Trial by example

It is a major challenge for contractors to efficiently present claims involving a large number of small items, such as when hundreds or even thousands of variations have been instructed on a project. One possible solution is to use a representative sample.

In Standard Life vs Gleeds and Others, the Technology and Construction Court considered the use of samples, albeit in the context of professional negligence. The case concerned a retail and residential development that Costain had been employed to build nearly a decade ago. Having gone significantly over budget, Standard Life settled with the contractor at double the original contract sum.

The employer’s post-mortem concluded the consultants were to blame; allegedly, there were too many variations, which were often vague and inadequately costed. The upshot was that proceedings were issued against the whole of the consultant team – six defendants including Gleeds, BDP and Carter Jonas.

One difficulty Standard Life faced when drafting its formal legal case was the sheer number of variation instructions – there were 3,600 in total. Rather than setting out full details of each, it prepared its case using a sample of just 122. The pleading recited full particulars of this limited number before extrapolating the results by reference to others, generating a claim of £24m. The defendants inevitably objected, arguing that no assumptions could be drawn from the sample and the case should be struck out.

A balance, of course, needs to be struck. While the court rules expressly require that cases should be dealt with at proportionate cost (indicating evidential flexibility), a claimant still needs to prove its case.

Even the defendants in this case agreed that a sample would sometimes be appropriate. The example given by them was that of a developer who builds 10 houses in exactly the same way and they all suffer the same defect. In such circumstances, the claimant would only need to prove its case by reference to one house, being representative of the whole. The sample house is said to be systemic, meaning the evidence indicates the contractor’s system of working had failed, and therefore is good evidence of them all.

In such circumstances, the claimant would only need to prove its case by reference to one house, being representative of the whole

However, the judge in the Standard Life case considered the emphasis on system failures too restrictive. He said extrapolation from samples may be appropriate in a wider set of cases and it was a matter of applying the ordinary principles of evidence. He gave the example of a situation where the evidence showed the defendant had regularly failed to perform in a certain way, such as consistently failing to read documents properly or report properly. An inference could then be drawn from the sample to be applied to the whole.

Turning to the more common situation of a contractor’s variations claim: It will often be rejected by the employer on the basis of a number of common defences; for example, that there is no formal written instruction. Where there are a large number of variations the only way to bring a case cost-effectively is to ask the tribunal to assess a representative sample. The guidance from the Standard Life case indicates that such an approach is appropriate, albeit the claimant would need to present evidence to show the sample was representative. In particular, that the employer’s reasons for rejection are consistent across the whole and turn on similar facts that can be assessed using the sample.

Where there are a large number of variations the only way to bring a case cost-effectively is to ask the tribunal to assess a representative sample

The sample itself is therefore key. In the Standard Life case, the judge found the sample of 122 was too small. His solution was to allow the defendants to nominate an additional 160 items to ensure it was more representative of the whole.

Cases of professional negligence such as the one brought by Standard Life are on the increase. For consultants, it is worth remembering that professional appointments often specify that the contract administrator must obtain client approval before instructing variations. Failure to do so can amount to a breach of contract. The problem sometimes for consultants is knowing whether an email directing that certain work should take place is indeed a contract variation.

This is one reason why it is becoming increasingly common for construction contracts to stipulate that variation instructions must be in a specified form. For example, under the new 2017 FIDIC forms of contract the instruction must self-identify as a variation. An email that just directs extra work (without specifying it is a contract variation) will not qualify.

This is a very significant change and while employers favour such provisions because they improve the budgeting of projects they also reduce the risk for consultants, meaning claims such as the one Standard Life has brought less likely to arise.

Michael Sergeant is a partner in the construction team at HFW and co-author of Construction Contract Variations