Contemporaneous records: build the claim while building the project

The project has two products

Every construction team is producing a physical asset and an account of how it was produced. The second product is usually neglected until a claim arrives. By then staff have left, email is scattered, programme updates contain unexplained logic changes, and the cost ledger cannot distinguish ordinary work from the effects of an instruction.

Good record-keeping is not a claims campaign. It improves control while the project is live. A precise record of restricted access helps the team remove the restriction. Installed quantities against labour hours reveal a productivity problem before the budget is gone. A programme narrative forces the project manager to explain why the critical path moved.

The six categories

Technical guidance The Society of Construction Law Delay and Disruption Protocol organises project records into six groups, and they make a durable spine for any project’s minimum standard. The Protocol is guidance, not law, and its first core principle is that records relevant to progress and to delay and disruption should be created as the works proceed rather than assembled afterwards.

• Programme records: baseline, updates, look-aheads, logic and change reports, narratives and native files.
• Progress records: diaries, quantities, inspections, approvals, milestones, photographs and as-built dates.
• Resource records: labour, plant, shifts, locations, activities, idle time and the reason for it.
• Cost records: cost codes, payroll, invoices, plant, subcontractors, accruals and change accounts.
• Correspondence and administration: notices, instructions, requests for information, minutes, transmittals and decisions.
• Contract and tender material: executed terms, amendments, scope, pricing assumptions, clarifications and risk allowances.

The categories are only useful when they are linked. A notice with no affected activity proves communication, not critical delay. A timesheet with no activity or quantity proves hours, not disruption. A photograph with no place or time may be vivid and nearly useless.

What a good daily entry looks like

Weak: “Client delay. Area unavailable. Twenty men affected.”

The stronger entry fixes time, place, people, work, cause, corroboration, mitigation and linkage. Just as importantly, it avoids a premature legal conclusion. A site diary says what was seen, who was present, when, and from what source. It does not retrospectively declare ten days of excusable compensable critical delay.

The programme must remember why it changed

A monthly update that changes dates without explaining logic is not a contemporaneous analysis. It is a moving picture with its labels removed. For every update, preserve the native file and a read-only snapshot, the data date and update period, actual starts and finishes and remaining durations, every logic, activity, calendar, constraint and coding change, the treatment of out-of-sequence progress, the longest and near-critical paths, a progress and critical-path narrative, links to events and instructions, and who prepared, checked and approved it.

If the employer rejects a programme, preserve the submission, the reasons and the resubmission. Rejection does not make planning disappear, but the contract may attach consequences to acceptance or compliance, so keep planning while the contractual status is resolved.

Record productivity, not merely attendance

Disruption is an efficiency question, so record input and output at a usable level: crew hours plus plant hours plus location plus activity plus installed quantity plus conditions plus interruptions plus rework plus shift pattern. Consistency beats retrospective perfection. If the cost ledger uses one work breakdown, the programme another and the diaries free text, build a translation table at the start and give every event a stable identifier used in notices, schedule narratives and cost accounts.

The event record

Every potential delay or disruption gets an identifier and these fields: first observed date and time, location, affected work, workforce and plant, source or direction, what changed, immediate effect, constraint or critical-path hypothesis, productivity hypothesis, mitigation, photographs and files, programme activity and update, cost code, notice clause, notice date, owner, next action, closure. Facts and contemporaneous observations stay separate from later opinion.

Seven ways records fail

1. Generic entries. “Delay by client” identifies neither the work nor the effect.
2. Late reconstruction. Memory substituted for time-stamped evidence.
3. Programme theatre. Percentages change while logic, actual dates and reason codes do not.
4. Unlinked systems. Notice, schedule activity and cost account use three different names for the same event.
5. Quantity blindness. Hours recorded without installed output, making productivity unknowable.
6. Uncontrolled photographs. No location, date, author, direction or chain of custody.
7. Advocacy in the raw record. Opinion and blame obscuring observable fact, and damaging credibility when it is read back years later.

The evidence graph

One chain runs through the whole system: event, contract clause, notice, affected activity, contemporaneous record, schedule effect, resource and output effect, cost ledger, mitigation, conclusion. An orphaned conclusion is a warning sign. A records health check scores coverage, not claim strength, and identifies missing links without opining on entitlement.

The legal and human reality

Contemporaneous does not mean infallible. A diary author can be mistaken. Automated metadata can be wrong. An email may be advocacy. Tribunals weigh provenance, consistency and corroboration. Nor does a record prove entitlement by itself: notice clauses can demand particular content, recipient and timing, and privilege, confidentiality, personal data and cross-border hosting all require a lawful governance design. The record system must preserve evidence without becoming uncontrolled surveillance.

The operational lesson

The record made on Tuesday morning may be worth more than the polished opinion written three years later. Not because early documents always win, but because they let the later expert test what actually happened. A project that records event, activity, resource, output, notice and cost together has handed its future decision-maker a map instead of a box of paper.