The evaluation committee and tender time limits

A committee is appointed where the award is made on the most economically advantageous tender criterion, because it is then that the tender calls for a qualitative assessment. A qualitative assessment calls for a collegiate and competent body.

Under the lowest price criterion the assessment is arithmetical and needs no collegiate body: checking the documentation and drawing up the ranking fall to the tender board in the composition the authority determines. The tender board performs a checking function rather than an evaluative one.

An immediate reading point follows: the presence of an evaluation committee signals that there is a qualitative component and therefore a competitive space other than price. Reading the criterion therefore precedes the decision to take part.

Appointment takes place after the deadline for submitting tenders has passed. Early appointment is a defect of legality and not a formal irregularity.

The reason is one of safeguard and is worth spelling out: if the members were known before the deadline, knowing their identity could shape how tenders were built, and the assessment would lose its impartiality. The rule therefore protects the authority as much as the bidders.

For a bidder this means it is not possible to calibrate a technical tender to known preferences of the assessors, and that the only reference available is the criteria stated in the documents. It is a further confirmation that a tender is built on the criteria and not on assumptions about who will read it.

Aspect	Tender board	Evaluation committee
Criterion that requires it	Lowest price	Most economically advantageous tender
Nature of the work	Documentary and arithmetical checking	Discretionary qualitative assessment
When it is constituted	Not tied to the deadline	After the tender deadline has passed
Main constraint	The rules in the tender documents	Self binding to the criteria and weighting
Protection for a bidder	Regularity of the check	Scores whose reasoning can be reconstructed

The constraints on composition pursue two objectives: technical competence relative to the subject matter and the absence of conflicting positions. The two objectives are checked against different documents.

The code governs conflicts of interest separately, with disclosure obligations on those taking part in the procedure and provisions identifying the situations requiring abstention. The disclosures are placed on the file and are accessible.

One systemic feature introduced by the 2023 code deserves a note: the rules on conflicts of interest were recast, reducing the automatic character of the relevant situations and giving greater weight to the duty to disclose and to a concrete assessment. The effect is to place responsibility on those involved rather than to impose preventive disqualification.

Here sits the principle governing the whole of the committee's work and the main protection available to a bidder. It is the authority's self binding to its own tender documents.

The evaluation criteria and their weighting are fixed in the tender documents and bind the committee. An authority that has bound itself to its own documents cannot depart from them, and an assessment conducted on criteria other than those stated is unlawful.

The self binding also operates on elements that appear to be matters of detail, such as how scores are attributed, any rescaling and minimum quality thresholds. If the documents provide for them they apply, if they do not they do not.

For a bidder it follows that the tender documents are both a tool for preparing the tender and a benchmark for reviewing what the committee did, and that keeping them together with any answers to clarification requests is part of managing the bid. Successive versions of the documents are worth keeping when they are amended.

The committee's technical discretion is wide but not arbitrary, and the ground on which it is measured is the reasoning. A score that cannot be reconstructed from the criteria is contestable.

Where criteria are defined verifiably, with descriptions of the performance expected at each score level, the attribution is reconstructible and a numerical score may suffice. Where criteria are generic, the reasoning calls for a discursive explanation, and its absence is the defect most frequently identified.

Access to documents allows a bidder to inspect the minutes and the assessments, and is the means of checking whether the score received can be reconstructed in the light of the stated criteria. The request must be filed within the periods the rules provide.

One mechanism the tender documents may provide deserves separate attention because it produces exclusions before any comparison of prices. It is the minimum threshold on the technical score.

A minimum quality threshold is the technical score below which a tender does not proceed to the next stage. Where provided, it operates as a preliminary qualitative filter, and a bidder competitive on price but weak on the technical component never gets to bring its discount to bear.

Two qualifications. The first is that a threshold operates only where the tender documents expressly provide for it, by virtue of the self binding. The second is that constitutional case law has struck down the imposition by statute of a mandatory minimum score on the technical tender, holding that it infringes the autonomy of contracting authorities: a threshold therefore remains a choice of the individual procedure and not a general rule.

For a bidder, checking whether a minimum quality threshold applies is among the first steps, because it determines whether the contest is fought on quality or on the balance between the two components. The check is done by reading the paragraph on evaluation criteria.

Between award and signature of the contract the code interposes periods with different functions. The standstill periods and the signature periods run consecutively.

There are standstill periods within which the contract cannot be signed, intended to allow unsuccessful bidders to pursue remedies. And there are periods within which signature must take place, whose expiry produces the effects the code governs.

Within the same interval sit the checks on the successful bidder's requirements, carried out by drawing on public databases, and the formalities relating to securities. A missing document at this stage moves the signature date.

For a successful bidder the consequence is one of programming: the award date is not the start date, and the commitment of resources must be calibrated on the full sequence rather than on the outcome of the competition. The gap between the two dates is measured in weeks.

Not checking whether a minimum quality threshold applies is the first, and can render the whole investment in the financial tender pointless. An excellent financial tender does not recover a technical score below the threshold.

Not seeking clarification on ambiguous criteria is the second, and leaves the committee an interpretative margin the bidder could have narrowed. Answers to clarification requests become part of the governing documents.

Not exercising access to documents after a lost procedure is the third, and deprives the operator of the most informative source for later bids. The minutes show which elements the committee actually valued.

Three actions have a concrete effect on a bidder's position at this stage. They concern clarifications, keeping the documents and access.

Submitting clarification requests within the time allowed, where the evaluation criteria are ambiguous. The answers become part of the tender documents and bind the committee.

Keeping the documentation in order, including answers to clarification requests, which is the material against which any departure of the assessment from the stated criteria is checked. An incomplete set makes any departure hard to demonstrate.

Exercising access to documents after the award, which serves more than a litigation purpose: reading the minutes of a lost procedure is the most informative source for calibrating later tenders to the same authority. The cost of the request is negligible against the value of the information.