In the first article, I discussed why multi-custodial data aggregation has become a strategic necessity for wealth management firms. In this follow-up, I want to go one level deeper to explain why this problem is so difficult to solve and why most solutions fail long before firms realize it.
From the outside, multi-custodial aggregation can look like a data plumbing exercise: connect to custodians, pull files, normalize fields, and present a unified view. In practice, it is one of the most complex engineering and operational challenges in financial services.
Custodian Data is Hard by Design
Every custodian produces multiple categories of files, including positions, transactions, pricing, tax lots, accounts, and accruals. Although these labels sound standardized, they are not. Across custodians, files differ in naming conventions, data types, identifier schemes, formats, delivery timing, and completeness. Even within a single custodian, formats and logic can change over time.
Take something as simple as a dividend reinvestment. The same economic event can be represented in several ways, including: a single "Dividend Reinvest" transaction that increases units and records the reinvested amount, a two-line pattern (a cash dividend followed by an automatic buy for the same amount), a journal entry that moves income to principal and updates units, or a non-cash/DRIP code that increases units without a separate cash line.
Depending on the custodian, the cost basis may be available, partially populated, or missing entirely. Each variation requires contextual interpretation, not just technical parsing.
Custodial systems were built independently and optimized for internal processing rather than for external users. Any system attempting to unify them must accommodate that reality. Below are common areas where discrepancies occur across custodians, data, and file types.
1 | Identified Inconsistency: Security identifiers are the foundation of accurate reporting and one of the most common sources of failure. Custodians use different identifiers: CUSIP, ISIN, SEDOL, Tickr, proprietary symbols, RICs, or sometimes nothing at all. International securities, private investments, and alternative assets frequently lack clean, persistent identifiers. Without price mapping logic, these inconsistencies result in position mismatches, pricing errors, performance distortions, and broken reporting.
At PCR, identifier normalization requires crosswalk tables, fuzzing matching, security master enrichment, and edge-case handling for marketable, OTC, alternatives, and fractional securities. This isn't a one-time mapping exercise. It's a living system that evolves as custodians change behavior and new instruments emerge.
2 | Transaction Codes: Transaction codes represent another layer of complexity. There is no universal standard. A buy might appear as BUY, BY, B, or a numeric code. Corporate actions are even more fragmented, with each custodian applying its own taxonomy and logic.
PCR translates and maintains thousands of custodian transaction codes into approximately 200 industry-standard codes with aligned mappings and adjustment rules. This standardization ensures consistent treatment across custodians and supports accurate performance, accounting, and reporting.
3 | T + 1 Completeness: Being "T+1 complete" is far more than receiving a daily file. It requires all positions, transactions, cash movements, income events, prices, tax lots, and accruals to be available and internally consistent every day. In practice, custodians frequently miss components. Files arrive late, partially, or not at all. Accounts may drop. Pricing may be delayed. Income accruals may lag positions. Tax lots may appear days after trades settle.
PCR addresses this with missing file detection, unit roll-forward logic, price hierarchy enforcement, transaction inference algorithms, and structured exception routing back to custodians.
4 | Reconciliations: True multi-custodial readiness requires daily reconciliation across multiple areas: completeness checks, position movement validation, pricing consistency, cash reconciliation, and historical anomaly detection. It also requires operational workflows for missing files, partial files, structural inconsistencies, and unexpected values with clear escalation paths across automated systems, internal operations teams, and custodian counterparties.
This is where many aggregators fall short. Technology alone is insufficient. Without a strong operational infrastructure and direct counterparty engagement, errors persist and trust erodes.
How PCR Solves the Problem End-to-End
PCR was built specifically to solve this problem on an institutional scale. We operated through LOA-based, direct custodial feeds that provide authorized, reliable access to data. That data flows through a multi-layer normalization engine that harmonizes transactions, identifiers, cash movements, and prices. The data is reconciled through institutional-grade validation logic and governed by automated exception management workflows. Where appropriate, we apply AI-driven anomaly detection to surface issues faster and improve accuracy over time. The result is a platform that delivers T+1 ready data consistently.
A Hard Problem Worth Solving
Multi-cusotdial data processing is one of the hardest problems to resolve in WealthTech. At PCR, we've spent years solving it because the firms we support cannot afford to get it wrong. Accurate, complete, reconciled data is the prerequisite for better advice, better decisions, and better outcomes.
