Primary · Internal
Choosing our
CRM.
Why the choice matters, what's broken today, and the real trade-offs behind every option.
The stakes
A people
business.
Our network is the asset. The CRM is where it lives — or where it leaks.
All we do here is find people — but they get lost.
Our Affinity-Lovelace marriage is rocky.
—Logging is still manual
Logger was a huge win, but capture is hand-work — and a CRM's whole value is automation.
—Pipeline lives in two places
Lovelace is closer to what we want, but now what-to-sync and which-way is a constant battle.
—Not everyone's on it
Seats cost too much, so half the firm's network never makes it in.
The field
Two shapes
of CRM.
Get the category right first — most CRMs solve a problem we don't have.
We have no customers.
We have a network.
—Sales CRMs
Salesforce and kin drive a pipeline toward a close. Built for customers.
—Relationship CRMs
Map who-knows-whom across years. Built for networks like ours.
—The tell
A VC never “closes” a founder. The relationship is the product.
—So the field narrows
Relationship-native is the only family worth our time.
Affinity — the
relationship CRM
we already run.
+Pro — Purpose-built for VC
Auto-captures every email and meeting, out of the box.
+Pro — Shared graph
Enriched by network data every other fund feeds the platform.
−Con — The pipeline stays manual
It captures contacts, but the master deal pipeline we live in won't auto-update.
−Con — Not builder-friendly
A walled API: rate limits cap what we build, and scale costs more.
Ring — the
VC-native
challenger.
+Pro — Deep enrichment
Backfills 20–30 years of email; maps cap tables, boards, advisors.
+Pro — Built to extend
50–60 tools and a REST API — some firms run it headless.
−Con — Unlock's not shipped
Ring-to-Ring deal sharing is still in design — targeted, not guaranteed.
−Con — Still rented
Seat-based paid pilot, and LinkedIn is manual CSVs.
Zero — AI-native,
wrong shape.
+Pro — Genuinely AI-native
Zero-click capture across email and LinkedIn. Ex-Smartly.io, 20VC-backed.
−Con — But it's a sales CRM
Built to prospect and close — the Salesforce shape, not ours.
−Con — And too early
Design-partner stage: no general release, no public price.
—Verdict
Inspiration for an AI-native tool — not one to adopt today.
The build-your-own option
Own
the stack.
Twenty is open source — the one option we can shape. Two ways to run it.
Twenty, hosted.
They run it.
+Pro — $9 a seat
Custom objects, APIs, workflows, and the app builder — all included.
+Pro — Vibe-code on top
Build our own apps and AI agents inside it, no infra to manage.
−Con — App builder is brand-new
The custom app-building feature we'd build on is ~3 months old and changing weekly. Expect churn.
−Con — And boxed into it
Customization is restricted to that app builder — we could lose things like event listeners for agents.
Twenty, self-hosted.
We run it.
+Pro — Free license
Open source. Infra runs ~$200/mo, right beside Lovelace.
+Pro — One database
Lovelace and Twenty share one Postgres — no sync lag or engine to build.
+Pro — Anyone can vibe-code it
We control the source, so people across the firm could vibe-code solutions to their own problems.
−Con — But it's ours to run
Upgrades, backups, security — we own all of it, and that's real, ongoing work.
The economics
~$200/mo
self-hosted for the whole firm — flat forever, while hosting charges by the seat.Where we land
Toward
Twenty.
The only relationship-native option we can fully control — likely self-hosted.
What's left to decide
One direction, two open questions
- 01PrivacySelf-hosting means we hold everyone's emails and calendar events ourselves — encrypted at rest, but still ours to protect.
- 02A long-term commitmentSelf-hosting is significant, ongoing effort — are we ready to own it for the long haul?