Rebuilding a $2bn single-family office around the general ledger

Profile: single-family office, roughly $2bn across liquid, private, and real assets, ~40 legal entities. Names and identifying details are withheld; the structure and decisions are representative of a pattern we see repeatedly.

01The situation

The office had grown up portfolio-first. A performance-reporting platform sat at the centre, fed by custodial and bank feeds, and it produced attractive dashboards. The problem surfaced at every quarter-end: the dashboards and the accounting records disagreed, and no one could say which was right. Capital activity in the private book was tracked in spreadsheets. Inter-entity loans were reconstructed from memory. The annual audit took months.

02The diagnosis

The platform was never designed to be a book of record. It reported on positions; it did not account for them. Because there was no authoritative general ledger beneath the reporting layer, every number was an assertion rather than a balance that tied out. The portfolio-first architecture had optimised the thing the principal looked at and neglected the thing that made it true.

03The rebuild

1A general ledger as the system of record

The office stood up a proper multi-entity general ledger as the single book of record, with the reporting platform repositioned downstream of it rather than beside it. Positions, capital activity, and inter-entity flows all posted to the ledger first.

2Reconciliation as a standing workflow

Daily and monthly reconciliation of custodial and bank activity against the ledger became a defined process with named owners and exception queues, not a quarter-end scramble.

3Private assets brought into the books

Capital calls, distributions, and valuations for the private book moved out of spreadsheets and into the ledger, with primary documents retained against each entry.

04The outcome

Within two reporting cycles the office had one set of numbers that reconciled, an audit timeline measured in weeks rather than months, and dashboards the principal could trust because they descended from a ledger that tied out. The lesson was not that the old platform was bad. It was that reporting is a consequence of accounting — and a family office that inverts that order eventually pays for it at every close.