Investment reporting is only as reliable as the data that feeds it. Family offices hold assets across several custodians, currencies and legal entities, and their platforms are populated by onboarding custodian statements into a general ledger and then reconciling the ledger against the source statements. This paper studies that cycle empirically using aggregated operating data from a family-office WealthTech firm over June–August 2026: 2,161 reconciled transactions, 82 client accounts and four investment modules (direct equity, mutual funds, bank/cash and credit, and tax lots). The overall break rate was 7.1% (154 breaks). Break rates differed across modules, from 4.3% for mutual funds to 11.8% for tax lots (χ² = 19.0, df = 3, p < .001), and the mix of causes also differed by module (χ² = 70.9, df = 12, p < .001). Data-entry errors and timing differences explained 63% of breaks. Tax lots combined the highest break rate with the longest turnaround and, with bank/cash, absorbed an estimated 73% of resolution effort. The monthly break rate fell from 8.9% to 5.6% (trend z = −2.48, p = .013), consistent with a learning curve; the single-family versus multi-family difference was not significant (p = .137). A worked example shows that the cost-basis method alone can change a realised gain by 44.6%. A scenario analysis sizes the benefit of shifting control effort from detection to prevention.

Keywords: reconciliation, internal control, operational risk, tax-lot accounting, cost basis, family office, WealthTech, data quality