Where Recompete Signals Actually Live

Most of what this industry sells as opportunity intelligence is an agency forecast with formatting. Agencies publish forecasts because they must; the forecasts say what a program office hoped in the quarter someone filled in the spreadsheet.

The record is more honest. A recompete cannot happen without leaving a public trail, and the trail is ordered.

This post is the standing explanation of how we read that trail, and how the expiration pipeline behind our monthly Pipeline edition gets built. We will link back to it every month rather than re-explain.

The sequence

Every competed follow-on walks the same road (chart below), and every step is public the day it happens:

1. The clock. A contract's period of performance end date sits in FPDS and USAspending for anyone to read. This is the earliest signal and the weakest one; it says a decision is coming, not what the decision is.
2. Sources-sought. The agency asks the market who could do this work. The first sources-sought touching an incumbent's scope is the earliest public confirmation that a real competition is being scoped, and it typically lands 12 to 24 months before the end date.
3. Draft RFP, then final RFP. Structure decisions surface here: single or multiple award, set-aside or full-and-open, evaluation scheme. By final RFP the field is forming and the protest lawyers are already reading.
4. Award. The announcement is where trade press coverage starts. As a signal it is nearly the end of the story.
5. The protest window. A protest triggers a stay and a statutory 100-day GAO clock, so every open docket has a checkable status and a known outer date. Last week's release of the Serco decision (the incumbent's challenge to the Army's MCTP award, denied) is what the end of this stage looks like: the stay lifts, the transition proceeds.

And then the loop: when any stage slips, the agency papers the gap with a bridge or extension mod, the end date jumps, and the sequence resets.

The bridge is itself a signal, often the loudest one in the whole sequence. It tells you the follow-on is late, roughly how late (bridges have stated lengths), and that the incumbent's position just got a year stickier.

Two weeks ago NGA published a justification for a one-year sole-source bridge whose stated purpose, quoting the notice, is that it "is essential to complete the development of the SABER II recompete requirement." An agency telling the market, in a legally required document, that a recompete is real and running late: no forecast will ever be that candid.

How the pipeline is built

Our expiration pipeline applies the sequence at scale. The mechanics, plainly:

• Pull every DoD prime contract above $50 million in obligations whose period of performance ends 6 to 24 months out, from the USAspending record. Today that is 1,197 contracts.
• Re-sweep continuously. Rows that stop appearing were awarded, extended, or fell out of the window; each one gets resolved by hand and stays in the record.
• Advance each entry's status as SAM.gov notices and the GAO docket touch it: watch → sources-sought → solicited → awarded or protested. The status changes, not the roster, are the product.
• When a tracked recompete is decided, it moves to the scoreboard: incumbent held or flipped, protest or clean. That record compounds monthly.

A live example of the front end: NAVAIR posted a notice of intent on July 23 to recompete the software engineering support contract at NAWCAD Pax River. No RFP exists yet, and the entry is already on watch, because stage two just happened in public.

The daily reading list behind those status changes is short and unglamorous: the Pentagon's contract announcements every business day around 5pm Eastern, new and updated SAM.gov notices in the coverage NAICS, fresh GAO filings and decisions, and the FPDS mod stream on tracked contracts. Four sources, checked in the same order, every day.

Anything that survives that filter and touches a tracked incumbency becomes either a post or a status change. Usually the status change comes first and the post follows when the story matures.

What this method cannot do

Honest limits, so you can weight the output correctly:

• End dates are noisy. FPDS "current end" reflects exercised options only; unexercised options, in-scope mods, and administrative extensions all move it. An approaching end date flags a candidate, it does not promise a competition.
• The pipeline does not predict award dates. Nothing public does. It tells you where the pressure is accumulating and when the decision window opens.
• Agency forecasts are aspirational, and we treat them as color, never as data. A forecast entry becomes interesting to us only when a notice confirms it.
• Not everything competes. Some incumbencies (interagency arrangements, OEM lock-ins, OTAs outside protest jurisdiction) will ride the loop indefinitely, and part of the read is knowing which rows those are.

Final Thoughts

None of this is secret. Every source named here is free, public, and updated daily, which invites the obvious question of why anyone should pay for the reading.

The answer is the same as for any market data: the signal is public, the watching is work. The sequence only means something if the same 1,197 clocks get checked every day, by the same method, with the misses logged. That is the job this desk does so you can cancel the chore.

Thanks for reading.