Ask SALConsolidated Portfolio Intelligence for a Multi-Asset Family Office
The Problem
A second-generation family office managed roughly $45 million across three asset classes: twelve multifamily residential properties, six express car wash locations, and three self-storage facilities. The patriarch built the real estate portfolio over two decades. His children expanded into the operating businesses. The finance infrastructure never caught up.
Every entity lived in its own QuickBooks file. Seventeen of them. The family's controller exported data from each instance, reformatted it by hand, and pasted it into a master Excel workbook for quarterly reporting. That workbook had grown into a 40-tab artifact with color-coded sheets, manual formulas linking across tabs, and a formatting logic that only one person understood. When the controller went on vacation, reporting stopped.
The principals wanted simple answers. How did the portfolio perform this quarter? Which assets are generating the most cash relative to invested capital? Where should we allocate the next dollar? Those questions required the controller to spend two to three weeks pulling, cleaning, and consolidating data across all seventeen QB files. By the time the quarterly PDF packet landed in the family's inbox, the numbers were already a month stale.
Cash flow forecasting was worse. Real estate throws off predictable rent rolls with seasonal maintenance spikes. Car washes swing hard with weather and traffic patterns. Self-storage has long, slow occupancy ramps followed by steady revenue once a facility stabilizes. Each asset class has a fundamentally different cash profile, and the family was forecasting all of them the same way: the controller's gut plus a flat-growth assumption in Excel.
The family had recently passed on two acquisition opportunities because they couldn't model the portfolio-level impact fast enough. That was the breaking point.
The Solution
We built a consolidated data pipeline and reporting layer that ingests from all seventeen QuickBooks instances, normalizes the data across asset classes, and delivers portfolio-level intelligence through an automated reporting dashboard. The work broke into four phases.
Phase 1: Automated ingestion. We connected all seventeen QuickBooks Online instances through the API using a Python-based extraction layer. Each entity's chart of accounts was mapped to a normalized schema that preserves the operational detail each asset class needs—unit-level occupancy for real estate, revenue per bay for car washes, price per square foot for storage—while rolling everything up to a common financial model. The extraction runs nightly. No more manual exports.
Phase 2: Normalized data model. The seventeen charts of accounts used inconsistent naming, different account structures, and varying levels of detail. A car wash location might book chemical supplies under "Cost of Goods Sold." A residential property might book similar maintenance supplies under "Operating Expenses." We built a mapping layer that translates each entity's native account structure into a unified taxonomy. New entities onboard through a configuration file, not a code change.
Phase 3: Automated reporting. The master Excel workbook was retired. In its place: a reporting layer that generates consolidated financials at the portfolio level, the asset-class level, and the individual-entity level. Budget vs. actuals comparisons run automatically against targets the family sets once per year. The reports update daily and are accessible through a lightweight web dashboard with role-based access for the principals, the controller, and their outside CPA.
Phase 4: Asset-class-aware forecasting. We built three distinct forecasting models tuned to each asset class's cash flow behavior. The real estate model uses lease roll schedules, historical maintenance patterns, and local rent comp data. The car wash model incorporates historical weather correlation and same-store traffic trends. The self-storage model tracks lease-up curves against facility maturity. All three feed into a portfolio-level cash flow projection that updates weekly.
The forecasting layer also includes a scenario modeling tool. When the family evaluates a new acquisition, they can plug in the deal terms and see the projected impact on portfolio-level cash flow, debt service coverage, and return metrics before making an offer.
The Result
Quarterly reporting went from a two-to-three-week manual process to a dashboard that updates daily. The controller's time shifted from data consolidation to exception review and analysis. The family now reviews portfolio performance weekly instead of quarterly, and the conversation moved from "what happened" to "what should we do next."
The scenario modeling tool got used within the first month. The family evaluated a seven-unit self-storage acquisition, modeled the lease-up curve against their existing portfolio's cash flow, and closed the deal in three weeks. Before the system existed, that analysis alone would have taken longer than three weeks to assemble.
Total engagement cost came in under $90K across a 10-week build. The controller estimated the annual time savings at roughly 600 hours, and the family's outside CPA cut their quarterly review time by about 40% because the data arrives clean and reconciled.
The broader lesson: family offices often have the data. What they lack is the connective tissue between entities. The value here wasn't AI or machine learning—it was solid data engineering, a normalized data model, and automation applied to a problem that had been solved manually for years because nobody had stopped to redesign the workflow.