What belongs in a handover file

At completion, a set of documents changes hands — to a buyer, to a society, to a facilities team. It is called the handover file, the O&M manual, the completion set. The name varies. The problem does not.

The file is almost always assembled in the last fortnight of a job that took years, from documents created by people who have since left. That single fact explains every gap in it.

What it should contain, and why#

The test for each item is not "was it important during construction". It is "will somebody need it in three years, when nobody who built this is reachable".

As-built drawings. Not the tender drawings, not the good-for-construction set — the drawings as the thing was actually built, with the deviations marked. The first person to drill into a wall needs to know where the conduit actually went.

Approved drawings and the approvals themselves. The stamped set and the sanctions attached to it. Later — at resale, at a loan, at any alteration — somebody will ask what was approved rather than what was drawn. Requirements differ by state and by authority, so verify what your own jurisdiction demands rather than copying a list from a previous project.

Equipment manuals and warranty documents, with their start dates. A warranty without its start date is not a warranty; it is a leaflet. The date it begins, what it covers, what voids it, and the name of the party who honours it. That party is often the supplier, not the contractor.

Test certificates. Concrete cubes, water, electrical, lift, fire, pressure tests on plumbing. These are the evidence that the thing was verified when it could still be verified. The cube results in particular are unrepeatable — the argument about why they matter and what they actually prove is in the concrete cube test.

Commissioning records. Who started the system, on what date, with what settings, and what was observed. When a pump behaves oddly two years later, the first useful question is how it behaved on day one.

Service provider contacts. The firm that installed the lift, the one that serviced the fire panel, the electrical contractor. Names and numbers, and whose contract they sat under. It is the item most often skipped and the one most often needed first.

Meter readings at the moment of handover. Electricity, water, gas, with the date and the meter serial. This one line prevents a dispute about who owes for consumption before the handover, and it is impossible to reconstruct afterwards.

The snag list in the state it was on the day. Not a clean sheet. The open items, their owners and the agreed dates. A handover claiming there were no open snags is believed by nobody — the discipline is in a snag list that actually closes.

A signed acknowledgment of what was actually handed over. An index, signed by both sides, listing what is in the file and what is not. A file with an honest list of its gaps is far more useful than one implying a completeness it does not have.

Keys, access cards and administrator credentials. For the access control, the BMS, the CCTV. These are routinely handed to one individual and lost when he moves on.

Why it is always incomplete#

Because it is assembled at the end. That is the whole of the answer; the mechanics are worth one level of detail.

The documents were created over years by many parties. A test certificate went to a site engineer who filed it in a cabin since dismantled. A warranty came by email to a purchase officer who has left.

The people who could identify a document have gone. Somebody finds a stack of certificates in the last fortnight and cannot tell which pour they belong to, because the detail lived in the head of the man who received them.

Nobody's performance depends on it. The project is measured on completion and on cost. The file is a task assigned late, to a junior, with no authority to compel the subcontractors who hold half the documents — and those subcontractors have already been paid, or are arguing about their final bills.

The commercial pressure runs the other way. Everybody wants the handover to happen, and a missing manual is not going to stop it. The file is accepted as it stands and the gaps become somebody else's, later.

File at creation, not at completion#

There is one habit that fixes this, and it is not a bigger effort at the end.

When a document is created, file it then, in the place it will finally live, with the identifying detail attached while somebody still knows it. The cube result goes against the pour on the day it arrives. The warranty goes against the equipment on the day it is received. The as-built markup is made when the deviation happens, not reconstructed a year later.

This is not more work in total. It is the same work, done when it is cheap rather than when it is expensive, by the person who knows the answer rather than the one who has to guess it.

The practical form is an index defined at the start of the project — the sections the file will have on the last day — and a rule that every certificate, manual and approval enters its section on arrival. Then the file is already assembled, and somebody checks it instead of building it.

A copy is not the record#

There is a difference between handing over a copy and handing over the record, and it decides what the file is worth in a dispute.

A copy is a scan or a photocopy in a binder. It proves what somebody chose to put in a binder. Anybody who has assembled such a file knows how easily a page is swapped, a date corrected, a version quietly replaced with a later one.

Handing over the record means the receiving side gets the document with its provenance: when it was created, by whom, what it superseded, and evidence it has not been altered since. That is the difference between a document you possess and one you can rely on, and the mechanics are in tamper-evident documents.

It matters most for the items argued about later — as-built drawings, test certificates, warranty start dates — the pages where a small change is valuable to somebody and invisible to everybody else. A file whose contents cannot be shown to be the originals is evidence of good intent, not of fact, and what makes a handover fail usually comes down to that gap.

The short version#

The handover file is needed years after handover, by people who cannot ask anybody. Its contents should be chosen on that basis: as-builts, approvals, manuals and warranties with dates, test and commissioning records, contacts, meter readings, the open snag list, and a signed index naming the gaps.

It is incomplete because it is assembled at the end. File each document at creation instead, in the section it will finally sit in. And hand over the record with its provenance, not a binder of copies.