I ask which object, version and moment the record represents.

A 2023 reconfiguration at Branvella replaced forty-eight former lease units with ninety-seven current lease areas. The portfolio extract retained the earlier identifiers and added the later ones. A proposed current count therefore showed 145, treating two generations of the space account as if they described separate current areas.

Nora Bellavern challenged the total by tracing the predecessor and successor relationships. The reporting sponsor wanted to remove the older records entirely because that would produce a simpler current extract. I wanted the current count corrected without discarding the history needed to understand earlier leases and reports. The new areas did not supply a reason to erase the former units, but the former units could no longer be counted as additional current space.

I supported A$3.2 million of data work to maintain dated relationships between the two sets and retire the earlier units from the current view. The mapping retained the relevant source and change period. Leasing and surveying owners remained responsible for the validity and meaning of the boundaries used in their work. My team maintained the account connecting them; it did not create legal boundaries or establish a new physical survey by assigning identifiers.

The result was a current count of ninety-seven lease areas with a traceable relationship to the forty-eight former units. We gave up both the inflated 145 total and the convenience of deleting the past. I did not describe the ratio between the counts as growth in floor space, rental income or asset value. Reconfiguration can change how space is represented without changing those quantities in the same proportion. The case mattered because an apparently simple count had combined records from different states of the property. Preserving the relationship made the current view clearer while keeping the earlier account available for the questions it still needed to answer.

A 2021 leasing report listed 120 signed agreements. By the reporting cutoff, eighty-four had commenced and thirty-six had future commencement dates. The proposed summary described all 120 as occupied, using the existence of a signed agreement to make a statement about a different condition.

The commercial-report sponsor preferred the single signed total because it showed the commitments already secured. That was a legitimate account to retain. I objected to the occupancy label rather than the signed count. Commencement and physical occupation were also not interchangeable: reaching a contractual date did not, by itself, prove how the tenant was actually using the space. The report needed to identify its evidence instead of allow one familiar property word to bridge several meanings.

I chose to preserve signed status and commencement status separately, with A$1.4 million of reporting work. The current account showed eighty-four commenced agreements and thirty-six awaiting their stated commencement, while retaining the total of 120 signed. Leasing and legal colleagues interpreted the agreements and relevant conditions. Building operations retained observations about actual use. The data function did not infer occupation from a signature or settle contractual questions merely by comparing dates.

We gave up a more optimistic-looking occupancy headline while keeping the evidence of signed commitments visible. The output was a clearer status account, not a measured improvement in leasing or a new rental-income result. I did not claim that the thirty-six future agreements were uncertain or invalid because they had not commenced. They simply belonged to a different part of the account at the cutoff. A property-data leader helps when the summary lets management see both present status and future commitments without making either population disappear or allowing one to stand for the other.

At Ternavell Infrastructure in 2009, I approved a project-information account containing ninety-six floor plans. The selection treated the latest uploaded file as the latest approved version. A review against the revision register identified twenty-eight of those plans as superseded. Upload order had been used as a shortcut for a documentary status it did not establish.

Evan Tarswell challenged the selection by returning to the revision evidence. My initial preference had been to keep the latest-file convention because it was easy for users to understand and maintain. I had not adequately distinguished a recent upload from an approved revision. An older file could be uploaded again, or a new file could exist without the status the report implied. The convenience of a chronological file list did not make it an approval record.

I accepted responsibility for the assumption and supported A$480000 of review and correction. We retained the approved revision reference, source, date and document status instead of deriving approval from upload position. The twenty-eight superseded plans were identified in the reviewed account. The other sixty-eight were not flagged by that exercise, but that fact was not presented as a universal certification that every drawing was current or correct for every purpose. Project engineering and the responsible document owners retained their approvals.

The change gave users a more defensible version account and required me to explain why the earlier selection rule was inadequate. It did not amount to a structural inspection, safety assurance or confirmation of as-built conditions. The case remains in my public narrative because the data function can make a consequential documentary claim without changing the contents of a drawing. My error was in the status I allowed the account to imply. The correction needed to preserve the evidence of approval, not merely replace one convenient latest-file list with another.

A 2017 review described one Branvella building as 22000 square metres on an external-extent basis and 18000 square metres in its lease schedule. The proposed exception report treated the 4000-square-metre difference as missing space. That interpretation assumed the two sources were measuring the same boundary for the same purpose.

The portfolio sponsor wanted a single area value to remove disagreement between reports. I wanted the surveying and leasing owners to explain the bases first. An external-extent account and a lease-schedule account can include different parts of the place. Their numerical difference does not establish a physical loss, an error in every source or an opportunity to charge rent for the remainder. Selecting one number without its definition would only move the ambiguity into the chosen field.

A$1.7 million supported data work to retain the purpose, measurement basis, source version and responsible owner beside each area account. We did not claim adoption of a particular measurement standard or invent a conversion that made the two views identical. The relevant specialists confirmed which definitions were appropriate to their use. The data team maintained the relationship and made the difference intelligible to a reader who had previously seen only two apparently conflicting totals.

The result was a clearer explanation of the two quantities, not newly discovered lettable space or a revised valuation. We gave up the convenience of an unqualified single area and withdrew the missing-space interpretation. The case taught me that reconciliation does not always mean forcing records to agree numerically. Sometimes it means preserving why they differ and preventing the next user from stripping away that reason. A property account becomes more useful when it tells the reader what was measured and for which purpose, rather than present precision as evidence that every source has the same boundary.

The 2025 portfolio review contained eighteen valuation records. Thirteen were dated 30 June 2025 and five were dated 31 December 2024. The proposed summary applied a uniform 30 June label to the collection, making it appear that every assessment referred to the later date.

The reporting sponsor wanted one clean portfolio heading and argued that the earlier values were still the latest available for those properties. I agreed that they belonged in an account of the latest available records if their dates remained visible. I did not agree that availability made them contemporary with the other thirteen. A current report can contain evidence from different dates; calling the report current does not update the underlying assessments.

I authorised A$2.6 million of data work to preserve the date and source of each valuation record and to separate the two dated groups in the review. We withheld the uniform 30 June valuation label. Professional valuation and finance owners retained their assessment and reporting conclusions. The data function did not roll the five earlier values forward automatically, infer appreciation or claim that a common date could be created through a presentation change.

The cost was a less uniform-looking account and a more explicit conversation about which evidence was available. The delivered output was a dated portfolio view, not a new valuation or a conclusion about portfolio fair value. I did not describe the five earlier records as wrong simply because they were older. Their suitability depended on the question and the relevant owners’ assessment. The case matters because a heading can add an unsupported temporal claim to otherwise accurately copied numbers. Keeping the dates attached allowed the reader to see where further valuation evidence might be needed rather than assume the data report had already supplied it.