Seed pitch deck, answered.

Twelve to fifteen body slides plus an appendix. Seed investors will read further than pre-seed investors will, but only if the first three slides earn it. A bloated twenty-five-slide deck reads as: founder doesn't know what the bet is yet.

There's no universal bar. The patterns are: $10–30k MRR with month-over-month growth and visible retention, or five to fifteen design partners on signed pilots, or strong consumer usage if that's your motion. Investors buy the slope more than the level. The shape of the proof matters more than the size of it.

Most US software seed rounds in 2026 land between $2.5M and $6M. The common median is $3.5–4M. Size to eighteen to twenty-four months of runway and the specific Series A milestone you're underwriting, usually $1–2M ARR with healthy retention. Don't raise to a round size. Raise to a milestone.

A seed deck has to defend numbers. Pre-seed leans on insight and team. Seed adds three burdens: a real GTM story (channels, CAC, payback), retention curves you can defend, and a financial model that survives a junior associate's spreadsheet. Same narrative shape. Heavier proof.

Slide three or four. Before market sizing, before product. If you bury traction at slide ten, you've spent the partner's attention on slides they would have skipped if the numbers had carried it. Lead with proof.

Yes. Not the 2x2 grid where you sit in the top right. That slide has been satirized for ten years. Use a 'who picks what' frame: which customer profile picks each competitor today, and why your wedge wins one specific profile decisively. That's the only competitive slide a partner will remember an hour later.

Eighteen to twenty-five percent on a priced seed is typical. Add residual SAFE conversion from any pre-seed and you can quietly find yourself at thirty percent before Series A. Model dilution at your projected Series A cap before you sign. It's the most-skipped check in a seed term sheet.

If you've sold to more than thirty customers, yes. If you haven't, replace it with a unit-economics-by-channel sketch: which channel works, what payback looks like, what changes at ten times scale. A confidently wrong CAC slide is worse than no slide.

Showing growth without showing retention. A hockey-stick chart with no cohort retention next to it tells a sophisticated investor that either you don't track retention or it's bad. Pair every growth chart with the retention that earned it.

Eight to fourteen weeks from first meeting to wired in 2026, for rounds that close. Partner meetings cluster in weeks four to eight. References and diligence sit in weeks six to ten. If you're past sixteen weeks without a term sheet, recut the deck. The read is broken, not the market.