Tender evaluation: how to score bids on price and quality without picking the wrong winner
Most disputes about a tender are not about the winning bid itself but about how it was chosen. Evaluation that looks objective, with weightings, scores and a spreadsheet, can still produce an arbitrary result if the scoring scale is vague or the price formula behaves badly. This guide sets out how to build an evaluation model before the tender is issued, how to score quality and price, and where the common traps lie.
Design the evaluation before you issue the tender
The evaluation model belongs in the tender documents, not in a spreadsheet built after the bids arrive. Bidders write better proposals when they know what will be valued, and a model fixed in advance protects the buyer against the suspicion, or the reality, of shaping the criteria around a favoured supplier. Our guide to RFIs, RFPs and RFQs explains where the evaluation section fits in each type of request.
For public buyers this is a legal duty. In the UK, section 23 of the Procurement Act 2023 requires award criteria that relate to the subject matter of the contract, are "sufficiently clear, measurable and specific" and are proportionate. The contracting authority must describe how tenders will be assessed against them, including whether failing a criterion disqualifies a tender, and, where there are several criteria, indicate their relative importance by percentage weightings, by ranking them, or in another way. Section 19 then allows the contract to be awarded to the "most advantageous tender": the one that meets the requirements and best satisfies the award criteria under that published methodology. The Act's wider changes are covered in our guide to the UK Procurement Act 2023.
In the European Union, Article 67 of Directive 2014/24/EU requires contracts to be awarded to the most economically advantageous tender, identified on the basis of price or cost, which may include life-cycle costing, or the best price-quality ratio. The weighting of each criterion must be stated in the procurement documents, possibly as a range; where weighting is not possible for objective reasons, the criteria must be listed in decreasing order of importance. Private buyers have no such obligation, but copying the discipline costs little.
Three layers: gates, quality, price
A robust model usually separates three kinds of test.
- Pass/fail gates. Mandatory requirements that are not scored: a required certification, acceptance of key contract terms, insurance at a stated level, a delivery date that cannot move. A bid that fails is out, however cheap it is. Keep the list short and genuinely essential, and state clearly that failure disqualifies.
- Quality criteria. Scored judgements on how well the bid meets the need: technical solution, implementation plan, service levels, risk management, sustainability or social value. Each criterion gets a weight, and larger ones may be split into weighted sub-criteria.
- Price or cost. Ideally a whole-life figure rather than the unit price alone, using the methods in our guide to total cost of ownership.
Supplier selection questions, such as financial standing or past experience of the company, normally belong in an earlier stage or a separate pass/fail check. Mixing them into award scoring blurs the question being answered: who can do the work, versus which offer is best.
Choosing the weights
There is no correct split between price and quality. A standard commodity with a tight specification can be awarded mostly or entirely on price; a complex service where poor delivery is costly justifies a heavy quality weighting. Two checks help. First, ask what the organisation would pay for a better bid: if quality carries 60% of the marks, a bid that is meaningfully better on quality may beat one that is far cheaper, and stakeholders should accept that in advance. Second, avoid counting the same thing twice, for instance scoring "experience of the team" under two criteria, or rewarding a feature under quality that is already priced.
Scoring quality: anchor every number
Quality scores are only as consistent as the scale behind them. A common pattern is a 0 to 5 scale in which each score has a written descriptor, for example:
- 0: no response, or a response that does not address the question;
- 1: serious gaps or reservations; the requirement is unlikely to be met;
- 2: partly meets the requirement, with significant weaknesses;
- 3: meets the requirement, with minor weaknesses;
- 4: fully meets the requirement, with clear evidence;
- 5: fully meets the requirement and adds value in ways that matter to the buyer, with strong evidence.
Evaluators should score each bid on its own against the descriptors, not by ranking bids against one another. Each evaluator scores independently first, writing a short justification for every score. A moderation meeting then reaches a consensus score for each criterion, recording why where views differed. Averaging individual scores without discussion hides disagreements that often reveal a misread requirement. The written justifications are also what the buyer will rely on to debrief losing bidders or defend a challenge.
Scoring price: how the formula changes the result
Price is usually converted into points with a formula. The most widespread is the proportional method: the lowest price receives the full price weighting, and every other bid receives the weighting multiplied by the lowest price divided by its own price. It is simple, but it has a property every buyer should know: each bid's price score depends on the other bids.
An illustration, with price weighted at 40% and quality at 60%:
| Bid | Price | Quality (out of 60) | Price score (out of 40) | Total |
|---|---|---|---|---|
| A | £100,000 | 50 | 36.0 | 86.0 |
| B | £120,000 | 54 | 30.0 | 84.0 |
| C | £90,000 | 36 | 40.0 | 76.0 |
A wins. Now suppose a fourth supplier, D, bids £50,000 with a weak proposal scoring 20 for quality. D becomes the lowest price, so every other price score is recalculated against £50,000: A falls to 20.0 (total 70.0), B to 16.7 (total 70.7), C to 22.2 (total 58.2), and D scores 60.0. B now wins, although nothing about A or B has changed and D finishes well behind both. A bid with no chance of winning has decided the outcome.
Ways to limit the effect include scoring price against a fixed reference, such as the budget or a should-cost estimate, rather than the lowest bid; using a linear scale between the budget and a stated floor; or reporting the award as a cost-per-quality-point comparison. Whatever the method, test it on invented bids before publishing it, and publish the formula itself so bidders can check their own scores.
Abnormally low bids
A bid that is far below the others, or below a credible estimate of the cost of the work, may signal a misunderstanding of the requirement, a plan to recover margin through change requests, or a supplier in difficulty. The rules on this are explicit. Under section 19 of the Procurement Act 2023, a UK contracting authority may disregard a tender whose price it considers abnormally low, but only after notifying the supplier and giving it a reasonable opportunity to show it can perform the contract for that price; if the supplier does so, the tender cannot be rejected on that ground. Article 69 of the EU directive similarly requires buyers to ask for explanations before rejecting a tender as abnormally low. In the private sector, the same step makes sense: ask the supplier to walk through its costs before deciding.
Records, clarifications and debriefs
- Log every clarification question and answer, and share answers with all bidders when they concern the requirement rather than one bid.
- Keep each evaluator's initial scores and notes, the moderated scores and the reasons for changes.
- Check the arithmetic: a second person should verify the scoring spreadsheet before the decision is approved.
- Debrief unsuccessful bidders with their scores and the reasons, against the published criteria. A clear debrief reduces challenges and helps suppliers bid better next time.
A short checklist
- Pass/fail gates, criteria, weightings, scoring scale and price formula written into the tender documents.
- Each criterion linked to the subject of the contract, with nothing counted twice.
- Price formula tested on invented bids, including one extreme low bid.
- Independent scoring with written justifications, then a recorded moderation meeting.
- Abnormally low bids queried before any decision.
- Arithmetic checked, decision approved, debriefs sent.
Sourcing & negotiationRFI, RFP or RFQ?
Which request to send, when, and what each document should contain.
Sourcing & negotiationTotal cost of ownership
Which costs to count before, during and after the purchase, and how to compare offers.
Sourcing & negotiationNegotiation basics
Preparation, BATNA, ZOPA and trading concessions without giving value away.