Somebody already holds the work you want. Here is when it runs out.
Almost everything a public body buys, it has bought before. Award registers say who holds it, what the term was and which option years have been exercised — and that is enough to say which contracts are approaching the end of their runway, and how likely each is to be competed again.
- Produces
- Predictions — inferred, never published
- Built from
- Award registers and option-year history
- Carries
- A window, a confidence and its evidence
A recompete you hear about at publication is a recompete you have lost.
Displacing an incumbent takes a year of work: understanding what the buyer is unhappy about, getting a hearing before the requirement is redrafted, and building the past performance that makes a switch defensible. None of that fits into a thirty-day response window.
The information needed to start that year on time is sitting in public award data. A contract with three option years exercised and none left is a contract with an end date. A buyer that has competed the same scope every five years since 2011 will probably do it again. A term that has twice been extended instead of competed probably will be extended again — which is just as useful to know, because it says where not to spend the year.
Almost nobody reads registers that way. It is tedious, it is spread across systems that do not talk to each other, and it produces a list that has to be maintained rather than an answer.
- A pipeline that stops at the horizonForecasting from published notices alone gives a quarter of visibility, which is not enough to plan hiring or teaming.
- Capture started too late to matterThe work that would make a displacement bid credible has to begin before the requirement is rewritten.
- Effort spent where nothing will be competedChasing an award that will quietly be extended for two more years costs the same as chasing one that will not.
From an award register entry to a dated pursuit.
Predictions are inferences, and every step is designed so the inference can be checked rather than trusted.
Read the award record
Incumbent, value, period of performance, option structure and any modifications are collected from public registers and disclosures.
A contract record with a term you can see, rather than an end date somebody remembers.
Work out the runway
Option years exercised are subtracted from options available, and extensions are distinguished from exercised options.
The date the award actually runs out, and how much authority remains before it does.
Look for the buyer’s pattern
Prior competitions for the same scope, the cycle they have followed, and whether a cooperative award could be used instead.
A history that either supports “this will be competed” or argues against it.
State a confidence and its evidence
High, moderate or low, with the sentence that justifies it — including when the reason for low is that the record is incomplete.
A prediction that can be argued with, and a low confidence that is a finding rather than a shrug.
Schedule the capture work backwards
A capture start date is derived from the expiry and the lead time your team needs for a pursuit of that size.
A dated task on the pipeline, months before anything is published.
Four contracts, and how much runway each has left.
The bars are the periods: what has been served, what option year is running, and what remains unexercised. Each prediction carries the evidence behind it — including the one held at low confidence because the register does not say how much extension authority is left.
- Student transportation routing softwareCounty school districtconfidence: High
The evidence behind this prediction
Two of three option years exercised on schedule; the third is the last. The same scope has been competed every five years since 2011, and the district publishes a procurement forecast that lists it.
- Dispatch console maintenance and supportRegional emergency communications authorityconfidence: Moderate
The evidence behind this prediction
Proprietary console hardware makes a sole-source justification available and the authority has used one before. The prediction is held at moderate for that reason, not because the expiry is unclear.
- Custodial services, courthouse complexCounty facilities authorityconfidence: High
The evidence behind this prediction
Every option exercised, none remaining, and a solicitation for the adjacent annexe under the same scope was published four months before that award expired.
- Clinical interface engine supportPublic hospital districtconfidence: Low
The evidence behind this prediction
The two prior terms were extended rather than competed and the register does not show how much extension authority remains. Recorded as low, and it stays low until something says otherwise.
Student transportation routing software — expected recompete
- Capture start
- 9 months before expiry — Sep 2026
- Draft solicitation expected
- Feb – Apr 2027
- Incumbent
- Named on the award register
- Object class
- Prediction, not an opportunity
A prediction is its own object class. It never appears in the opportunity feed, it carries a window rather than a deadline, and if the buyer extends instead of competing, the prediction is closed with that outcome recorded against it.
Period served Option exercised, running Option remaining
A prediction is an inference from public records, not an announcement. If the buyer extends instead of competing, the prediction is closed with that outcome recorded against it.
Nothing above is a live query or another customer’s pipeline. Every record in it was written for this page, and no buying organization named in it is real.
What makes a prediction usable rather than decorative.
A number with a percentage sign on it is easy. These are the properties that make one worth putting on a plan.
The option structure, visible
Base period, each option year, and which have been exercised — drawn, so a contract with one year of runway is distinguishable from one with three at a glance.
Extensions counted separately
An extension is not an option year. Conflating them is how a contract appears to be ending when the buyer has two more years of authority.
Incumbent on the record
Who holds it, since when, and through how many renewals — the single most useful fact about a recompete, and the one that decides whether a displacement is realistic.
Cooperative alternatives flagged
Where an existing cooperative award covers the same category and is open to this buyer, the prediction says so, because that is a competition that may never happen.
Capture lead time, backwards from expiry
A start date derived from your own lead time rather than a generic ninety days, so the work lands on a plan instead of a wish list.
Outcomes recorded either way
Competed, extended or cancelled — the outcome is stored against the prediction, so a pattern of buyers who never compete becomes visible over time.
Who plans a year out.
Predictions matter most where the work is incumbent-held and the sales cycle is longer than a response window.
A services firm in a mature category
- Every target buyer already has a supplier, and new opportunities are almost always somebody else’s renewal.
- Predictions across the territory, sorted by expiry, with a capture owner assigned to each high-confidence one.
- A twelve-month pipeline of dated pursuits, rather than a list of open notices and a hope.
A firm deciding where to hire
- Growth requires committing to a region or a practice area before the work exists.
- Aggregate view of expiring awards by category and geography, with confidence and value ranges.
- Hiring and teaming decisions made against a dated picture of what will be competed, not a hunch.
An incumbent defending its own work
- Holds contracts that will be recompeted, and finds out about its own renewals from the buyer.
- Predictions on its own awards, with the capture date set early enough to influence the requirement.
- Defence planned on the same timeline the challengers are using, instead of starting at publication.
Predictions start at Professional.
Recompete predictions arrive with Professional, alongside buying signals: they are the capture-planning half of the product, and Starter is built for responding to what is already open.
- Professional and above include recompete predictions.
- Starter includes discovery, matching and response, and does not include predictions.
- Business and above add the analytics needed to compare predicted competitions against what actually happened.
- Agency adds unlimited company profiles, so predictions can be tracked separately for each client.
Plan names and allowances are mirrored from the billing catalog. Amounts are on the pricing page, which reads them from Stripe rather than from a number typed into a marketing page.
What people ask about predictions.
Where does the contract data come from?
Public award registers, contract disclosures, spending records and published budget documents. Every prediction links back to the entries it was built from, so the inference can be checked against the record rather than taken on trust.
How far ahead does it look?
As far as the option structure allows. A contract with two option years remaining has a runway that can be described today; one whose extension authority is not disclosed is recorded at low confidence with that gap stated as the reason.
What is the confidence based on?
What the register shows about options remaining, whether the buyer has competed the same scope before and on what cycle, whether a cooperative vehicle could be used instead, and whether the record is complete. Each of those is stated on the prediction, not summarised into a score with nothing behind it.
Is a prediction the same as an opportunity?
No, and they are different object classes for that reason. A prediction has a window rather than a deadline and cannot be responded to. If it becomes a published solicitation, the opportunity is created by discovery and the prediction is attached to it.
What happens when a buyer extends instead of competing?
The prediction is closed with that outcome recorded. Over time this is how the picture of a buyer improves — an organization that has extended three predicted recompetes in a row is telling you something useful about the fourth.
Find out what is running out in your territory.
Look at the awards expiring in the next eighteen months across the buyers you sell to, with the incumbent and the option history on each one.