How to find public contracts before they come back up for renewal (Germany & the EU)

Most bid teams work from the tender feed: a notice appears, the clock starts, and they have three to six weeks to write a response against an incumbent who has been inside the account for four years. That is a losing position, and it is entirely avoidable — the information needed to arrive two years earlier is already published.

A contract award notice states who won, for how much, and — often — for how long. Add the duration to the award date and you have an estimate of when the buyer has to run the procurement again. That estimate is the entire basis of pipeline work in public procurement, and almost nobody builds it, because the data arrives as tens of thousands of XML documents a year rather than as a list.

Where the data actually lives

• TED (Tenders Electronic Daily) is the EU-wide publication of record. Every award above the EU thresholds appears there, in eForms XML since 2023. It is free and it is complete for what it covers.
• National platforms carry what sits below the EU thresholds. In Germany that is the Bekanntmachungsplattform des Bundes plus the Länder systems; other member states have their own. Coverage varies, and below-threshold awards are frequently not published at all.
• Buyer websites and the annual reports of large public bodies sometimes disclose framework contracts that never produced a separate award notice.

For anything above threshold — which is where the money is — TED is enough. The difficulty is not access. It is that an award notice is a legal document written for compliance, not a record designed to be queried.

How to estimate when a contract expires

1. Start from the award or conclusion date, not the publication date. Notices are frequently published weeks or months after the contract was signed.
2. Read the duration from the lot, not the notice. A notice can cover several lots with different terms, and the headline figure is often the sum.
3. Add the extension options. A "4 years" contract with two 12-month options is a 6-year commitment if the buyer is satisfied, and buyers usually are. Treat the base term as the earliest possible re-tender and the fully extended term as the likely one.
4. Treat frameworks differently. A framework agreement does not expire into a single re-tender; it expires into call-offs that may already have moved elsewhere. Its end date still matters, but so does whether call-offs are still being made against it.
5. Discount anything where the value looks wrong. Published values contain keying errors of several orders of magnitude, and a contract "worth" more than the buyer's entire budget will distort any ranking built on value.

The output of that exercise is a date per contract, with a confidence attached. It is an estimate, not a commitment — buyers extend, cancel and re-scope — but an estimate twelve months out beats a certainty that arrives with the tender notice.

What to do with the twelve months

• Qualify the account before the notice exists: who the buyer is, which department runs it, what the incumbent is charging, whether the requirement has changed since the last award.
• Ask for a market consultation. Buyers are allowed to speak to the market before a procurement starts, and are often glad to — the constraint is that they must treat everyone equally once the process is live.
• Watch for the prior information notice. Many buyers publish one, and it is the strongest available signal that a re-tender is genuinely coming rather than being quietly extended.
• Decide early whether to bid. Half the value of a pipeline is the contracts it lets you decline before you spend a bid budget on them.

Doing this at scale

Manually, this works for one buyer. It does not work for a sector: Germany alone produces hundreds of award notices per working day, and the parsing is where the effort goes — resolving that "Stadtwerke München GmbH" and "STADTWERKE MUENCHEN GMBH" are the same buyer, that a value of 4.12 trillion euro is a typing error, and that a lot without an explicit end date still has a computable one.

That is the problem Awantia solves. It parses every TED award notice into buyers, suppliers, lots and values, derives an expiry estimate per contract, and alerts you when a contract you care about crosses the horizon you set. If you would rather build it yourself, the guide above is the whole method — the data is public, and nothing here depends on us.

Frequently asked questions

Is contract award data free to access?

Yes. TED publishes every above-threshold award notice in the EU, free of charge, in machine-readable eForms XML. What costs money is turning it into a usable dataset — entity resolution, value validation and expiry derivation are the work.

How accurate is an estimated expiry date?

Accurate enough to plan against, not accurate enough to promise. Base terms are usually stated; extension options frequently are not exercised on schedule. Treat the base term as the earliest re-tender and the fully extended term as the likely one, and re-check when a prior information notice appears.

Does this work for contracts below the EU thresholds?

Partially. Below-threshold awards are published on national and regional platforms with inconsistent coverage, and in some cases not published at all. For above-threshold contracts — which is where most of the spend sits — TED is complete.

What is the difference between a framework agreement and a contract?

A framework sets terms under which future call-offs can be made, often with several suppliers and no guaranteed volume. Its end date tells you when the buyer must re-run the framework, but individual call-offs against it may already have moved to another supplier on the same framework.

How far ahead should an alert fire?

Twelve to eighteen months before the estimated expiry for anything you would need to write a serious bid for. Public buyers typically begin internal planning a year out, which is the last point at which the requirement can still be influenced.

See how Awantia’s Expiry Radar automates this