Automating waterfall calculations
Replacing the waterfall spreadsheet without losing the ability to explain every dollar.
How do you automate a distribution waterfall?
Model each tier and definition from the LPA in a waterfall engine, test it against historical distributions until it reproduces what was actually paid, apply investor-level overrides from side letters, then run every future distribution through the engine with a per-investor trace, versioned configuration and approval controls.
Step 1: model from the document
Start from the LPA's definitions, not a template: pref basis and compounding, catch-up rate, tier thresholds, treatment of fees, GP commitment, recallable distributions, and any deal-by-deal features.
Step 2: prove it against history
Run the model against every past distribution and reconcile to what was actually paid. Differences are either model errors or historical errors — both need to be found before go-live.
Step 3: investor-level overrides
Apply side-letter terms as overrides on specific investors rather than building separate models.
Step 4: run with controls
Scenario runs for planning; booked runs for real distributions, with approval before posting. Every run produces a per-investor trace and a variance to the prior run.
Step 5: connect it
The waterfall run should generate the distribution notices and post to the ledger, so no one re-keys its output.
Questions
How long does it take to model a waterfall?
Days to a couple of weeks, depending on complexity and history.
Can we run what-ifs?
Yes; scenario runs never touch the ledger until approved.
Want to see this on your own fund?
Name and work email. A person replies within one business day, usually with times for a demo on a structure like yours.
Talk to us
Name and work email. We reply within one business day, from a person.
Sent.
Someone from FUN/D will email you within one business day. If it's urgent, call +1-212-658-0862.