the reading, the sorting, the tracking and the first draft — the work that is expensive because it is voluminous. The judgement stays with the person whose name is on the response.

Before an agency solicits, FAR Part 10 requires it to conduct market research — to find out who can do the work and how the commercial market does it. Sources sought notices and requests for information are how that research reaches industry.

Responding is the highest-leverage thing a small business can do in the federal market, and it is the thing that gets skipped. There is no award attached, so it never feels urgent. But this is the only point where you influence how the requirement gets written, and it is how the contracting officer learns whether enough capable small businesses exist to set the work aside at all. Silence from small business is read as absence.

It is also, mechanically, the easiest response to automate a draft for. A sources sought response is short, formulaic, and assembled almost entirely from things you have already written: what you do, what you have done, your codes, your certifications, your point of contact. Turning a two-hour task into a fifteen-minute review is what changes behaviour — the ones you respond to are the ones that were already drafted when you sat down.

This is not a market survey of capture tools. It is a description of software we built and operate, and the first firm it had to work for was ours: 693 federal notices scored to date, on the pipeline that feeds our own federal business development.

We are an SBA-certified Service-Disabled Veteran-Owned Small Business, which means the set-aside logic was written against a real eligibility, ours, before it was ever generalised. We know what the gate feels like from the inside: which notices are genuinely reachable as a prime, which are teaming plays, and which are a polite way of describing an incumbent.

We ran the playbook on ourselves: this company — brand, website, legal, pipeline, and product — was built with its own platform in 16 days.

Read this section as carefully as the rest. It is the part vendors usually leave out, and it is the part that tells you whether the thing fits.

It does not submit

Nothing is filed to SAM.gov, to an agency portal, or to a prime on your behalf. Submission stays a deliberate human act, performed by your people in the government’s own systems.

It does not decide bid / no-bid

A score orders your reading. It does not authorise a pursuit, and it does not close one. Those are business decisions with money attached and they belong to you.

It does not predict awards

No win probabilities, no confidence percentages on outcomes decided by evaluators we cannot see. We will not generate a number we cannot defend.

It is not a compliance authority

It does not certify your size status, your representations, your clearances or your registrations. It works with what you tell it, and you remain accountable for what you assert.

It does not replace a capture manager

If you have one, this makes their reach larger. If you do not, this is the reading and tracking work — not the relationships, the agency conversations, or the judgement.

Small businesses and SDVOSBs that are registered, eligible and genuinely capable, but do not have a full-time capture manager. If checking notices is something that happens between other jobs — and if you have found out about a good opportunity too late more than once — this is aimed squarely at you.

It fits best when you already know what you are good at and can say so in writing, when your codes and certifications are current, and when there is one person who will own the approval queue. That last one is not optional. A gate nobody stands at is just a delay.

It is a poor fit if you want software that decides for you, if you are hoping a tool will substitute for eligibility you do not yet hold, or if you already run a full capture desk on an enterprise platform. In that last case you do not need this — you need people, and we would tell you so on the call.