Round operations

The meeting note format that makes future follow-up obvious

A six-field decision note turns an investor call into a follow-up you can write in two minutes, even three weeks later.

Jun 15, 20266 min readRound operations

You finish a good investor call. You felt smart, the partner leaned in, you exchanged a few jokes about the cap table. You paste the auto-transcript into a doc, title it "Meridian - call 1," and move on to the next meeting.

Three weeks later that partner emails: "Great to reconnect, where did we land?" You open the doc. It is 2,400 words of you and a stranger talking. Somewhere in there is the thing they cared about and the thing you promised to send. You scroll. You skim. You give up and write a vague follow-up that adds nothing, because reconstructing the meeting would take longer than the follow-up is worth.

The transcript did not fail you. It was never a note. A transcript is a recording of what was said. A note is a decision about what to do next. Most founders confuse the two, and the cost shows up exactly when a round needs momentum: at the follow-up, at the update, at the prep for meeting two.

What raw notes actually cost you

Storage-shaped notes feel productive because they are complete. Nothing is lost. But completeness is the problem. When everything is captured, nothing is prioritized, and the act of deciding gets pushed from the moment you remember the meeting to the moment you have forgotten it.

Three specific failures repeat across founders running their first serious round:

The follow-up arrives late and empty. You wait until you "have time to process the notes," which never comes, so the follow-up goes out a week late with no specific hook. The investor's attention has already moved.

The promise gets dropped. Somewhere in the call you said "I'll send you the cohort retention by month." It is one sentence inside 2,400. You never extract it, so you never send it, and the one thing that would have moved diligence forward dies in the doc.

The second meeting starts cold. You walk into call two without re-reading call one, because re-reading means re-watching a transcript. So you repeat your intro, miss the objection they raised last time, and the partner concludes you are not tracking the conversation. That is the read they take into the partner meeting.

None of these are memory problems. They are format problems. You captured the meeting in a shape that cannot drive the next action.

The fix: write notes as decisions, not records

Right after the call, before the next thing, you write one short note with six fields. Not a summary. Six specific decisions about what this meeting changed.

Signal - what did this investor reveal about their real interest or their thesis? Not "seemed positive." The specific thing: "Keeps circling back to retention. Their thesis is durable consumer, and our month-6 number is their gate."

Concern - the one objection, stated or implied, that stands between this meeting and a yes. If you heard three, pick the load-bearing one. "Worried our CAC will break at scale because current growth is referral-driven."

Promise - exactly what you committed to send or do, in your words, with enough detail to act on. "Send month-by-month retention cohorts and a one-paragraph CAC-by-channel breakdown."

Next action - the single move that advances this relationship. Often it is the promise. Sometimes it is "ask for the intro to their portfolio founder who solved this."

Deadline - a real date, not "soon." "Send by Thu Jun 18."

Owner - who does it. On a two-person team this still matters, because "we'll handle it" is how things fall through.

Six fields. Four to six minutes while the call is fresh. The discipline is that you decide now, when you still remember the room, instead of forcing your future self to decide from a wall of text.

A filled example

Here is the same Meridian call as a decision note instead of a transcript.

Template
MEETING: Meridian Capital - Partner: Dana Liu - Call 1 - Jun 13
STAGE: First call, warm intro via [insert real referrer]

SIGNAL:   Thesis is durable consumer. She returned to retention
          three times. Month-6 cohort is her real gate, not growth rate.
CONCERN:  CAC durability. Current growth is ~70% referral; she thinks
          paid CAC at scale is unproven and that's the bear case.
PROMISE:  (1) Month-by-month retention cohorts, last 12 months.
          (2) One-paragraph CAC-by-channel, current vs projected.
NEXT:     Send both in a single tight email. Offer call 2 in 2 weeks
          with the channel data as the centerpiece.
DEADLINE: Send by Thu Jun 18.
OWNER:    Me (founder). Cofounder pulls the cohort numbers by Tue.

Compare reading that in three weeks against scrolling a transcript. The follow-up writes itself: you open the note, you see the promise and the signal, you send the cohorts with one line that names her concern directly. "Dana - the month-6 retention you asked about, plus the CAC-by-channel breakdown since that's the piece you flagged." That email takes two minutes and lands as if the call were yesterday.

Why these six fields and not more

You could add ten fields. Sentiment, attendees, valuation hints, fund size, next-round timing. Resist it. The note has a job: make the next move obvious. Every field that does not serve the next move is storage creeping back in.

Signal and concern capture why this person matters and what is in the way. Promise, next action, deadline, and owner capture what happens now and who owns it. That is the full causal chain from "we talked" to "the round moved." Anything else belongs in your investor record, not your meeting note.

The format also compounds across the round. Three calls in, your notes are not three transcripts. They are a stack of signals and concerns you can scan in 90 seconds before any partner meeting. The patterns jump out: if four investors all flagged CAC, that is not four objections, that is one narrative gap in your deck, and you would never see it buried in three thousand words per call.

How the note powers the rest of the round

A decision note is the input to three things you will do dozens of times during a raise.

Follow-ups. The promise and signal fields are the whole follow-up. You are not composing from scratch, you are executing a decision you already made.

Investor updates. When you write the monthly update, your concern fields tell you what skeptics are watching. If retention and CAC keep showing up, those are the two charts that belong at the top of the update, aimed at the people still deciding.

Meeting prep. Before call two, you read one note, not one transcript. Signal tells you what they care about, concern tells you what to close, promise tells you what you owe them. You walk in picking up exactly where you left off, which is the single clearest tell that a founder is running a tight process.

Where this connects to RoundOS

The manual version of this works. Six fields, written by hand after every call, is enough to run a clean round. The friction is that you have to do it every time, the notes live in twelve different docs, and nobody surfaces the promise you owe before its deadline passes.

RoundOS works from the sources where the round already lives: your email, calendar, and raw meeting notes or transcripts. Instead of asking you to re-type, it pulls the call into the same six-field shape: signal, concern, promise, next action, deadline, owner, each one tied back to the investor and the thread it came from. The promises with dates become a queue, so the thing you committed to send shows up as a next move before the deadline, not after the investor has to chase you. When you sit down to write the follow-up or the update, the decision is already structured, and the draft starts from what the meeting changed.

The point is not to replace your judgment about what mattered in the room. It is to stop that judgment from evaporating between the call and the follow-up three weeks later.

Try this on your next call

After your next investor meeting, do not save the transcript and move on. Write the six fields first: signal, concern, promise, next action, deadline, owner. Time it. If it takes more than six minutes, you are summarizing instead of deciding. Then, in three weeks, write the follow-up from that note and notice how much faster it is than reconstructing the call.

If you want the same six fields pulled automatically from your existing meeting notes and turned into a dated queue of next moves, that is the part RoundOS handles. Upload one call's notes and see the decision note it produces.

Write the decision note before the transcript goes cold.

After the next investor call, fill signal, concern, promise, next action, deadline, and owner before you save the transcript and move on.