I am new to federal contracting, so when I needed a capability statement I did what I do with any unfamiliar system: I went looking for the specification rather than the tutorials. In this case the specification is a set of training documents produced by government small business offices for their own outreach events. They are dry, some of them have not been revised in a decade, and they say things that flatly contradict most of what gets sold to small businesses as capability statement help.

Here is what I found and what I did with it.

The document has one job, and it is not to win you work

Contracting officers use capability statements during market research under FAR Part 10. The question they are answering at that stage is whether a requirement can reasonably be satisfied by small business, and if so under which socioeconomic category. They are counting capable firms. Two or more responsible small businesses at fair market prices means a set-aside becomes possible, and your document exists to make you countable.

Nobody is scoring you at this point. There is no evaluation, no technical volume, no source selection. That reframing changed what I put on the page, because a document written to persuade reads differently from a document written to be classified correctly, and the second one is what is being asked for.

The formatting conventions follow from that. One page. PDF, under 1MB, because these get emailed and attached to internal notes and dropped into vendor folders. A filename like CompanyNameCapabilityStatement.pdf, so that when it lands in a directory with two hundred others it is findable by the company name rather than by capstatement_v4_FINAL.pdf. None of this is aesthetic advice. It is about surviving contact with somebody else’s filing system.

The conventional structure is five blocks: header, core competencies, past performance, differentiators, and company data. Use it. Buyers scan these by position on the page, and a creative layout costs you the one thing you have, which is instant legibility.

The differentiators section is where everyone fails

This is the part worth the whole article.

The Air Force guidance I read is explicit that socioeconomic certifications do not belong in your differentiators. The reasoning is obvious once stated and I had not thought of it: on a set-aside, every firm competing holds the same certification. Being small business certified on a small business set-aside distinguishes you from nobody in the room. The certification belongs down in your company data block, listed alongside your CAGE code and your NAICS codes.

The same guidance rules out superlatives. “Best in class.” “World class.” Anything that describes quality without evidence. And the one I see most often, which the government documents name specifically: “over twenty years of combined experience.” Combined experience is an arithmetic trick. Five people with four years each produces the same number as one person with twenty, and every contracting officer has read the phrase a thousand times.

What counts as a real differentiator is anything a buyer can verify and that most of your competitors cannot claim:

  • Facility clearance level, and the number of cleared personnel by level.
  • Named certifications with dates: CMMI maturity level, ISO 27001, FedRAMP authorization status, CMMC level and assessment status.
  • Experience with a specific system at a specific agency, named.
  • An authority to operate you helped a client achieve, with the system type and the framework.
  • A delivery metric with a number attached and a scope that makes the number meaningful.

Notice that every item on that list is a fact somebody could check. That is the test I now apply to any line I am tempted to write. If a reasonable person could not disprove it, it is not telling them anything.

Two or three strong differentiators beat a column of mediocre ones. A list of nine items signals that you could not decide which two mattered, and it costs you the scan.

Past performance, and the line I keep coming back to

The Air Force document says it about as plainly as a government publication says anything: if past projects do not relate to the targeted agency’s needs, do not list them.

That instruction is harder to follow than it sounds, because your best project is your best project and leaving it off feels like leaving money on the table. But the ordering the guidance recommends is by relevance to the reader: work for that same agency first, then related agencies, then other government at any level, then commercial. A commercial engagement that closely matches the requirement outranks a federal one that does not. An impressive project in an unrelated domain is filler, and filler on a one-page document is expensive.

I have no federal past performance. None. That is the honest position of most firms entering this market, and the temptation is to disguise it with vague language about “extensive experience supporting mission-critical environments,” which is exactly the sort of thing a contracting officer has learned to skip. What I did instead was describe commercial and regulated work concretely, with the system type, the compliance regime, the scale, and the outcome, and let the reader decide whether it maps. If it does not map to their requirement, I would rather they know that quickly than discover it after a wasted conversation.

How buyers actually find you

Two mechanisms, as far as I can tell.

The first is keyword search. Buyers search SBA’s Small Business Search, which replaced the Dynamic Small Business Search and now lives at search.certifications.sba.gov, and they search their own agency vendor files and prior sources sought responses. That means the words in your profile and on your document are a retrieval problem before they are a persuasion problem. I rewrote our core competency bullets after realizing that I had described our work in the vocabulary I use at conferences rather than the vocabulary that appears in a performance work statement.

The second is that when a buyer already has a draft requirement, they read your bullets looking for their own language reflected back. This is why generic capability statements underperform and why the practitioners who do this for a living insist on a tailored version per agency, or per requirement if you are serious about it. Same firm, same facts, different emphasis and different words drawn from the documents that agency publishes.

The rule of thumb people in this field repeat is that the first scan lasts about six seconds. I have no study to point to and I would treat the specific number as folklore rather than measurement. As a design constraint it is still useful: assume the reader decides whether to keep reading before finishing your header, and put the identifying facts where a six second scan lands.

What I think the consultant market gets wrong

There is a large industry selling capability statement services to small businesses, and much of what it produces is a well-designed page full of unfalsifiable claims. Attractive typography, a stock photo of a data center, three bullets about commitment to excellence, and the socioeconomic certifications set in large type across the top because that is what the client asked for.

The government’s own material tells you not to do that. It has told you not to do that for years, in free PDFs sitting on contracting squadron websites. I do not think this is a conspiracy. I think design services are easier to sell than the uncomfortable work of admitting you have two real differentiators and one relevant past performance reference.

My own document is now one page, has three differentiators, names two commercial engagements by system type and compliance regime, and does not claim we are best in class at anything. It is less impressive than the version I first drafted. I expect it to work better, and I will find out.