Who on site should own the records?

Each record should belong to one named person, whoever stands closest to the facts it captures. The site diary belongs to the site manager or foreman who watches the job happen, because a diary only counts if the person who was there writes it. The registers, of variations, of drawings and revisions, of notices given and received, belong to whoever runs the commercial side, usually the QS, because registers are read against the contract. The photo routine belongs to whoever already walks the job on a rhythm. What fails is shared ownership: a record everyone keeps is a record nobody keeps, and the gap surfaces at the account, each person having assumed another wrote things down. Above the owners sits the second half: a senior monthly check that the records are being kept, because ownership without inspection decays quietly, and by the time a claim exposes it the missing months cannot be rebuilt. Give the check a fixed date and an hour: sample the diary against what the checker knows happened, test the registers against the month's drawings and correspondence, and ask the question that catches most losses early: did anything happen this month that should have triggered a notice nobody sent?

Consider a groundworks contractor running two near-identical jobs. On the first, record-keeping is everyone's job: the contracts manager assumes the foreman keeps the diary, the foreman writes only when something big happens, and the QS keeps a variation register current for six weeks and then quietly stops. When an access claim arises in month nine, the file for the critical period is three diary entries and a photograph. On the second, the same company names names: the foreman owns the diary daily, the QS owns the registers, the contracts manager samples both on the first Friday of every month. In month four the check catches the drawing register slipping and a handover delay nobody had notified; the register is caught up in an afternoon and the notice goes out inside its window. Same company, same systems, same kind of claim. The first settled thin; the second proved its claim. The difference was never the paperwork, it was the names on it.