8 Nov 2026
When a public contract goes wrong, people look for a culprit: a failing supplier, a careless official, a fraud. Those cases exist. But most losses that could have been avoided build up differently, through a series of ordinary decisions taken at the wrong time or without enough information. Each one looks harmless. Together they turn into amendments, delays, relaunched tenders, expenditure rejected at audit and, in the end, a public service that arrives late or not at all.
This article follows a contract from the first estimate of the need to the archive, and stops at the five moments where those decisions are made. For each one, it suggests a warning sign that is easy to track and a question to ask before moving on.
Before the tender: a need estimated in a hurry
The costliest mistake is often the first. A need estimated quickly, from an old contract or a single quotation, distorts everything that follows: the choice of procedure, the division into lots, the budget set aside and the timetable. If the estimate is too low, bids exceed the budget and the tender fails. If it is too broad, the entity pays for quantities or services it will not use.
Timing causes the same trouble. A tender launched late pushes teams to shorten the bidding period, which keeps serious bidders away, or to use urgent procedures that are harder to justify.
Warning sign: the same need has gone through more than one procedure in the year.
Question to ask: who approved the estimate, and on what data?
The bidding documents: criteria that don’t measure what matters
The bidding documents set the rules of the game. Three weaknesses come up again and again. Technical specifications are copied from a previous contract without being checked against today’s need. Evaluation criteria are too vague to be scored consistently, or their weighting does not reflect what the entity actually wants. And qualification requirements are out of proportion, such as a high minimum turnover, very specific references or rare certifications. They narrow competition without improving quality.
Warning sign: few bids received, or only one found compliant.
Question to ask: could two different evaluators score each criterion and reach the same result?
The evaluation: a decision that is hard to justify
A sound award can be read in the evaluation report. It should show the link between each score and its criterion, explain why a bid was rejected and confirm that committee members declared any conflict of interest. Bids with an abnormally low price deserve a written analysis: accepting a price that cannot be held often means buying a delay or an amendment.
Warning sign: a large price gap between bids, with no comment in the report.
Question to ask: could an auditor who attended none of the meetings rebuild the reasoning from the file alone?
After signature: a contract nobody owns
Much of the effort goes into the award. Once the contract is signed, monitoring is sometimes left to no one in particular. Goods are received without real checks on quantity or quality, amendments pile up until they change the economics of the contract, and late payments accumulate. Late payment has a hidden cost: suppliers build it into their prices at the next tender.
Warning sign: the cumulative value of amendments is getting close to the ceiling set by the applicable rules.
Question to ask: who is responsible for this contract until it is closed, and which indicators do they have?
At closure: the missing document on audit day
In donor-funded projects, expenditure that was properly committed can still be declared ineligible if proof of the procedure followed cannot be found. The entity then often has to refund it from its own funds. The causes are mundane: documents spread across several departments, undated successive versions, unsigned minutes, important exchanges left in personal mailboxes.
Warning sign: rebuilding the complete file of a closed contract takes more than a day.
Question to ask: where is the official version of each document kept, and who can access it?
A scenario to illustrate
Fictional scenario, written for teaching purposes. A public agency needs to equip its regional offices with computers. The need is estimated from the invoice of an old purchase, without allowing for price changes. The tender, launched at the end of the financial year to use up the budget, gives bidders little time. Two bids arrive. One exceeds the budget; the other is much cheaper but promises unrealistic delivery times. The committee chooses it without asking for an explanation. Six months later, half the equipment has been delivered, an amendment extends the deadlines, and the donor funding the project disputes part of the expenditure because the evaluation report lacks detail. None of these decisions was unlawful on its own. Together they cost time, money and the partner’s trust.
Self-assessment: ten questions before your next tender
- Was the need estimated from recent data and approved by a named person?
- Does the procurement plan allow for the real duration of each step?
- Were the specifications checked against today’s need?
- Is every evaluation criterion measurable and weighted according to the entity’s priorities?
- Are qualification requirements in proportion to the size and risk of the contract?
- Have committee members declared any conflict of interest?
- Can the evaluation report be followed without having attended the meetings?
- Is a contract manager named at signature?
- Are amendments and late payments tracked with indicators?
- Is the complete file archived in one place, with dated versions?
Treating procurement as a performance cycle
These mistakes have one thing in common: they appear when each stage is handled as a separate formality. They recede when procurement is managed as a cycle, with an initial diagnosis, realistic planning, contract monitoring, evaluation and lessons carried into the next tender. Three indicators are often enough to start: the average number of bids per procedure, the share of contracts that needed an amendment and the average time between delivery and payment.
To go further, our program E-procurement: governance and risk control covers these controls on an electronic platform, and our training on the financial management of donor-funded projects deals with eligible expenditure and audit preparation. To take stock of your own procedures, see our assessments and diagnostics or our support for delivery and management.