पाठशाला Pathshala · धन Dhan, Money · Lesson 07 · Start

The pitch deck: twelve slides and the one that matters

A seed deck is read in under two minutes by someone deciding whether to spend an hour. Build it on what has happened and what you know that others do not, and put the number where it is found.

Pathshala, The Founder Library · 11 October 2026 · 9 min read

A seed deck is not read. It is scanned, by a person with forty others in the same folder, for about two minutes, to decide whether to spend an hour. Everything about how the deck should be built follows from that sentence, and most decks are built as if it were false: as a presentation to be delivered, forty slides long, with the number that matters on slide nineteen.

This lesson gives the twelve slides in order and what each one has to do, the reading-time data that explains the order, the slide a partner turns to first and how to build it so that it holds them, the design rules that make a deck legible to a tired reader, the Indian specifics that belong in a deck raised here, and the quarterly habit of rebuilding it from the numbers up.

What a deck is for

The deck has one job: to get the meeting, and then to help the meeting go well. It does not raise the money. Y Combinator’s guide to designing a better pitch deck puts the point in its first line: investors invest in teams not slides, and the slides exist to make your ideas more clear. Sequoia’s long-standing outline says the same thing from the other side of the table, opening with the instruction to define the company in a single declarative sentence, and the reason is that a partner who cannot say in one sentence what you do cannot repeat it to the partner next door, and a company that cannot be repeated does not get to the Monday meeting.

The consequence is a deck that is short, numerical and ordered for a reader rather than a listener. Sequoia’s outline has ten headings: company purpose, problem, solution, why now, market potential, competition and alternatives, business model, team, financials, vision. The twelve below keep that spine and add two slides a seed deck needs that a general outline does not: a traction slide, because at seed the evidence is the argument, and a go-to-market slide, because the question a seed investor asks after “does it work” is “how will people find out”.

The twelve slides, in order

One: purpose. The company in one declarative sentence, with the name and a logo and nothing else. If the sentence needs a comma it is two sentences; cut one. Two: problem. Whose problem, how often it occurs, what it costs them today and how they cope now. A number on this slide beats an adjective: “a Surat textile trader reconciles 400 invoices a month by hand” is a problem; “payments are broken” is a mood. Three: insight. Not the product yet. The thing you know about the problem that the obvious solutions miss, usually learned by doing the work or by the [customer interviews](/library/the-customer-interview-done-properly). This is the slide investors remember when they describe you to someone else. Four: why now. What changed in the last two or three years that makes this possible or necessary: a rule, a rail, a price, a behaviour. In India this slide often writes itself from the public stack, and DocSend’s 2023 data found investors giving it the third-longest viewing time of any section. Five: product. Show it. Two screenshots and a caption that says what the user did, or a thirty-second video linked from the slide. Not a feature list. Six: traction. The one that matters, covered below.

Seven: market. Sized from the bottom up, in counts and prices that reconcile to the traction slide, as the [market sizing lesson](/library/market-sizing-an-investor-will-believe) describes. A ₹2 lakh crore figure from a consultancy report is a reason to stop reading. Eight: business model. Who pays, how much, how often, what it costs to serve them and what is left. Three numbers: price, gross margin and payback. Nine: competition. Every alternative the customer actually uses today, including the spreadsheet and the cousin, and the specific reason you win against each. The two-by-two with you alone in the top right persuades nobody; the honest table does. DocSend’s 2023 data found time spent on competition sections up 88 per cent in a year, which is investors telling you they have stopped believing the two-by-two. Ten: go-to-market. The one channel that has worked so far, with its cost and conversion, and the next one you will test. Eleven: team. Why these people, for this problem, in two lines each, with the thing each has done that is relevant rather than impressive. Twelve: the ask. The amount, the milestone it reaches, when, and the three things the money buys. A seed fund reads this slide to check that the round is sized to a milestone they would pay for next.

Everything else, including the financial model, the detailed cohort tables and the cap table, goes in an appendix that is sent on request. The deck that gets the meeting and the pack that survives diligence are two documents, and confusing them produces the forty-slide deck that is read for ninety seconds.

Why this order: what the reading data says

DocSend has tracked how investors read decks since 2015, when its study with Harvard Business School, reported by TechCrunch, found an average of three minutes forty-four seconds per deck, with financials, team and competition the most-studied pages, and a successful order close to Sequoia’s with product and team moved earlier. By 2019 the average had fallen to three minutes twenty-seven seconds while successful seed decks stayed at about twenty slides. By 2023 it was under two minutes for the first time, at one minute fifty-six seconds, with time on traction up a third and time on competition up 88 per cent compared with the year before.

Read together the data says three things. Investors spend their minutes on the pages with numbers, and at seed the financials page is the traction page, because there are no financials worth the name. They have become more sceptical about competition and want the honest table. And the whole read is now shorter than a song, which means the deck must deliver its argument in the first six slides and use the last six as proof. That is the logic of the strip below.

The slide that matters, and how to build it

A partner opening a seed deck is looking for one thing: evidence that something is already happening that would be worth more if it happened faster. That is the traction slide, and the way a partner reads a deck is to find it, judge it, and then read the rest in the light of it or not at all. Build it accordingly. One number, the one that most honestly measures whether the company is working: monthly revenue for a company that charges, weekly active users for one that does not yet, gross merchandise value only if the take rate is on the same slide. Its growth, as a line over at least six months, with the rate written on the chart; the [growth lesson](/library/growth-rate-is-the-only-number-that-matters-early) explains why the rate is the number investors chase. And the period, so that nobody has to ask whether the chart is weeks or quarters.

Then the caption. YC’s design guide makes the point with a growth graph: without the caption a reader has to study the chart to reach the conclusion, so state the conclusion. “₹4.2 lakh monthly revenue, growing 18 per cent a month since March, 61 paying customers, 94 per cent retained” is a caption. “Strong early traction” is not. If retention is good, the second chart on the slide is a cohort curve, because a partner who sees revenue growing and customers staying has already decided to take the meeting. If the company is too early for any of this, the traction slide is still the sixth slide, and it carries the strongest evidence you have: a waitlist that behaves like demand, letters of intent with named buyers, a pilot with a measured result. The one thing it may not be is absent.

A partner does not read a deck from the front. They find the number, decide what they think of it, and then read the other eleven slides to confirm or overturn the decision.

Design for a tired reader

YC’s three rules are legible, simple and obvious. Legible means large type, high contrast and the text near the top of the slide, and for a seed deck the reader to design for is a partner on a phone between meetings, not an audience in a hall. Simple means one idea a slide, and if the idea needs two slides the guide still asks you to try to make it one. Obvious means the slide states its conclusion as the headline so that a reader who sees only the headline gets the point. Kevin Hale’s Demo Day advice is five to seven slides for five to seven ideas; a seed deck sent cold can run to twelve because it has to stand without a speaker, and should go no further.

Three practices from the fundraising lesson belong here too. Send the deck after the meeting rather than before, so that it reminds rather than replaces. Keep one version, dated, so that a partner who shares it internally shares the current numbers. And make every figure in the deck reconcile to a metrics sheet in the data room, because the first thing an associate does with a deck they like is check it against the numbers, and a deck that does not tie out is a company that does not know its own figures.

What an Indian seed deck adds

Four things. Rupees in lakh and crore with the dollar figure once on the ask slide for a fund with foreign partners; a deck that is entirely in dollars reads as a company that has not met its customers. The regulatory slide, where the business touches one: a lending product shows its licence route and the capital it requires, a health product its registrations, a payments product the rails it rides and whose permission it rides on; a seed investor here has been burned by the regulator before and wants to see that you have not been surprised. The distribution reality: if the customer is a small business in a tier-two city the go-to-market slide says how it is reached in that city, not in a model of it. And DPIIT recognition, which is a line on the ask slide rather than a slide of its own, because it tells the fund the convertible note route is available and that the Fund of Funds money some of them manage can be deployed.

A quarterly rebuild

A deck is a snapshot of the company’s argument, and arguments age. On the first working day of each quarter, whether or not a raise is planned, open the metrics sheet and rebuild the traction slide from it. Then ask of each of the other eleven: is this still true, is this still the strongest version, and does the number on it reconcile to slide six? Rewrite the purpose sentence last, because the company may have become something slightly different while nobody was looking. A founder who does this four times a year is never more than a fortnight from a raise, and the [six-week process](/library/fundraising-process-six-weeks-not-six-months) begins with a deck that already exists. The one who rebuilds it the week the money runs low is writing the deck under the worst possible conditions, which is how forty-slide decks happen.

Nothing here is investment advice. The reading-time figures are DocSend’s, from the years stated; the slide outline is Sequoia’s with two additions, and the design rules are Y Combinator’s. The sources are below.