How to prepare a pitch deck investors actually read
A practical guide for pre-seed and seed founders: what goes on each slide, the five mistakes that kill most decks, and how to pressure-test yours before it reaches an investor's inbox.
Your deck has exactly one job
A pitch deck does not close a round. It gets you a meeting. DocSend's research on thousands of decks puts the average investor's reading time at about three minutes, most of it on a phone. That constraint drives every rule below: the deck must survive a fast, skeptical, partial read by someone who sees dozens of decks a week.
The test is simple: if an investor reads only your slide titles, do they get the argument? If they stop at slide four, have they seen your strongest evidence?
The 12 slides, in order
This is the standard seed narrative. You can reorder around your strengths (traction-heavy decks lead with traction), but every question below must be answered somewhere.
The five mistakes that kill decks
These come up in nearly every deck we review, and each one is enough to lose the reader:
If you have revenue or growth, it goes in the first three slides, not slide nine. Investors decide in minutes; lead with the strongest card you hold.
'1% of a $50B market' convinces nobody. Count your actual buyers, multiply by a price you already charge or can defend, and show the arithmetic on the slide.
'Our team' says nothing. 'Two exits in payments, ex-Stripe CTO' is a headline. Every slide title should be the takeaway, so the deck works when someone only reads titles.
One idea per slide. If a slide needs a paragraph to make its point, it is two slides or it is cut. Investors skim on a phone between meetings.
Ten to fourteen slides for the send-ahead deck. Everything else moves to the appendix. A long deck does not look thorough, it looks unedited.
Test the deck before an investor does
Every deck looks convincing to the person who wrote it, so the feedback loop matters more than any single rule in this guide. The realistic options, roughly from free to expensive:
Set a timer and skim your own deck on a phone. Whatever you did not reach in three minutes, an investor will not reach either. Free, takes five minutes, catches length and ordering problems, but it cannot catch what you are blind to.
Someone who knows nothing about your company reads it without your commentary, then says your business back to you. Where they stumble, the deck fails. Free, but the honest version is rare: friends soften the verdict.
The best free feedback there is, because they know which slides investors pushed back on. Slow to arrange, and you can spend a favor on a draft that was not ready for it.
AI reviewers score a deck against how investors actually read. Several are free (usually as lead magnets for other services); we run one as a Telegram bot: verdict, 0-100 score, and a fix for every slide in about 10 minutes for $5, with a public sample of its full ElevenLabs deck review so you can judge the depth first. Fast and unsentimental; useful on every revision, not just the final one.
Fiverr critiques run $65-150, productized reviews $150-300, ex-VC consultants $300+. Worth it once the deck is near-final; wasted on early drafts that still have structural problems a cheaper pass would catch.
Whichever route you take, one trick multiplies its value: paste the feedback into the AI you draft your deck with and have it apply the fixes while keeping your facts. Collecting the critique is the hard part; the edits are cheap.
Rules of craft
- Numbers over adjectives. "Fast-growing" is noise; "31% month-over-month for 6 months" is a signal. If you have no number, get one before you raise.
- The deck must read alone. You will not be in the room when it is forwarded to the partner meeting. No slide can depend on your voiceover.
- Design is hygiene, not decoration. Consistent fonts, aligned elements, readable charts. A sloppy deck signals sloppy operations. But no amount of design saves a weak argument.
- Write for a smart outsider. No internal jargon, no unexplained acronyms. If a 15-year-old cannot follow the story, an investor skimming at midnight cannot either.
- Dates on everything. Traction without dates is fiction. "10k users" means nothing; "10k users, 8 months after launch" is a growth rate.
The pre-send checklist
- Slide titles alone tell the full story
- Strongest proof point appears in the first three slides
- Market is bottom-up with visible arithmetic
- Every claim has a number, every number has a date
- Competition slide names real alternatives and your winning axis
- The ask states the amount and the milestone it buys
- 14 slides or fewer; the rest lives in the appendix
- Readable on a phone at arm's length
- Sent as PDF, under 10 MB, filename includes your company name
The deck is done when it survives a three-minute skeptical read by someone who owes you nothing. The founders who get meetings are rarely the ones with the best businesses; they are the ones whose decks pass that test.