Why do global claims keep failing?

Because blaming everything at once tends to prove nothing in particular, so a single cause the claimant is responsible for can sink the whole number. A global claim (sometimes called a total cost claim) says the job cost more than was allowed for and puts the whole difference down to the other side's events lumped together, without linking any one cause to any one pound of the extra cost. That structure hands the defence its answer on a plate: some of the overrun was almost certainly the claimant's own tender error, inefficiency or subcontractors, and because the claim cannot separate those causes from the rest, a court or adjudicator can conclude that none of the sum is properly proved. Global claims can succeed, but they carry a heavy risk and usually only survive in narrow circumstances. The reliable cure is to break the claim down, event by event, each with its own established effect and quantified cost, and to concede honestly the parts that are genuinely yours, because a smaller claim that is fully proved is worth far more than a large one that collapses on its weakest link.

Picture a steelwork contractor whose job overran by twelve weeks and cost £300,000 more than tendered. Its first draft claim simply asks for the whole £300,000, saying the employer's late drawings, late access and design changes together caused the lot. In defence the employer points out that part of the delay was the contractor's own late steel deliveries and that its tender had under-priced the connections. Because the claim never separated the employer's events from the contractor's own problems, the adjudicator cannot tell how much of the £300,000 is truly the employer's fault, and the claim is at serious risk of being rejected. Rebuilt properly, the contractor links four weeks and £95,000 specifically to the late drawings with dated records, three weeks and £60,000 to the access problem, and honestly drops the part driven by its own deliveries. The smaller, event-by-event figure is far more likely to be recovered than the impressive but unproved total.