How to bid on a federal contract

Most people think bidding starts when the RFP appears. By then the requirement is written, the budget is set, and often an incumbent or a shaping vendor has already told the customer what good looks like. You can still win from there, but you are answering someone else’s questions.

If you take one thing from this page: answer sources sought notices for work you want, even when there is no money attached and no obligation. That is the moment the government decides whether to set the work aside for small businesses, and what to ask for.

None of the rest matters until these are true. Expect a few weeks, mostly waiting.

Take the size standard seriously. Small is defined per NAICS code, by employee count or average revenue. The same company can be small in one code and large in another, and the code the contracting officer picks determines which you are for that bid.

Certifications are worth the paperwork. An open competition might draw 40 bidders. The same work as an 8(a) or HUBZone set-aside might draw four. Nothing else you can do changes your odds by that much.

Before a contracting officer writes a solicitation they have to understand the market. They do that by publishing a sources sought notice or a request for information. Both ask industry to describe capability. Neither asks for a price, and neither commits you to anything.

This is the highest-leverage moment in the whole process, for two reasons:

1. It decides the set-aside. If enough capable small businesses respond, the work is likely to be set aside for small businesses, and the field shrinks dramatically. If nobody responds, it goes out full and open.
2. It shapes the requirement. What you describe as standard practice can end up reflected in the eventual statement of work.

What a good response looks like: two to four pages. Who you are, your UEI and CAGE, your socioeconomic status, two or three directly relevant contracts with enough detail to be checkable, and a plain statement that you can do this work. Answer the specific questions they asked, in their order.

Also do the unglamorous part. Find the contracting officer and the program office, introduce yourself before anything is published, and ask what they are trying to achieve. This is allowed and expected, right up until the solicitation is released, after which contact is restricted to the official channel.

A serious proposal costs real money and weeks of senior attention. The most valuable skill in this business is declining work that was never winnable. Answer these honestly before committing:

• Do we meet every mandatory requirement?One unmet “shall” can make you ineligible regardless of how good the rest is.
• Do we have past performance that matches? Relevant means similar size, scope and complexity, not merely similar words.
• Did we know about this before it published? If the answer is no, assume somebody else did.
• Is there an incumbent, and are they doing badly? A happy customer rarely switches.
• Can we actually deliver this if we win? A win you cannot perform damages the past performance record you need for everything after.
• Is there enough time? Starting a complex proposal with a week left is usually how you spend money to lose.

A no-bid is a real decision with a reason, not a failure. Record it. The pattern in your no-bids tells you what to go fix.

Federal solicitations follow the Uniform Contract Format, lettered A to M. Most people start at the beginning and work through. Do not. Read M first, then L, then C.

Section M tells you where the points are. If past performance is weighted more heavily than technical approach, your effort belongs in past performance. Teams routinely write a beautiful technical volume for a bid decided on price, or shave price on a bid decided on technical merit.

Know which kind of evaluation you are in. Lowest price technically acceptable means exactly what it says: clear the bar, then be cheapest, and elegance earns nothing. Best value tradeoff means they may pay more for a better offer, and your discriminators matter.

Use the question period. There is a deadline for written questions, and answers are published to everyone as an amendment. If a requirement is ambiguous or looks written around one vendor, ask. Asking costs nothing and the answer binds the government.

A compliance matrix is a table with one row for every single thing the solicitation requires. Go through L, M and C line by line and pull out every shall, must, will and the offeror is required to.

Each row carries: the requirement, where it came from, which proposal section answers it, and who writes it.

This sounds like bureaucracy. It is the single highest-return hour in the entire process, because the most common way to lose is not writing badly. It is failing to answer something, in a document nobody had time to cross-check at midnight.

Mirror their language and their order. Evaluators score with a checklist drawn from Section M. If your headings match their structure, they find each answer where they expect it. If you reorganise things to suit your own narrative, a tired evaluator may simply not find a point you earned.

Most proposals break into volumes, each usually with its own page limit:

Write to the evaluator, not to yourself. They may read thirty of these. They are scoring against a sheet, under time pressure. Short sentences, their headings, the answer first, and the evidence underneath it.

Win themes are claims with proof attached.“We are committed to quality” scores nothing, because every bidder says it. “The same project manager who ran this agency’s last migration, delivered three weeks early” is a discriminator, because nobody else can say it.

Past performance is the asset you build over time. Keep a record of every contract as you finish it: customer, value, dates, scope, outcome, and a reference who will pick up the phone. This is the content you will be asked for on every future bid, and it is painful to reconstruct years later.

Price is where unprepared bidders lose money twice: by bidding too high and losing, or by bidding too low and winning something they cannot deliver.

• Price the work, not the ceiling. Build it up from labour categories, hours, rates and materials.
• Know your indirect rates. Overhead and general and administrative costs are what turn an apparently profitable contract into a loss.
• Make it consistent with the technical volume. If you promise five engineers and price three, evaluators notice, and it reads as either a mistake or a bluff.
• Check the arithmetic, then have someone else check it. Price errors are both common and fatal.
• Understand the contract type. Firm fixed price puts the risk on you. Cost reimbursement shifts it, and brings accounting requirements you must be able to meet.

Established bidders run colour team reviews. Even as a small team, run at least two of these with someone who did not write the thing.

The red team is the one that earns its keep. Hand someone the solicitation and your compliance matrix and ask them to score you honestly. It is uncomfortable and it is cheaper than losing.

The deadline is absolute. A late proposal is normally not evaluated at all, however good it is and whatever the reason. Portals fail, uploads stall, files are larger than you expect. Submit a day early and treat the official deadline as already passed.

• Follow Section L exactly on file naming, format and how many separate files.
• Check every page limit. Overflow pages are often simply discarded unread.
• Confirm you acknowledged every amendment. Missing one can make your offer unacceptable.
• Keep the confirmation receipt, with its timestamp.

Evaluation takes weeks or months. They may ask clarifying questions, or open discussions with the bidders still in contention and request revised offers. Answer quickly and precisely.

Ask for a debrief, whether you win or lose. There is a short window to request one after you are notified, so put it in the calendar the day the notice arrives. A debrief tells you how you were scored and where you fell short. Losing bidders who skip the debrief lose twice.

Protests exist, and are rarely the right move. If something was genuinely improper there is a process, with short and strict deadlines measured in days rather than weeks. Get advice immediately rather than later, but understand that protesting a customer you want to work for has a cost beyond the legal fees.

Then do the thing almost nobody does: write up what happened while it is fresh. If you won, it becomes past performance. If you lost, the debrief notes tell you what to fix. Both are worth more than the next opportunity you chase.

How the government buys depends on the size of the purchase. These figures were checked against FAR 2.101 on acquisition.gov on 26 October 2026. They are adjusted for inflation periodically, so confirm before relying on them.

Lower micro-purchase figures apply to construction subject to wage rate requirements ($2,000) and to services under the Service Contract Labor Standards ($2,500). Higher figures apply to contingency and disaster response work.

Where a new company should start: work under the simplified acquisition threshold. Shorter proposals, fewer bidders, faster decisions, and every award builds the past performance record you need to compete for the larger ones.

Very few bids are lost on the quality of the writing. They are lost on these:

Every one of these is preventable with a compliance matrix, a calendar, and one honest reviewer who did not write it.

For how this platform helps with the parts it can automate, see the user guide.