
A Midwest real estate investment manager with over $10 billion in AUM maintained a vertically integrated fund operations model. Their in-house team owned every step from capital activity to LP reporting, which gave them strong control and operated as a real differentiator in fundraising. The cost: an operational stack built on Yardi IM and a layered set of Excel workbooks that grew more fragile with each new fund and side letter.
Quarter-close required the controller team to reconcile capital account balances calculated in Excel against Yardi IM outputs. Discrepancies surfaced every cycle. The fee model that handled side-letter overrides for the firm's institutional LPs was held together by one senior associate's formula structure. When she went on parental leave, the team rebuilt the model from scratch over four weeks.
Investor reporting demands compounded. Pension funds and sovereign wealth LPs asked for sector-level performance breakdowns, vintage-blended returns across their commitments, and quarterly fee transparency disclosures that Yardi IM could not produce natively. The finance team built each report manually, which meant a $2B follow-on fundraise was competing for staff time with the quarterly close.
Scaling the firm meant either doubling the finance team (proportional headcount growth they had explicitly rejected) or rebuilding the operational stack.
The CFO's mandate to the finance team was specific. Replace the Excel layer entirely. Keep Yardi IM as the property-level general ledger, because the asset management team's workflows depended on it. Build a parallel fund-level computation system that could handle the firm's specific waterfall structures, side-letter overrides, and scenario modeling needs without forcing manual reconciliation against the GL.
The team evaluated three categories of tools: fund admin platforms that would have required outsourcing the operational layer the firm had built as a moat; standalone waterfall engines that solved one calculation problem but left the LP-level allocation work in Excel; and integrated fund management platforms that could replace the Excel layer wholesale while connecting to the existing GL. Maybern fit the third category.

Implementation ran in parallel with quarterly close. Maybern's Solutions Architects encoded the firm's three fund vintages, their hybrid waterfall structures, and the side-letter terms that had previously lived in a private OneNote. Historical data from the prior eight quarterly closes was loaded and reconciled against the existing Yardi records.
The first measurable change was time. Waterfall and investor allocation prep, which had taken nine to eleven working days per fund per quarter, dropped to two to three. The team redirected the recovered time to scenario analysis, fee strategy review, and LP communication.
The second change was the kind of question the finance team could answer.

The catch-up calculation that previously required a senior controller to build a one-off Excel model over a day and a half now runs as a query, with the result traceable to the underlying capital activity and the LP's specific side-letter terms.
The third change was structural. The fee model that had depended on the senior associate's formula structure is now a configurable rule set inside Maybern, with logic documented at the level of the LP's contractual terms rather than at the level of Excel cell references. The key-person dependency is gone.
Audit trails for every calculation are automated. The fund's auditor receives traceable lineage rather than a stack of workpapers. The firm's first full close on the new system completed two days ahead of schedule.
The firm is now underwriting fund four with operational headroom they did not have before. Two structural decisions previously deferred (a parallel sidecar vehicle for a strategic LP segment, and a co-investment program with vintage-blended fee economics) are back on the table because the operational cost of running them is no longer prohibitive.
The bet on vertical integration was a contrarian one in an industry trending toward fund admin outsourcing. The firm is doubling down on it.
Quick reference for this topic.
The firm in this story chose to replace the Excel layer entirely while keeping Yardi IM as the property-level general ledger. The operating principle was straightforward: outsource what is commoditized, keep in-house what is differentiated. Fund administration is increasingly commoditized. The fund-level computation that handles waterfalls, allocations, and LP-specific terms is the moat, and that stayed in-house, just on better infrastructure.
For a $10B RE manager with three vintage funds and roughly 50 institutional LPs, the in-house operational stack required two senior controllers plus a fee model maintained by a third. Before modernization, those three roles spent the majority of their time on reconciliation and report production. After, those same three roles handle materially more work because Excel reconciliation has been removed from the workflow entirely.
The signal is usually a fundraising cycle the firm cannot support without doubling finance headcount. For this manager, the trigger was the IR team's increasing volume of LP-specific scenario requests during a $2B follow-on fundraise. The Excel stack could service the fund. It could not also service the questions an institutional LP was going to ask before allocating.
The catch-up is the contractual provision that allocates carried interest to the GP after LPs receive their preferred return, until the GP reaches their full carry share. The math is straightforward but path-dependent: side-letter terms, individual LP commitments, and fund-close timing all change the answer. In Excel, recalculating the catch-up for one LP at a specific fund-close scenario is a half-day project. In a configurable rule system, it is a query.
For this firm, no. Shadow accounting exists because GP teams do not trust the fund administrator's outputs and run parallel records to catch errors. When the GP team operates as the source of truth and Yardi IM serves as the property-level GL, there is no third-party administrator to shadow. The reconciliation work shifts from finding the admin's errors to confirming the property-level GL matches the fund-level allocation logic, which is structural verification, not error-finding.
