Audit where your time goes, delegate or automate reactive tasks, set clear boundaries on requests, and agree with partners on the strategic priorities your role should own. Leverage comes from designing systems and making decisions, not being the default for every ad hoc request.
Platform and ops roles at venture firms have a gravity problem. You're helpful, you're reachable, and you know where everything is. So every request finds its way to you: the LP's portal login, the founder who needs a lawyer, the partner who needs a deck reformatted by noon, the event that needs a venue.
I was a startup chief of staff before Cura, so this one feels familiar. Each request is small. Together they fill the week, and the work that would make the role more valuable keeps sliding to next quarter, like redesigning how the fund collects portfolio data or building a real talent program. Platform leads and chiefs of staff describe the same week to me all the time. Many also tell me their role has no written goals ("the role is evolving with me"), which makes every request look equally valid.
Working harder rarely fixes it. Changing what lands on your desk does. My rule of thumb: platform time belongs with founders, and admin belongs to software. Just do the thing, and don't create homework for yourself afterwards.
Step 1: Audit two weeks of your time
You can't move time you haven't measured. For two normal weeks, log every block of work longer than 15 minutes. Keep it rough. A note in a spreadsheet is enough.
For each entry, capture four things:
- What it was. "Pulled Q3 numbers for partner meeting."
- Who asked. A partner, a founder, an LP, yourself.
- How long it took.
- Type. Reactive (someone asked) or proactive (you chose it).
At the end of two weeks, total the hours by type and by requester. Most people are surprised by one of two things: how much of the week was reactive, or how much of it came from one person.
Step 2: Sort the work with an Eisenhower matrix
Take the log and place each recurring task in one of four boxes.
| Urgent | Not urgent | |
|---|---|---|
| Important | Do it yourself. A founder crisis, a fundraise deadline, an LP issue. | Schedule it and protect the time. System design, programs, partner strategy. |
| Not important | Delegate or automate. Data pulls, scheduling, routine formatting. | Drop it or batch it. Nice-to-have reports nobody reads. |
The top-right box is where the higher-leverage work lives, and it always loses to the top-left. So grow it on purpose, by shrinking the bottom two.
Step 3: Delegate or automate the reactive work
For each task in the bottom row, ask three questions in order:
- Can it stop? Some recurring reports exist because someone asked once, years ago. Ask whether anyone still uses it. If nobody notices it's gone, it was never needed.
- Can a system do it? Data collection, meeting notes, intro tracking and newsletter drafts are often good candidates. The post on which platform tasks to automate has a scoring method for this.
- Can someone else do it? An EA, an analyst or a fund admin.
Be careful with self-serve. A resource hub for founders sounds efficient, but in my experience founders rarely go to portals (nobody remembers the link). Answer them where they already ask, with a reply drafted from your own notes. The same goes for partners. Instead of a dashboard they won't open, let them ask the portfolio data directly in Slack or in Claude. Teams chasing each other for internal data is one of the most common time sinks I hear about. A shared data layer is what stops "can you pull the numbers?" from reaching you.
Step 4: Put a front door on requests
Ad hoc requests arrive through Slack, email, hallway conversations and texts. That's what makes them impossible to prioritize. You can't compare what you can't see in one place.
Set up one intake channel. A dedicated Slack channel works well, because people already live there. Skip the form. It's one more place for a request to get lost. Ask for three things:
- What is needed
- By when, and why that date
- Which fund priority it supports
Then set a response rule people can rely on. For example: "Requests in the channel get a reply within one business day with either a date or a trade-off." Partners care more about predictability than speed, and a clear date beats a vague "on it."
Sample wording for the Slack channel description:
Need something from platform or ops? Post it here with what you need, by when, and why. We reply within one business day with a date or a suggested alternative.
Step 5: Agree with partners on what you own
This is the step most people skip, and it's the one that makes the others stick. Once a quarter, sit down with the managing partner, or the partners who matter most to your role. Bring your time audit. Then agree on two or three things your role owns this year, and on how the GP will measure them.
Push for explicit goals and KPIs. I feel strongly about this one. Without them you can't prove impact, and you have no clear path to promotion. Pick the projects with the best return, not the loudest requester. And revisit the list each quarter, because priorities tend to go stale after two or three quarters.
Then count the work. Keep a running value-add log: intros made, hires closed, founder requests answered. At review time, that log makes the case for you.
A RACI chart helps here. For each major area, write down who is Responsible, Accountable, Consulted and Informed. For example:
| Area | Platform lead | Managing partner | Deal partners |
|---|---|---|---|
| Portfolio data collection | Accountable | Informed | Consulted |
| Annual founder summit | Accountable | Consulted | Informed |
| LP quarterly update | Responsible | Accountable | Consulted |
| Board deck reformatting | Not involved | Not involved | Responsible |
Look at that last row. Writing down what you don't own is as useful as writing down what you do. It gives you a reason to say no that isn't personal.
As the function grows, these agreements matter more. When a platform function goes from one person to a team, clear ownership is what keeps the new hires from inheriting the same reactive pile.
What higher-leverage work looks like
It varies by firm, but it tends to share one trait: it keeps paying off after you stop working on it. A few examples from platform and ops roles:
- Designing a portfolio data process that runs every quarter without manual chasing or founder forms.
- Building a talent process that sends candidates the day a founder posts a role.
- Creating an intro process with tracked outcomes, so the fund can show founders proof of value.
- Setting up the reporting that lets partners make faster decisions in their Monday meeting.
What these have in common is that you build them once and they keep running. An hour spent on any of them saves many hours later. An hour spent as the default for ad hoc requests doesn't carry over to next week.
How do chiefs of staff at VC firms prioritize?
Strong chiefs of staff prioritize against a short, written list of the managing partner's goals for the year. Every request gets checked against that list. Work that serves none of the goals gets delegated, batched or declined, and the list gets reviewed each quarter.
In practice, that's a weekly 30-minute check-in with the managing partner. Walk through the open requests, confirm the top priorities for the week, and flag anything that conflicts. Keep a running decision log, so neither of you has to remember why something got pushed back.
How do you say no to partners?
Say no by showing the trade-off. Name what the request would displace, offer a smaller version or a later date, and let the partner choose. Most partners accept a clear trade-off. What frustrates them is a silent delay or a yes that never ships.
Sample wording:
Happy to do this. To hit Thursday I'd need to push the LP update draft to next week. Or I can get you a lighter version by Thursday and the full one Monday. Which works better?
That message doesn't refuse anything. It makes the cost visible and hands the decision back to the person who owns the priority. Over time, partners learn that your time has a cost, and the "quick favors" drop off on their own.