Tender evaluation software that stands up to scrutiny.

What evaluation software is for

Tender evaluation software takes the submissions from a competition, splits them into the questions and lots each evaluator is responsible for, collects individual scores with written rationale, runs moderation to a consensus, applies the published weightings and price model, and produces the outcome together with the assessment summaries each bidder receives. It replaces the scoring spreadsheet, the shared drive of marked up bids and the moderation meeting where someone types notes into a document afterwards.

The mechanical benefits are real. Arithmetic errors disappear, weightings cannot be mistyped, and a bid cannot be scored against a criterion that was never published. But the substantive benefit is different: the software forces the reasoning to exist at the moment the judgement is made. An evaluator who has to write why a response is a four rather than a three thinks harder about the difference, and the authority ends up with feedback it can send without redrafting.

It also protects evaluators. Scoring a live tender is uncomfortable work, often done by service managers who will have to work with the winner afterwards. A structured environment where scores are individual before moderation, conflicts are declared, and the rationale is anchored to the published scoring guidance gives them cover. Their judgement is recorded as their judgement, and the consensus is recorded as a consensus.

• Question level allocation of evaluators by lot and expertise
• Independent scoring before any scores are shared
• Mandatory written rationale tied to the published scoring guidance
• Automated moderation flags for score divergence
• Weighting and price model applied automatically to final scores
• Assessment summaries generated from the moderated record

The legal backdrop under the Procurement Act 2023

The Procurement Act 2023 came into force on 24 February 2025. It requires award to the most advantageous tender assessed against award criteria that are set out in advance, that relate to the subject matter of the contract, and that are proportionate. Criteria and their relative importance have to be published, and the assessment has to follow them. That is not new in spirit, but the transparency regime around it means the gap between what was published and what was done is far easier for a bidder to spot.

Bidders receive an assessment summary explaining the outcome, and the standstill period gives them a window to act on it. In practice this changes the drafting standard for evaluation records. Notes written for internal comfort are not enough, because a version of them is going to a supplier who may be well advised and who has every commercial reason to test whether the reasoning holds together against the published criteria.

The competitive flexible procedure gives authorities freedom to design multi stage processes, including refining requirements and reducing the number of bidders as the process runs. The freedom comes with a duty to follow the process you published. If you said you would down select after stage one on stated criteria, the record has to show that you did exactly that, on those criteria, on the evidence in front of you.

Building a scoring model that survives scrutiny

A defensible model starts with criteria that a bidder can actually respond to. Vague headings such as quality or approach invite inconsistent scoring, because five evaluators will privately define them five ways. Break the requirement into the things that genuinely differentiate suppliers, describe what a good answer contains, and publish the scoring guidance so bidders and evaluators are working from the same definition of a strong response.

The price model deserves the same care. Lowest price scores maximum with others scored proportionately is common but behaves badly when one bidder submits an outlier. Whatever method you use, publish the formula, test it against plausible bid patterns before you publish, and make sure the software applies exactly that formula rather than a spreadsheet approximation of it. Discovering at moderation that your price mechanism produces a perverse result is a bad moment.

Weightings should reflect what actually matters to the service, not what the last tender used. If social value carries a meaningful weighting, the questions behind it must be answerable and assessable, otherwise every bidder scores similarly and the weighting does nothing except add pages. The same discipline applies to any criterion: if it cannot separate bidders, it is decoration.

Moderation and consensus that mean something

Moderation exists to reconcile honest differences of judgement, not to average them. Averaging scores is quick and it is indefensible, because the resulting number represents nobody's assessment and has no rationale attached. A proper moderation session takes each divergence, re reads the response against the guidance, and reaches an agreed score with an agreed reason. Software should make that the default path and should record the outcome as it is agreed.

Independence before moderation is what makes the session worth holding. If evaluators can see each other's scores while scoring, the senior person in the room effectively scores the bid and everyone else confirms it. Locking individual scores until all evaluators have submitted is a small control with a large effect on quality, and it also demonstrates independence if the process is later examined.

The chair matters too. Moderation needs someone who owns the process rather than the outcome, who can ask an evaluator to point to the sentence in the bid that justifies a score, and who will send the panel back to the document when the discussion drifts into what everyone knows about the incumbent. Systems can enforce structure. They cannot supply judgement, and they should not pretend to.

• Never average scores to resolve a difference
• Lock individual scores until all evaluators have submitted
• Require evidence references from the submission for each agreed score
• Record the moderated rationale during the session, not afterwards
• Keep the chair independent of the service that will manage the contract

Challenge risk and where it actually comes from

Most challenges do not come from a bidder believing the wrong supplier won. They come from a bidder believing the process was unfair or opaque, often triggered by feedback that reads as thin or inconsistent. A well evidenced assessment summary that quotes the published guidance and points to what the response did and did not contain closes the conversation. A summary that says the response lacked detail invites the next letter.

The recurring evidential problems are familiar: scores that changed between moderation and award with no recorded reason, a criterion applied that was not published, an evaluator who scored a bidder they had a relationship with and never declared it, or a scoring spreadsheet with a formula error nobody spotted. Every one of these is a systems problem before it is a judgement problem, and every one is preventable.

There is also the quiet risk of doing everything correctly and being unable to prove it. If the panel's reasoning lived in a meeting and the file records only final scores, you may have run an exemplary evaluation and still be unable to demonstrate it eighteen months later. Contemporaneous records are worth more than reconstructed ones, and only a system creates them reliably.

Sector specifics: NHS, councils, housing and education

NHS evaluation frequently involves clinicians whose time is scarce and whose availability does not align with a procurement timetable. Asynchronous scoring with clear deadlines works better than trying to assemble everyone in a room, and quality criteria often need clinical input while price and mobilisation sit with commercial staff. Provider selection rules for certain health services sit alongside the Act, so the applicable regime needs establishing before the criteria are drafted rather than during evaluation.

Local authority evaluation is typically the broadest, from construction and highways to care and culture, with panels drawn from service teams who evaluate rarely. Guidance embedded in the scoring screen, so the definition of each score is in front of the evaluator as they work, does more for consistency than a training session held two months earlier. Cabinet and committee cycles also mean the evaluation record often has to support a public report.

Housing associations weight contractor competence and building safety assurance heavily, particularly for repairs, planned works and compliance contracts, and the evaluation needs to connect to the supplier assurance record rather than rely on a self declaration in the bid. Education trusts usually evaluate small panels on framework mini competitions, where the risk is not complexity but informality: quick decisions made sensibly, with nothing written down.

Evaluating on Dynamic Markets and framework call offs

The Procurement Act 2023 replaced the Dynamic Purchasing System with Dynamic Markets. Because a Dynamic Market is permanently open and membership cannot be capped, evaluation happens at two levels: assessment of applications against the published conditions for membership, and assessment of tenders in the competitions run from it. The first is a rolling assurance task, the second is a normal evaluation, and confusing them is a common error.

The timing rule bites here. Applications must be assessed within a reasonable time, and any pending application must be considered before a competition concludes. So a supplier who applies during your tender window cannot simply be told they were too late. Your membership assessment queue and your competition timetable have to be managed together, which is far easier when both live in the same system with the same visibility.

Framework mini competitions carry a different trap. Because they feel routine, the discipline slips: criteria are lifted from the framework without being made specific, evaluation is done informally, and the record is a short email. The mini competition is still a competition, and the same evidential standard applies. Treating call offs as first class evaluations, with the same templates and controls, costs very little and removes a real exposure.

Audit, reporting and what good evidence looks like

The evaluation record should stand alone. Someone who was not involved should be able to open it and follow the logic: these were the published criteria and weightings, these evaluators were assigned and declared no conflicts, these were the individual scores with rationale, these differences were moderated and this is why, this is the price calculation, this is the resulting ranking, this person approved it and on this date. If any link in that chain is missing, the record is an assertion rather than evidence.

Beyond individual competitions, evaluation data tells you things worth knowing. Which criteria never separate bidders and should be redesigned. Which evaluators score consistently high or low, which is a training conversation rather than an accusation. How long moderation takes by category. How often the highest quality score also wins on price. Authorities rarely look at this because the data has historically been trapped in spreadsheets.

Retention matters as well. Evaluation records may be needed years later for audit, freedom of information requests or a dispute during contract delivery. Records held in one system with defined retention are manageable. Records spread across personal drives and mailboxes are not, and they become a liability the moment a subject access request or an FOI arrives.

How eSourcing Data supports evaluation and moderation

eSourcing Data covers source to contract in a single platform, and evaluation sits at the centre of it. Criteria and weightings published in the tender flow directly into the scoring model, so the two cannot drift apart. Evaluators score independently with mandatory rationale, moderation is run and recorded in the system, and assessment summaries are generated from the moderated record rather than written from scratch under standstill pressure.

Because evaluation connects to supplier assurance and to contract management in the same platform, the assessment does not end at award. The commitments a supplier scored well on become obligations you can monitor during delivery, which is the only way social value and method statement promises turn into something real rather than something scored. Data is held with UK data residency and the platform is GDPR compliant.

Public buyers can buy eSourcing Data software through RM1557.15 G-Cloud 15, with 28 software services listed on the Digital Marketplace plus cloud support services. Framework purchases are made as call off contracts. If you want to test whether the scoring model handles your price mechanism and your social value weighting, the most useful thing is to walk through a live competition of your own in a demonstration.