Fix the things that compound or carry legal weight first: one source of truth for portfolio data, with a link to each number's source, then audit readiness, then a clear working relationship with your fund admin. LP reporting gets easy once those three hold. Most other ops requests can wait a quarter without real harm.
I hear some version of this question every week from ops people at smaller funds. There's usually one of them, maybe with a fractional CFO for a few hours a week, and requests are coming in from every partner, the fund admin, the auditor and LPs at once. Everything is labeled urgent. Some of it actually is.
The trouble is that urgency and importance come from different people. A partner wants a new dashboard by Friday. The auditor wants valuation support in three weeks. Miss the second one and there are real consequences. Miss the first and you mostly have a partner who's mildly annoyed. So you need an order that doesn't depend on who asked loudest.
The order I'd fix things in
Here's the sequence I'd suggest, and why each step comes before the next:
- One source of truth for portfolio data, with provenance. Every other job pulls from it.
- Audit readiness. It has a fixed deadline and real consequences.
- The fund admin relationship. Their numbers need checking against yours, and you can't check them without step 1.
- LP reporting. Once the first three hold, this becomes assembly instead of research.
As a chain: clean data → audit evidence → reconciled books → LP report. Fix them out of order and you'll do the later work twice. A beautiful LP report built on numbers nobody can trace needs rebuilding the first time an LP or auditor asks where a figure came from.
Start with one source of truth, and know where each number came from
At a lot of small funds, portfolio data lives in a partner's inbox, a few board decks in Drive, a spreadsheet someone last touched two quarters ago and a WhatsApp thread. Nobody can say for sure which revenue number is current.
The fix isn't a new reporting template for founders. Founders already report. Their numbers are sitting in monthly updates, board decks and financial statements you've already received. Asking them to retype it all into your format costs goodwill and gets you nothing new.
Instead, build one place where each metric is stored with its context. An illustrative row:
| Field | Example |
|---|---|
| Company | a Series A fintech |
| Metric | ARR |
| Value | $2.4M |
| Period | Q1 |
| Source | link to the March board deck, revenue slide |
| Received | April 12 |
| Checked by | ops lead, April 14 |
The source and "checked by" fields are what make this a source of truth instead of one more spreadsheet. A number without a source is a liability. When an auditor, an LP or a partner asks "where did this come from?", you click once and show them.
If a number really is missing, ask the founder for only that number. Send it from a real person's email, pre-fill what you already have and let them reply in whatever format they like. Founders answer faster when the request shows you did your homework.
Get audit-ready before the auditor asks
Audit pressure is rising for smaller funds, especially as they reach a third fund or register as an investment adviser. And the pain looks the same everywhere: ten days before fieldwork, someone is digging through Drive and old inboxes for the document behind a valuation, then emailing a CEO who's on a plane.
Build the evidence as you go. For each portfolio company, keep a short evidence index:
- The latest priced round documents, and the cap table after it.
- The most recent financials or board deck you relied on for the mark.
- Your valuation memo, with the method and the inputs.
- Any event since the last round that could move the value: a bridge, a down round, a lost customer, a big new contract.
- A link to each source document, not a summary of it.
Auditors want evidence with some weight behind it: signed documents, board materials, statements from the company. A founder's casual Slack message might be true, but it's thin support for a mark. Know which of your marks rest on thin support before the auditor points it out.
Then run a dry audit a quarter ahead. Pick five companies at random and try to support each mark in under 15 minutes. Whatever you can't find is your to-do list.
Treat your fund admin as a partner you check
Fund admins do essential work, and at some point they will also make a mistake. Most ops people I talk to have a story about a wrong capital account, a missed expense or a distribution posted to the wrong period. That's not a reason to switch admins. It's a reason to check.
Keep your own shadow copy of the numbers that matter most: the schedule of investments, capital called and distributed, and fees. Reconcile it against the admin's reports every quarter. With a clean portfolio dataset from step one, this takes hours instead of days.
A few working rules help:
- One named contact on their side and on yours.
- A shared list of open items with owners and dates, reviewed every two weeks during busy season.
- Agreed turnaround times for quarter-end statements and LP notices.
- Written answers for anything that changes a number, so the reason travels with the change.
LP reporting comes last because it gets easy
LP reporting feels urgent because LPs are the customer. But a report is only as good as the data under it. Once the portfolio data is clean, the audit evidence is filed and the books reconcile, the quarterly letter is mostly assembly.
Don't force every LP into one template, either. Larger LPs often have their own format. With clean data and last quarter's version of each report, drafting the new one should take hours, not a week (and a person still reads every page before it goes out). The LP update guide covers what goes in the letter, and how LPs evaluate platform value covers the platform section.
This is where Cura fits for a lot of smaller funds. It sends metric requests to portfolio companies, follows up on its own, extracts metrics from the replies and documents, and drafts LP updates from that data.
A 30-day triage plan
| Week | Focus | Done when |
|---|---|---|
| 1 | List every open request, score it, agree on the top five with partners | Partners have seen and accepted the list |
| 2 | Pull the last two quarters of portfolio metrics into one place, with sources | Every current number links to a document |
| 3 | Build the evidence index for your ten largest positions | Each mark is supportable in under 15 minutes |
| 4 | Reconcile with the fund admin and log open items | Every difference is explained or assigned |
Thirty days won't finish everything. It does turn "everything is on fire" into a short list that everyone has agreed to.
Which fund ops problems can wait?
Anything that doesn't touch money, compliance or LP trust can usually wait: a new CRM, a prettier dashboard, an internal wiki cleanup, tool consolidation and most nice-to-have reports. Write each one down with a date to revisit, so partners know it's parked, not forgotten.
Some examples of what can usually wait a quarter:
- Replacing a CRM that works well enough.
- Building a partner dashboard nobody has asked for twice.
- Rewriting the internal wiki.
- Consolidating tools to save on a small subscription.
- A one-off analysis a partner mentioned once in passing.
Keep the parked list visible, with a revisit date next to each item. A dated list looks like a plan. An undated one looks like you're ignoring people. And if a parked item comes back three times, that tells you it matters more than you thought.
How do you triage ops work with partners who all want something now?
Put every request in one visible list, score it on risk, deadline and how many people it unblocks, and review the top five with partners together each week. When two partners want different things, let them settle the order in the room, not through you.
A simple score works. Rate each request from 1 to 3 on three things:
- Risk if late: 3 if it's legal, audit or LP-facing, 1 if nothing breaks.
- Hard deadline: 3 if there's an external date, 1 if it's "whenever."
- People unblocked: 3 if it helps the whole firm, 1 if it helps one person.
Add them up and sort. Then share the top of the list where partners already look, like a short Slack message each Monday:
This week's ops priorities: (1) valuation support for the audit, (2) Q2 capital call notices, (3) reconciling the fund admin's statement. Parked until July: CRM migration, new partner dashboard. Shout if the order looks wrong.
Most partners won't argue with a visible, reasoned list. When two of them do disagree, ask them to agree on the order between themselves. Your job is to make the trade-off visible. Picking a winner between partners is a job you don't want.