Most overruns are decided before anyone pours concrete.

Xploor was founded on a simple observation: on almost every troubled project in Rajasthan, the problem was already in the ground on day one. Water table, fill depth, bearing capacity, a boundary that did not match the title. Nobody had looked, because looking costs money and winning the tender means quoting low.

So we inverted it. We survey first, at our cost on projects we are shortlisted for, and we quote against what we actually find. It loses us the occasional job to a cheaper number. It has also meant that over fifteen years, our final account has landed within five per cent of the award figure on more than nine projects in ten.

Today we run our own batching plant, employ thirty-one engineers and supervisors on salary, and hold the machinery rather than hiring it. All of that exists for one reason: so the programme we promise is one we control.