This series of articles draws on our disputes experience and identifies 7 common Procurement Pitfalls. When we advise on procurement challenges we tend to find the same types of problems, irrespective of the sector in which they arise. Often these are problems which emerge from the content of the tender documents, and which lead to problems for evaluators. The objective of these articles is to forewarn so that early thought can be given to avoiding these issues.
We will be focussing on:
- Price/evaluation methodology
- Waiving requirements
- Imposing unreasonably high requirements
Procurement Pitfalls 1 – pricing formulae
Pricing formulae: There are an almost infinite number of ways to evaluate price. The most common in our experience is inverse proportionality in which each bidder is awarded a percentage score based on the degree to which its price exceeds the lowest. We focus here on two examples in which the pricing evaluation formula had unintended consequences that resulted in a legal challenge.
Our first example is where a formula did not anticipate and account for the possibility that a bidder might submit a “zero” price for part of its bid.
In this case total price was weighted at 30%. Bidders were asked to submit prices for three elements each of which was given a sub-weighting: block price (10%), additional unit cost (15%), and service cost (5%). Each element was evaluated on an inverse proportionality basis (lowest price / submitted price * weighting). One bidder submitted “zero” for its additional unit cost, on the grounds that all of its costs were covered by the block price. The problem is illustrated by the following table:
|Weighting||Bidder 1||Score||Bidder 2||Score|
|Additional unit price||15%||£10||0%||Nil||Undefined|
Any number divided by zero is undefined (i.e. it has no value) so a zero price is not capable of being evaluated on an inverse proportionality basis. This is not remedied by substituting a nominal price for zero, as the following table shows:
|Weighting||Bidder 1||Score||Bidder 2||Score|
|Additional unit price||15%||£10||0.015%||£0.01||15%|
Under no circumstances would Bidder 1 win if Bidder 2 were able to submit a nominal price for one element. This is despite the fact that Bidder 1’s total price was lowest. It could be that an authority puts a very high premium on minimising costs over and above the block price because they want budget certainty, in which case such a formula might be capable of identifying the most economically advantageous tender (“MEAT”). However, where an authority has expressly stated that the highest score would be awarded to the lowest price, such a formula may be found to be irrational on the grounds that it could be manipulated such that it is not capable of identifying the MEAT on that basis. In similar circumstances the Swedish Court of Appeal has declared unlawful an evaluation methodology that allowed bidders to “game” the system by loading costs onto a low-weighted criterion, thereby enabling them to submit an artificially low price, on the grounds that this failed to identify the MEAT as that had been defined by the contracting authority.
What should the authority have done?
The most common issue we are consulted about is where a pricing/quality evaluation formula is simply carried over from a previous procurement, but does not quite fit with the authority’s current needs. Once the authority starts to evaluate, it realises this but is then constrained not to amend the weighting. Conversely, whilst time and resource are of course scarce, careful modelling and “dry-running” of likely or possible bid responses before the documents are even issued is usually well repaid in terms of a smoother process.
There are myriad ways of dealing with the possibility of a zero price. One option would be to prohibit zero or nominal prices, but this may inhibit innovative or beneficial solutions. Another option would be to evaluate total price only. If there are elements for which an authority knows bidders may submit zero prices then another alternative would be to evaluate those prices on a ranked as opposed to inverse proportionality basis (e.g. 5% for the lowest price, 4% for the next, and so on). Careful thought would need to be given to the weighting of such an element to ensure that enabled the authority to identify the MEAT on the basis on which it said it would do so.
Our second example is where the authority adopted a method based on a sliding scale in which scores were awarded by reference to both highest and lowest prices. While this avoided the “middle price prejudice” inherent in inverse proportionality, it can also lead to an outcome incapable of identifying the MEAT.
In this case price was weighted at 40%. The evaluation formula was:
[(Highest price – submitted price) x 40] ÷ (Highest price – lowest price)
This results in the highest price scoring zero and the lowest price scoring 40 (full marks). A French court of appeal has declared this methodology unlawful (Communes de Lognes, no 15PA02953). In that case only two bids were submitted. Applying the above methodology the highest price would score zero even if it was only 1 Euro more expensive. On those rather extreme facts it is unsurprising the court declared the outcome unlawful, though it is reported that the court considered the fact that there were only two bidders was irrelevant (and therefore that the methodology was unlawful in principle). Whether an English court would go that far is not known. But what is clear is that the methodology can have unintended consequences that effectively make price the sole criterion.
What should the authority have done?
One of the difficulties with inverse proportionality is that the scores awarded to Bid A and Bid B are affected by the content of Bid C (which is irrelevant to the respective merits of A and B). One way around this problem is to anchor the evaluation to an objective standard, for example a maximum budget. Where a contracting authority is happy to disclose its maximum budget (which it arguably has to do when stating the estimated value in the contract notice) prices can be scored based on the extent to which they fall below that maximum. A similar approach would be to provide an indicative price (for which, say 20% would be awarded), with extra points being awarded or deducted to reflect the degree to which prices fell below or exceeded it. This would enable bidders to work out the score that they will be awarded for their price, and is very transparent.
It is instructive to see how the sliding scale formula adopted in Communes de Lognes compares with inverse proportionality when applied to the same prices. In the following example the highest price is £100 and the lowest is £50:
|Submitted price||Inverse proportionality||Sliding scale formula|
The sliding scale formula appears “fairer” in that the mid-price (£75) is awarded the mid-score (20%), and the distribution of the scores reflects the distribution of the prices. However, consider the effect when a small number of prices are very closely bunched:
In this scenario Bidder 2 loses a third of the points available despite its price being only 1% higher than Bidder 3. Bidder 2 would have to score 14% higher than Bidder 3 for quality, in order to win. It is very unlikely to have been the authority’s view that a solution that was 13% better quality – and only 1% more expensive – was not the most economically advantageous.
The position of Bidder 1 is even worse. Despite its price being only 3% higher than Bidder 3 it is effectively eliminated from the competition. For Bidder 1, evaluation has been on price only (rather than MEAT based on price and quality).
Our specialist procurement litigation team frequently bring and defend court challenges, both for suppliers and contracting authorities. In the past year alone, we have advised on a range of disputes including: health and social care, infrastructure and development, waste collection and disposal, pathology, defence and telecommunications.
 Malmö Case 5293-10, Swedish Administrative Court of Appeal
 The “middle price prejudice” describes how tenders in the middle of a range are awarded lower scores than might properly be said to reflect their value. For example, where three bids of £100, £75, and £50 are submitted, the middle priced bid would score 66.67% (50/75 x 100) despite the fact that it was positioned exactly in the middle between the lowest and highest bids (which would score 100% and 50% respectively).