A seed round is, in the limit, a sequence. You write a list of investors, you ship a deck, you take meetings, you open a data room, you sign SAFEs, the money lands. The steps are well known to anyone who has run twenty rounds. They are mostly invisible to anyone running their first.
Three things have changed since 2018 that first-time founders should internalize:
What follows is the canonical sequence — eight steps — with notes on what's changed, where founders stall, and how the tooling we built handles each phase. We use Arx as the example tool not because we made it but because we made it for this; the principles apply with or without the software.
01. Set the brief.
Before you write a single cold email, write a one-page brief. Company in one sentence. Round size and instrument. Target close date. Lead profile — sector, check size, geography. Three names of partners you'd ideally land. A handful of investors you'd refuse to work with.
Most founders skip this. The result is a pipeline of "everyone we can find" instead of a pipeline of "the right investors for this round." Skipping the brief is the single most common reason a seed round takes eight months instead of three.
A round is shaped before the first meeting. If you can't write the brief, you cannot run the round.
02. Build the pipeline.
Then a list. 80–120 names is reasonable for a seed. Some founders cap at 40 to focus; some build to 200 to wash out the soft yes. There is no universally correct number, but the shape of the list matters: roughly a third leads, a third co-investors, a third angels and friendlies.
The list should be active, not historic. A fund that did three seed checks in the last 90 days in your category is meaningfully different from one that did three checks two years ago. Tools that surface recent-activity signals — Arx, Pitchbook, Crunchbase — make this difference visible.
A common mistake. Sorting your pipeline by fund prestige instead of partner fit. The check comes from a partner. Partners specialize. The hottest fund in your category may have only one partner who actually writes seed checks in your sector.
03. Ship the deck.
The deck is your asynchronous self. Most partners now read it before deciding to take the first call, and many of them will skim it on a phone. That has implications: large type, one idea per slide, a strong slide 7. (More on the slide-7 phenomenon in a separate piece.)
Use a tracked link. The "are you ready" signal is whether your slide-by-slide dwell is U-shaped (everyone reads the intro and the team, drops off in the middle) or top-loaded (everyone reads the intro and quits). U-shaped is fine. Top-loaded is a sign the middle of your deck isn't earning its keep.
In Arx, the heatmap lives inside the pitch deck module — per-slide dwell, per-viewer, with a small badge that flags partners returning to the same slide twice. Whatever tool you use, do not send your deck as an attachment. You will lose the signal you most need.
04. Take the meeting.
The first call is for two things: confirming the partner is serious enough to do real work, and answering the three questions the partner had after reading your deck. Not for re-pitching your deck.
A useful tactic: book the first call yourself, not through a scheduling link, when the partner is from a top-tier fund. Send the link only after the second meeting is in motion. It signals importance early; it scales later.
05. Open the room.
A data room is a packet, not a Google Drive folder. The seed-stage standard packet has 38 items grouped into six sections — corporate, financials, product, market, team, legal — and most founders forget customer references and the 409A snapshot. Both are now table stakes.
Scope per investor. Northbeam doesn't need to see the references Catalyst is reading. Watermark everything with the viewer's email. Set links to expire on a date you set, not when you remember.
06. Model the round.
Once a lead is signaling intent, run a pro-forma. Pre-money, post-money, option pool top-up (typically 10% at seed, sometimes 15% if the lead pushes), founder dilution, SAFE conversion math. Share the pro-forma with the lead — they will run their own anyway, and being aligned on the math before the term sheet conversation saves you a week.
The forecast does double duty here: it shows the partner the runway, the milestones the round will fund, and the next-round defensibility. The model is the bridge between "yes, in principle" and "here's the term sheet."
07. Send updates that work.
During the round, monthly updates to your full pipeline. They keep momentum visible. After the round closes, monthly updates to your signed investors. They keep your next round warm.
The five-block structure most founders should default to: Headline (one sentence), **Highlights** (3–5 bullets), **Lowlights** (1–3 bullets), **Asks** (specific), **Traction** (one chart). Send from your real Gmail address. Investors recognize your name; open rates stay above 80%. Send from noreply@ and open rates collapse.
08. Close the wire.
The unsexy part. Issue and sign the SAFE inside Arx — native e-sign, no third-party tab. Track the signature. Track the wire. The cap table updates the moment the money lands. Send the welcome note. The investors who close fastest are the ones you can rely on to do this again next time; the friction of closing is a quiet signal.
A round that closes in ten days, post-term-sheet, is a healthy round. A round that takes six weeks to close from term sheet is a sign of something — typically a lead being slow, sometimes a side-letter dispute, occasionally cold feet. If it crosses four weeks, ask.
The meta-step.
Throughout all eight, the dominant emotional risk is not getting a no. It's slowly stalling on a soft yes. A "this is interesting, keep us posted" can absorb four months if you let it.
The way to avoid that is to maintain — clearly, visibly, both to yourself and to the partner — a sense of momentum. Updates that show traction. A list of other investors progressing. A close date that exists. Every step should answer the question is this round actually happening?
If you do all eight steps well, a seed round in 2026 closes in three to four months. If you don't, it closes in eight to ten. The difference is mostly process.
This piece is part of a series. Next month: a closer look at the seed diligence packet — what's in the 38 items, and the three documents most first-time founders forget.