Designing a multi-party Private Markets Closing Platform.
Multi-stakeholder closing workflow · LP / GP / Legal / Fund Admin
Private markets secondaries closings involve five or more parties, dozens of documents, and dependencies that nobody can see at once. The result is a closing process that runs on email threads, shared drives, and phone calls — slow, opaque, and prone to last-minute surprises.
This is an anonymized reconstruction of real product work. The workflow UI, data, names, and examples are synthetic; the stated outcomes reflect the original programme.
Five parties. Dozens of documents. No shared view of what's blocking the deal.
A secondary transaction involves LP sellers, GP consent, incoming buyers, legal counsel on both sides, and fund administrators — each with different information, different obligations, and different timing requirements.
Without a shared system, each party manages their own view of the deal. The result:
- Nobody knows what is actually blocking the closing
- Document versions diverge across email threads
- GPs spend hours chasing status from counsel and administrators
- Last-minute surprises delay closings by days or weeks
The product opportunity was to give every party a shared, role-appropriate view of the transaction — with blockers, documents, and closing readiness visible to everyone who needs to act.
Problem
A closing is a shared workflow with many owners.
Sellers, buyers, fund managers, counsel, and operations participate in the same transaction — but each has different actions, dependencies, and information needs.
Closing
- What needs to happen?
- Who needs to act?
- What is blocking progress?
- When is the transaction ready to close?
Four decisions that shaped the platform.
- 01
Role-based views, not a single shared interface.
LPs, GPs, counsel, and fund administrators need different information at different moments. A single view for everyone creates noise for everyone. Role context reduces cognitive load and surfaces only what each party needs to act.
- 02
Blockers as first-class objects, not status labels.
In multi-party workflows, a deal stalls when no one knows who is responsible for unblocking it. Making blockers explicit — with owner, type, and impact — converts ambiguity into an actionable queue.
- 03
Closing readiness as a live score, not a checklist.
A static checklist tells you what's done. A closing readiness score tells you how far you are and what matters most. This shifts the conversation from 'are we done?' to 'what's the highest-value next action?'
- 04
Audit trail built in from day one, not bolted on.
Private markets transactions are regulated and disputed. Every state change, document version, and decision needs a tamper-evident record. Treating the audit trail as infrastructure — not a feature — saved significant rework later.
Constraints that protected the product.
Scope decisions in regulated products are as important as features. These are the things I deliberately left out and why.
- 01
No automated signature sequencing
Signature order in private markets closings is legally significant and often negotiated. Automating sequencing without explicit configuration per deal would introduce legal risk the product cannot own.
- 02
No AI-generated closing conditions
Closing conditions are drafted by counsel with specific legal intent. Generating them with AI creates liability without adding speed — counsel will rewrite them anyway.
- 03
No cross-deal portfolio view in v1
GPs wanted portfolio-level visibility, but building it before single-deal workflows were reliable would have created a dashboard showing unreliable data — worse than no dashboard.
Where the product has to be unambiguous.
Potential risks
- Incorrect role assignment exposes sensitive deal data
- Blocker logic missing edge cases in complex fund structures
- Document versioning conflicts during concurrent edits
- Readiness score creating false confidence near closing
- Audit trail gaps under high-volume transaction periods
Controls
- Explicit role assignment with GP-level override
- Blocker ownership always assigned — never unattributed
- Document lock during active edit with conflict detection
- Readiness score caveated with outstanding blocker count
- Immutable event log with timestamp and user attribution
- All state changes require explicit user action — no auto-progression
The measurement definitions we used.
Each metric below is defined by its measurement method — what was counted, how, and against what baseline.
- Time from onboarding to first signature — median days per deal, from role assignment to first executed document
- Blocker resolution time — median hours from blocker creation to resolution, segmented by blocker type and owner
- Closing readiness score at open vs. close — delta between score on deal creation and score at signing
- Document round-trips — median number of document versions exchanged per closing
- Off-platform communications — proportion of closings where parties reported needing to communicate outside the platform
- Return rate within 48 hours of new blocker — proportion of assigned owners who acted within 2 days
Prove single-deal reliability before adding portfolio complexity.
- Phase 1
Single-deal closing
- — Role-based onboarding
- — Document upload & versioning
- — Blocker queue
- — Closing readiness score
- Phase 2
Multi-fund transactions
- — Fund-level grouping
- — Cross-fund blocker aggregation
- — Signature coordination
- Phase 3
Portfolio intelligence
- — GP portfolio view
- — Closing velocity benchmarks
- — Workflow pattern analysis
Where I was involved.
- Owned the product strategy and roadmap for the closing platform from initial discovery through delivery
- Led customer discovery and workflow mapping sessions with LPs, GPs, legal counsel, and fund administrators
- Defined the blocker model, role-permission logic, and closing readiness scoring system
- Translated complex multi-party transaction requirements into data-model logic, edge-case handling, and delivery priorities
- Partnered with engineering and design to deliver scalable workflow automation and operational controls
Cross-functional collaboration
- Product
- Engineering
- Design
- Legal Counsel
- Fund Operations
What I would do differently.
- 01I underestimated how much of the complexity lives in the edge cases of fund structures — specific LP classes, side pockets, and transfer restrictions that don't map to the happy-path workflow. I'd spend more discovery time on edge cases before defining the data model.
- 02The closing readiness score was added in iteration 2. It should have been the primary design surface from day one — it's what GPs actually care about, not the document checklist.
- 03I'd pilot with one fund administrator as a design partner before building the GP portfolio view. We designed for an assumed need; the real need turned out to be different.