The first-time fund manager's operating stack: what you actually need before your first capital call

Most first-time managers spend the months before a first close thinking about the thesis, the pitch and the LP list. Almost nobody spends them thinking about operations. Then the first capital call approaches and the questions arrive all at once: which LP committed how much, who has signed, how do I send a notice, where do I record the wire, and what am I going to show these people in ninety days.

This guide is the inventory I wish someone had handed me. It is not about fund formation, which your lawyer owns, or about the bank account, which your bank owns. It is about the operating layer in between: the tools and habits that let a one- to three-person fund run without a full-time operations hire.

The five jobs

Every fund, regardless of size, does five things. The size changes the volume, not the list.

Sourcing. Finding companies, capturing them somewhere durable, and knowing enough about each one to decide whether it deserves an hour.
Relationships. Knowing who you know, who your team knows, and who can make an introduction. In venture, this is the actual product of the firm.
Pipeline. Moving a small number of companies from first meeting to investment or pass, with the diligence and the memo that the decision requires.
Portfolio and performance. Recording what you own, at what price, in which vehicle, and what it is worth now, well enough to compute returns you would defend to an LP.
LP reporting. Capital calls, distributions, quarterly reports and the documents that go with them, delivered to the right people and only the right people.

The trap is that each job has a well-known point tool, so the natural move is to buy five subscriptions and a spreadsheet to reconcile them. It works for a quarter. Then the same company exists in four places with three spellings, the LP roster in the reporting tool disagrees with the one in the cap table tool, and you spend your Sunday reconciling instead of talking to founders.

What each job actually requires

Sourcing: one record per company, with provenance

The minimum is a single list of companies with a normalised website domain as the key, so the same startup cannot be entered twice. Everything else attaches to that record: who owns the relationship, where it came from, sector, stage, notes.

The thing people underestimate is enrichment. You will meet a company and know its name and a founder. Filling in the rest by hand, founding year, team size, prior funding, location, is an hour per company that nobody does, which is why most sourcing databases are half empty. Automated enrichment from public information solves this only if every filled field says where it came from and how confident the source is. A field that says “Series A, 2024” with no source is a liability in a memo. A field that says “Series A, 2024, from a press release dated March 2024, high confidence” is a fact you can use. If you adopt any AI-assisted tool, insist on this. It is the difference between a research assistant and a rumour mill.

Signals are the next layer up: hiring velocity, product launches, GitHub activity for developer tools, filings. They matter because they tell you which of the two hundred companies in your list deserves a call this week. You do not need them on day one. You need the single list first.

Relationships: metadata, not mail

The question you will ask most often in your first year is “who can introduce me to this person”. The answer lives in your team’s email and calendar history. The tools that answer it well read only metadata, who wrote to whom and when, never the bodies. That is a privacy stance, but it is also a practical one: founders and LPs will ask what you connected, and “headers only” is an answer you can give with a straight face.

Relationship strength should be computed, not typed. A field where you rate a contact from one to five decays the day you stop maintaining it. Frequency and recency of real interactions do not.

Pipeline: a board, a checklist and a memo with a guardrail

A Kanban board from Sourced through Screening, Diligence, IC and Term Sheet to Closed is enough process for a fund of this size. Two things make it useful rather than decorative.

First, a diligence checklist per deal that actually gets ticked, so you can see at a glance what has been done and can explain to a co-investor why you are comfortable.

Second, the investment memo. AI drafting a first pass from your notes, the enrichment record and the diligence items is genuinely useful and saves hours. It is also dangerous if the draft can slide into an IC pack unread. Whatever tool you use, the memo must carry a visible “review before finalizing” state until a human marks it reviewed. Nothing should reach an investment committee or an LP that a person has not read. This is not a compliance nicety; it is the thing that keeps an AI summary from becoming an investment decision.

Portfolio and performance: a ledger, not a snapshot

This is where spreadsheets fail last and worst. The instinct is a table with one row per company: cost, ownership, current value. It breaks the first time a company raises a follow-on round you did not participate in, because your ownership changed and your cost did not, and now the row has to remember two things.

Model rounds as a ledger instead. Every round is an entry: date, post-money, amount raised, whether you participated and for how much. Ownership after each round is derived from the entries, not typed. Your cost basis is the sum of what you put in. Current value is ownership times the latest valuation, unless you have a reason to mark differently, and that reason is recorded. Exits are entries too.

Fund-level numbers follow from the same discipline. Called capital comes from the capital call ledger, distributions from the distribution ledger, and TVPI, DPI and IRR are computed from those, not typed into a cell. When an LP asks how you got to 1.3x, you can show them.

Collecting KPIs from portfolio companies deserves its own sentence: send founders a link that needs no login, ask for four numbers, remind them once, and store the answers against the period. Anything heavier than that and the response rate collapses.

LP reporting: audiences, access and one source of truth

Capital calls first, because they come before any report. You need the LP roster per vehicle with commitments, the ability to issue a call as a percentage of commitment or a fixed amount, per-LP allocations, and a status per allocation: notice sent, paid. That is the whole feature. The detail that matters is that the notice, the tracking and the fund’s called-capital figure all read from the same record, so the number in your quarterly report is the number you actually called.

Then quarterly reporting. A report set is a folder per fund per period: the manager letter, the capital account statements, the portfolio update, supporting documents. Two things save you from embarrassment. Audiences, so a document restricted to one LP is never visible to another. And an access log, so you know who opened what, can nudge the ones who did not, and stop guessing whether anyone reads the letter.

Give LPs a portal rather than attachments. A magic-link sign-in with no password, a clean read-only view of their capital account and documents, and watermarked downloads. LPs remember the funds that make this easy.

The order to set it up

You do not need all five on day one. This order matches when each becomes urgent.

When	Set up	Why now
Before first close	Company record, contacts, pipeline board	You are already meeting founders; capture them once
At first close	Fund structure, LP roster with commitments	You cannot issue a call without it
First capital call	Call issuance and payment tracking	The number must be right the first time
First investment	Rounds ledger, position record	Ownership math starts here
First quarter end	Report sets, portal, access log	Your first impression with LPs as a manager
Second quarter	KPI collection, signals, memo drafting	Now you have volume worth automating

What this costs when stitched together

A representative stack for a two-person fund in the United States, at list prices, comes to something between $700 and $1,500 a month once you add a CRM, a cap table tool with LP features, a data room, a portfolio tracker, an enrichment provider and an e-signature seat. The dollars are not the problem. The reconciliation is. Every one of those tools has its own copy of your companies and your LPs, and the copies drift.

The alternative is one system where the company record, the LP record and the fund ledger are shared by every surface, and the AI layer is a co-pilot that proposes and never decides. That is what Investryx is, built because I needed it for my own fund. Pricing is flat per vehicle tier with unlimited seats, and the LP portal is included on every plan with vehicles.

Whichever way you go, do the ledger discipline and the provenance discipline from the start. Both are almost free to adopt at zero companies and expensive to retrofit at forty.

Ready to set it up properly? Start a 14-day trial and have your company list, LP roster and first capital call in place this week.