Every founder asks the same question before their first raise — what does a pitch deck that actually works look like? The honest answer is that it looks like a great many different things, and the founders who study only one example tend to produce decks that resemble a costume more than a document. What follows is a survey of twenty-three decks that closed institutional capital. Each is annotated with the single thing it does that you should take seriously.
- Airbnb — the discipline of a clean problem.
- Uber — selling a future that does not yet exist.
- Tinder — when the product is the pitch.
- Brex — what to do when traction can carry the deck.
- Notion — taste as a commercial asset.
- LinkedIn — naming the failure of your predecessor.
- Facebook — one number, well chosen.
- YouTube — the case for unembellished prose.
- Snap — the deck as manifesto.
- Mixpanel — repetition as proof.
- Eleven habits common to the entire survey.
- Five errors that disqualify an otherwise strong deck.
I.Airbnb. The discipline of a clean problem.
The most circulated pitch deck in living memory, and properly so. Slide two consists of a single sentence: price is an important concern for customers booking travel online. Slide three: hotels leave you disconnected from the city and its culture. The problem is so plainly stated that the solution lands harder when it arrives. There is no rhetorical fog to clear before the reader can evaluate the proposition.
II.Uber. Selling a future that does not yet exist.
The original deck was titled UberCab — long before the name change, and unusually long for a seed pitch at twenty-five slides. It reads less like a modern pitch deck and more like a business plan from the prior era. The deck's most durable strength is its vision section, which renders a city without the inefficiency of street-hailed cabs as something inevitable. The reader does not need the product to be operational to believe the world the deck describes is coming.
III.Tinder. When the product is the pitch.
Tinder's original deck was unusually visual for its era — screenshots of the product occupy most of every slide. The interface itself does the value-proposition work. Reading the deck without context, you understand within thirty seconds what Tinder does and why it would feel different from any predecessor.
IV.Brex. What to do when traction can carry the deck.
The Series B deck is a master class in numbers. Almost every slide contains a chart, and the growth curves are unambiguous. The team slide name-drops Stanford and a Y Combinator alumni network without apology. This is what a deck looks like when the company's recent history is itself the most persuasive argument.
V.Notion. Taste as a commercial asset.
The deck resembled a Notion page, which was, of course, the point. The artifact itself functions as a demonstration of the product's design philosophy. An investor evaluating Notion is not only buying a thesis about the market but a particular sensibility about how software should feel — and the deck makes that sensibility legible.
VI.LinkedIn. Naming the failure of your predecessor.
Reid Hoffman's now-public Series B deck is famously long and dense. Its most useful feature is the explicit attention paid to why LinkedIn would not become another Friendster — the social-network failure that was, at the time, the most recent counterexample in investors' minds. Hoffman's later annotations are, in places, more instructive than the original slides.
VII.Facebook. One number, well chosen.
The Accel pitch was carried by a single statistic: student users who signed up returned to the platform at extraordinary rates. The remainder of the deck follows from that one number. Once the reader accepts the retention figure as real, every subsequent slide becomes more credible by association.
VIII.YouTube. The case for unembellished prose.
The deck identifies a problem that is now difficult to remember: in MMV, there was no straightforward way to share videos online. The slides do not reach for hyperbole. They name a clear gap, propose an obvious solution, and trust the reader to grasp the implications.
IX.Snap. The deck as manifesto.
Snap's deck reads less as a pitch and more as a philosophical document. It opens with a case for ephemeral content as a more honest mode of communication than the permanent record social networks had become. The product, when it arrives, feels like the inevitable consequence of the worldview the deck has already established.
X.Mixpanel. Repetition as proof.
The deck is wall-to-wall growth charts. Every claim is accompanied by a chart, and every chart inclines upward. The deck spends comparatively little time on what Mixpanel does, and a great deal of time on how rapidly Mixpanel is growing. The repetition itself functions as proof.
XI.Eleven habits common to the entire survey.
The length is short, and not by accident.
Eighteen of the twenty-three decks studied here run between ten and fifteen slides. The outliers — LinkedIn at thirty-nine, Uber at twenty-five — had specific reasons rooted in their context. If you are a typical seed or pre-seed founder operating without those constraints, ten to twelve slides is the appropriate length, and shorter is generally preferable to longer.
The problem precedes the solution, without exception.
Every deck in the survey establishes the pain before introducing the product. The instinct to lead with the thing you have built is consistent and incorrect. Investors fund problems; the product is the response.
There is one number that anchors the document.
The best decks in the survey have a hero metric around which the entire pitch is organized. For Facebook it was retention; for Mixpanel, growth; for Airbnb, unit economics. Identify yours before you begin writing.
The team slide is positioned as a closer, not an opener.
Placing the team upfront reads as credential-dropping. Placing it at slide eight or nine — after the product, the market, and the traction — invites the reader to think I would back these people to do this. The order matters.
The ask is specific.
Every deck in the survey closes with a clear ask: the amount being raised, the allocation in percentage buckets, and the milestone that capital unlocks. Vague closings — open to discussing valuation, seeking strategic partners — signal weakness.
XII.Five errors that disqualify an otherwise strong deck.
- Buzzword inflation. Phrases such as revolutionary AI-powered platform leveraging cutting-edge machine learning read as evidence the writer is reaching for borrowed authority. Replace each instance with a specific, falsifiable claim.
- The omitted competition slide. Asserting that you have no competition reliably reads as evidence you have not looked. Name two or three plausible alternatives — including doing nothing — and explain your durable advantage.
- Hockey-stick projections without a starting point. A five-year revenue forecast reaching nine figures, without a credible mechanism explaining how the company reaches the first turning point, is taken as fiction.
- Wall-of-text slides. If a slide requires thirty seconds of reading, it has failed as a slide. The medium is visual; treat it as such.
- The missing ask. Always close on the amount, the allocation, the milestone. Not on thank you.
If you have a draft in hand, the fastest way to test these patterns is to run it against the canonical structure laid out in essay three — and to do so with fresh eyes after a day away from the document.
A deck of this calibre, built for you.
Five complete pitch decks across five distinct registers — minimal editorial, dark premium, bold consumer, data-led B2B, narrative founder-led. Pure HTML, fill in the brackets, ship in an hour.
Acquire the pack — $39Or — diagnose the deck you have.
Before commissioning a rebuild, paste your current draft into our complimentary diagnostic. It returns a 0-100 reading against the criteria institutional investors apply on a first read. No signup, no email, no telemetry.