Workflow 2026-08-25 6 min read

Bid or no-bid in ten minutes

The most expensive thing a small supplier does is spend a week writing a bid it was never going to win. Here is how to decide in ten minutes.

Updated 25 August 2026

TED (Tenders Electronic Daily) is the online version of the 'Supplement to the Official Journal' of the EU, dedicated to European public procurement.

TED alone published 113,173 tender notices open to bidders in the 90 days to 24 August 2026.

Bid teams have a name for the discipline of deciding what not to pursue: the bid/no-bid gate. It exists because the alternative is optimism, and optimism is expensive when a serious tender costs 20 to 40 hours.

Suppliers without a bid team rarely run a gate at all. A notice looks interesting, someone starts writing, and the decision gets made by the fact that work has already been done. That is the worst way to choose.

Here is the version that fits in ten minutes.

The three questions, in order

Stop at the first hard no. That is the point.

Is it real? A named buyer, a deadline in the future, a defined scope. Prior information notices, market consultations and already-awarded frameworks are not live opportunities, and they read almost identically to ones that are. Check what kind of notice you are looking at before anything else.

Is it winnable? Not "could we do this" but "can we evidence that we have done this". Evaluators score demonstrated delivery, not capability in principle. If the scope reads like a description of somebody's current service, there is an incumbent and you are pricing their renewal for them.

Is it worth it? Realistic value against the cost of bidding, plus where winning leads. A small contract with a buyer you want ten more from can justify a thin margin. A large one-off in a country you cannot service cannot.

Most tenders fail one of these in the first two minutes, which is exactly the outcome you want.

Then score seven factors

For anything that survives, score each 1 to 5 and write the numbers down. Writing them down matters: it lets a colleague disagree with one factor instead of the conclusion.

A low score on capability or evidence is usually fatal on its own. A low score on strategic value just means it has to pay for itself.

The deadline arithmetic people get wrong

Count working days, not calendar days. Then subtract the clarification-question deadline, which falls earlier and is the real constraint if anything in the documents needs explaining.

Then subtract the time to gather evidence, which always takes longer than writing. Certificates, insurance documents, signed references, CVs, financial statements. If you do not already hold these, ten working days is not enough, and the timeline is your binding problem rather than your capability.

Say the verdict in one of three forms

Never hedge between them:

Bid. State what has to be true for this to be won, and what happens first.

No bid. State the single decisive reason. Do not soften it with the six factors that scored well; that is how a no-bid becomes a bid by Thursday.

Conditional. Name the exact thing that would flip it, who answers it, and by when. A conditional with no owner is a no-bid that consumed a week.

Why a fast no is worth as much as a yes

Better bid/no-bid discipline is one of the few changes that reliably improves win rates, and it works by subtraction. You are not bidding better. You are bidding less, on things you can win, with the hours you freed up.

For a supplier doing two tenders a quarter, declining one badly-fitting tender buys roughly a week. Spent on the right tender, that week is the difference between an adequate bid and a good one.

Where an AI actually helps

Qualification is reading, comparison and arithmetic against a fixed set of criteria, which is the shape of task that AI handles well and humans find tedious enough to skip.

What it does well: pulling the buyer, deadlines, values and criteria out of a notice; counting working days; checking requirements against a stored profile; reading 60 pages of specification for the clause that would disqualify you; keeping the process identical every time so that Friday afternoon gets the same rigour as Monday morning.

What it should not do: decide. The strategic value question needs someone who knows which buyers matter and what the company is trying to become. A good qualification ends with a recommendation and its reasoning, and a person who agrees or does not.

The part worth automating is not the judgement. It is everything that has to happen before the judgement is possible, which is where the ten minutes usually go.