How to prepare a pitch deck for investors: the deck slide by slide

We do pitch deck analysis on early-stage projects, and the pattern repeats: most decks are assembled as a story about the company. First about us, then about the product, metrics somewhere in the middle, the amount being raised at the end in small type. An investor reading the same file for the first time is running a different sequence in their head — who are you, what is the problem, how do you solve it, who pays for that, what is already proven, how much do you need and what for. A deck built along that second sequence reads as a set of answers rather than as a presentation. That difference is the whole of what people mean when they say a deck is "well structured".

How many slides, and how a deck actually gets read

A working benchmark for an early-stage project is 12–15 slides in the main body. That is not a rule handed down from anywhere; it follows from how the file is consumed. A first read takes two or three minutes, and the entire story has to fit inside that window. Everything that does not fit is not deleted — it moves into the appendix after the last slide, where it waits for a reader who has already decided to keep going.

The same constraint applies inside a single slide. The reader looks at it for about twenty seconds and spends half of that on the headline. So write headlines as statements, not as section labels. "Three pilots renewed for a second term" delivers its content even to someone scrolling fast. "Traction" forces the reader to squint at the numbers underneath and work out the meaning on their own — and on a first pass, nobody does that work for you.

That gives you a test for the whole structure. Read only the slide headlines, in order, one after another. If they form a coherent narrative, the deck holds together. If they read as a table of contents, the headlines need rewriting into statements. This is the single cheapest edit in the whole process, and in our reviews it is the one that changes the most.

One more thing worth saying early: the deck is not a legal document and not a business plan. It does not have to be exhaustive. It has to be checkable. Every claim in it should be one the founder can defend in a sentence, and every number should be one the founder can reproduce on a calculator in front of a stranger.

Pitch deck structure: twelve slides

Slide 1. Title

The title slide needs the name, one line describing what you do and for whom, the stage of the project, and a contact. The one-liner is the main element on the slide: it has to be understandable to someone who does not know your industry. The formula that usually works is "we help [whom] [do what] [by means of what]" — with no superlatives and no internal module names that mean nothing outside your team.

The twenty-second idea: "I understand what this company does, I can keep reading." A title slide does not sell the project; it removes the need to guess. A version date is useful too. A deck lives for months and circulates by forwarding, so a reader needs to know how fresh the numbers inside are before they start quoting them to a colleague.

The slide works if: someone outside your industry can look at it once and describe your business back to you in their own words, with no terms in their description that need decoding.

Slide 2. Problem

Describe the problem in your customer's language, not as a market description written from inside your product. A working problem slide has four elements: the role ("head of sales in a company under 200 people"), the situation, the frequency ("every Monday", "on every deal"), and what the person does about it today — a spreadsheet, an agency, an extra hire, or nothing at all.

Anything from the field adds credibility: how many people you interviewed, how many called the situation regular, one short verbatim quote. Our own demo deck, StockPulse, carries a line of this kind — 24 production sites interviewed, 19 confirmed downtime caused by stock tracking at least once a month — and that single pair of numbers does more work than a paragraph about market fragmentation ever will.

The twenty-second idea: "this pain exists for specific people, and it repeats."

The slide works if: a specific role and a frequency are named on the slide, and at least one fact comes from talking to customers — a count of interviews, a share who confirmed, or a quote.

Slide 3. Solution

This slide answers exactly the problem stated on the slide above, ideally in the same words. The common mismatch looks like this: the problem is written as "two days to close the monthly report" and the solution as "a unified data management platform". Formally the second is about the first, but the reader has to build the bridge themselves — and on a fast read they do not build it.

Product features do not belong here; they have their own slide next. What you need is a sentence about what happens to the customer's job, and a line about how: the approach, the data, an integration, the automation of one specific step. If a startup pitch deck feedback session ever produced a single most-repeated note, it is this one — the solution slide drifts one level of abstraction above the problem slide.

The twenty-second idea: "I can see exactly how the pain on the previous slide goes away."

The slide works if: the wording of the solution literally reuses the key words from the problem slide, and the difference between "before" and "after" is expressed in something measurable — time, money, or a number of actions.

Slide 4. Product and demo

Two or three screenshots of the real interface are more convincing than any architecture diagram. Do not show the whole menu; show a short user path — where the person starts, what they do, what result they get. If the product is not visual, describe the same path in three steps and say what happens at each one.

State the readiness honestly: what works today, what is in development, what is planned. A "planned" tag next to a feature reads far better than the investor discovering the same fact in a meeting. And check that any demo link opens from a clean browser without an access request — a link that asks a stranger to request permission is a dead link.

The twenty-second idea: "the product exists, and I can see it being used."

The slide works if: the slide alone makes it unambiguous what runs today and what is marked as plan, and the user path fits into three steps.

Slide 5. Market

Three circles with a number lifted from an industry report give context, but they do not answer the question the reader actually cares about: where will your next customers come from, and how much money is reachable along that path. So next to the top-down estimate you need a bottom-up calculation — how many companies match your customer profile, how many of them you can physically reach in a year, what the average contract is, and what that multiplies out to.

The second mandatory element is the beachhead segment and one sentence on why that one: shorter sales cycle, you already have access, the pain is sharper there. At an early stage a narrow entry point reads as maturity, not as a shortage of ambition. Investors are not surprised by a small first segment; they are surprised by a founder who cannot name one.

The twenty-second idea: "this team knows who it sells to first and what that is worth in money."

The slide works if: alongside the overall market estimate there is a count of customers reachable by you and an average contract value, their product matches the figure you print, and the beachhead segment is named explicitly.

Slide 6. Business model

The model decomposes into four questions: who pays, for what, how much, and how often. If the payer and the user are different people, say so directly — the whole sales cycle depends on it. If you have several plans, put the main one on the slide and move the rest to the appendix.

One line about the sales channel earns its place: how does a customer reach you today. Even "through the founder's personal network" is clear and appropriate at this stage, whereas leaving it out simply parks the question until the meeting. Investors do not penalise an early channel; they penalise not knowing what the channel is.

The twenty-second idea: "I understand where the money comes from and how regularly."

The slide works if: the model reads as a single line of the form "who pays — for what — how much — how often", and that line needs no spoken explanation.

Slide 7. Traction

In this section the reader is looking for three things: money, repeat usage, and signed commitments. Sign-ups, downloads and followers honestly show that the team is active, but they answer a different question. If there is no revenue yet, that is a normal state for an early-stage project, and the plain sentence "no paying customers yet, three pilots are free" builds more trust than a large sign-up number standing in its place.

Format matters as much as content. Month-by-month values are more informative than cumulative charts: a curve that by definition only goes up tells you nothing about momentum. Every number needs a period and a source. A metric without a time window is not a metric, it is a decoration.

The twenty-second idea: "something here is confirmed by facts rather than by plans."

The slide works if: every number has its period written next to it, and at least one of them reflects money coming in, a renewal or a user coming back, rather than a one-off action.

Slide 8. Unit economics

For an early-stage project four quantities are enough: what it costs to acquire a customer, what that customer brings in, the gross margin, and how long the investment takes to come back. Anything computed on assumptions should be labelled "estimate", with a line on where the assumption comes from. A labelled hypothesis is material for a discussion; an unlabelled one becomes the subject of a cross-examination.

Arithmetic deserves separate attention here. This slide gets edited many times, the price and the conversion rate move, and the derived totals sometimes stay behind from an earlier version. A reader recomputes two or three formulas in their head, and any mismatch they find transfers distrust to every other number in the file. StockPulse, our demo deck, does exactly this: the deck states ARPA per year of 206,000 ₽ while 18,000 × 12 = 216,000 ₽, so two different numbers live inside one file. More on this in pitch deck mistakes.

The twenty-second idea: "this team can do arithmetic and knows what its customer costs."

The slide works if: every derived number can be recomputed on a calculator from the ones next to it, the table has a period under it, and every assumption is called an assumption.

Slide 9. Competitors

The customer is already solving this problem today — with a spreadsheet, a contractor, an adjacent tool, or by deciding to change nothing at all. All of those are alternatives, and the familiar manual method is usually the most formidable one on the list. A row called "how this gets solved today without us" makes the slide more convincing than any roster of company names.

A comparison table works when the alternatives have strengths marked too. A grid with ticks only in your own column reads as criteria chosen to produce the answer. State your own advantage narrowly and checkably — one or two parameters that matter to the segment you picked — rather than as a general claim of being better.

The twenty-second idea: "this team knows the field it plays on and understands why someone would pick them."

The slide works if: the comparison includes the current way of solving the problem without your product, every alternative has at least one strength marked, and your advantage is expressed as one verifiable parameter.

Slide 10. Team

This slide answers one question: why will these specific people build this specific product. Job titles do not answer it. Biography facts connected to the task do: worked in this industry and knows the procurement cycle, has shipped a similar product before, ran engineering in a team of this size, has access to the first customers.

Show the open roles directly. "Sales is run by the founder; hiring a head of sales is planned after the round" reads as a managed situation. Advisors belong on the slide only when it is clear what they actually help with and how often they are involved — a logo wall of names nobody talks to weekly subtracts credibility instead of adding it.

The twenty-second idea: "this team has a reason to be taking on this problem."

The slide works if: next to every name you can append a sentence containing the word "already" — already built, already sold, already worked in this industry — and open roles are named together with a plan for closing them.

Slide 11. Roadmap

A roadmap is a report plus a plan, not only a plan. Milestones whose dates have already passed need an explicit status: done, moved, cancelled. A move with one line of explanation looks entirely normal; a past date with no status forces the reader to draw their own conclusion, and the conclusion they draw is rarely the generous one.

It helps to cap the horizon at the length of the round: what you will achieve with the money you are asking for. Milestones for the next twelve months expressed as metrics read as a working document; three-year plans read as scenery. And cross-check the dates on this slide against the dates on the traction slide — contradictions inside one file are common, and every investor pitch deck review we run looks for them.

The twenty-second idea: "this team manages timelines and knows what comes next."

The slide works if: every date earlier than today carries a status, and the milestones for the coming year are expressed as metrics rather than as names of workstreams.

Slide 12. The round and the next step

A well-grounded ask ties three things together: the amount, the period it covers, and the metric the project will reach in that time. The test is a single sentence: "we are raising X for Y months to reach Z." If Z cannot be stated as a number, this section is not finished yet — and it is the most frequent fix we suggest in reviews.

Break down the use of funds in large buckets — team, product, sales — but tie the buckets to milestones rather than leaving them as bare percentages. Then close with one next step instead of three. A short conversation, access to the demo, or a folder of materials: pick one, and put a name and a way to reach that person beside it.

The twenty-second idea: "I understand how much is needed, what for, and what to do next."

The slide works if: the sentence "we are raising X for Y months to reach Z" can be written out in full with a number in place of Z, and the slide asks the reader to take exactly one action.

What belongs in the appendix, not in the main deck

The appendix is not a dumping ground for leftovers. It is the set of answers to questions that only arise after interest exists. The main body answers "is this worth looking into"; the appendix answers "how exactly does this work". That split is what lets the deck stay short without losing anything.

What usually moves to the appendix: the detailed financial model and its scenarios; the market sizing methodology with sources; cohorts and detailed usage analytics; technical architecture and data security questions; the legal structure and cap table; full team biographies and the hiring plan; descriptions of pilots and the terms agreed; additional product screens.

The selection rule is simple. If a slide answers a question the reader will only ask after the project has already interested them, it is an appendix slide. The appendix itself needs nothing more than one line on the round slide saying it exists.

Three versions of one deck: email, meeting, one-pager

The same material behaves differently depending on whether a person is standing next to it explaining. The three versions come from one common base and differ in density, not in content — building them as three separate stories is how numbers start to disagree with each other.

The email version has to read without you: denser text, headlines as statements, a period next to every number, terms defined on first use. The format is PDF — the layout does not shift, and the file opens on a phone. This version is what most people mean when they say "the deck".

The meeting version is built the opposite way: less text, larger numbers, one idea per screen. Here a slide is a support for the story rather than a standalone document. Note that in a meeting a deck is almost never walked through in order, so each slide still has to make sense on its own when someone jumps to it.

The one-pager is a single page for a first touch and for forwarding inside a fund. It carries: what you do and for whom, the customer's problem, stage and traction in two or three numbers, the model, the team in one line, the ask and a contact. It is not a shortened deck but a separate text, and its job is to earn a conversation rather than to tell the whole story.

What to check before you send it

A short list that takes about half an hour and removes most of the clarifying questions from a first meeting.

• All slide headlines, read one after another, form a coherent narrative.
• Every derived number has been recomputed on a calculator, and every metric has its period written down.
• Every date earlier than today has a status, and numbers do not contradict each other across slides.
• Abbreviations and internal names are defined the first time they appear.
• Links to the demo and to materials open from a private browser window with no access request.
• The file goes out as PDF, weighs a sensible amount, and is readable on a phone.
• The filename contains the project name and a date, not "deck_final_v7".
• The last slide contains one action, a name, and a way to get in touch.
• Someone outside your industry has read the deck and described the point back to you unprompted.

It is also worth running the questions that usually come up in a first meeting against your own file — they are collected in investor questions for startups. If each one has an answer sitting in the deck or in the appendix, the deck is ready to send.