How to build a pitch deck that raises investment: structure, examples and common mistakes

The structure of a 10-12 slide pitch deck (problem, solution, market, product, model, traction, team, competition and ask), what doesn't belong on each slide, and the mistakes that make an investor close the file.

What the deck is for (and what it isn't)

The deck has one job: get someone with little time to understand the business enough to want one more hour. It's not a business plan. It's not your product. It's not the place to "tell your journey".

There are two versions and they aren't the same file:

The one you send. Understandable without you. Each slide stands on its own. Little text, numbers with unit and date, one idea per page.
The one you present. Even fewer words. You are the thread. If the slide already says everything you were going to say, it's extra —or you're reading.

Don't send the presenting one by email. Don't present the sending one by reading paragraphs out loud.

Don't expect "with this deck you raise" either. A clean file doesn't fix a market that doesn't pay or a round that makes no sense. The deck is hygiene. The thesis comes from the business.

The 10–12 slide structure

If you go past 12, it's not that your company is more complex. It's that you didn't cut. The order below is the one an investor already has in their head; inventing another forces them to hunt for the information.

1. Cover

Name, one line of what you do (who + problem + how, in under 12 words), and the round context if public ("pre-seed, 2026"). Logo if it's readable on a phone.

Doesn't belong: a manifesto, a quote, "we're revolutionizing X", or three paragraphs of vision. If the cover needs explaining, the company doesn't have its sentence yet.

Useful example: "Dollar collections for exporting SMBs, without the bank in the middle." Useless example: "The comprehensive financial services platform of the future."

2. Problem

Who suffers, how often, what it costs them today. A concrete character beats "the Latin American market". "A 20-person importer in Guadalajara loses two weeks and a fee every time they pay Shenzhen" is understandable. "Financial inclusion is broken" isn't.

Doesn't belong: a newspaper clipping, a TAM on this slide, or five problems at once. A deck listing "also onboarding, also payroll, also invoicing" hasn't chosen yet.

3. Solution

How that pain disappears, in one sentence and one diagram. What the user does, what you do, what ends up solved. If you can draw the before and after in two columns, do it.

Doesn't belong: the feature list, the architecture, or "we use AI" as if it were the solution. The solution is the outcome for the client. The technology is the means and it lives in Product.

4. Why now (optional)

The external change that makes this unsellable five years ago: a payments rail, an API, a regulation, a cost that dropped. Without this slide, many LATAM decks sound like the same idea from 2017.

Doesn't belong: a timeline of humanity, or "AI changed everything" without the mechanism. If you can't name it in one sentence, delete the slide.

5. Market

This is where most decks in the region die. Don't start with all of Latin America. Start with the beachhead: the group you can sell to within 18 months, through a channel you already know.

One honest way:

Today we sell to X (manufacturing SMBs with 10–50 employees in Mexico already paying for foreign-trade software).
Then Y (the same profile in Colombia and Chile).
The ceiling, if everything works, Z (the region's exporters already operating in dollars).

Name the unit: number of companies × what they pay per year. If you don't have the fine-grained number, say how you approximated it. An investor prefers a small, collectible market to an invented continent.

Doesn't belong: "if we capture 1% of a 40 billion market". That isn't a model. Mixing free users with enterprise budgets doesn't belong either.

6. Product

One or two real screenshots. The happy flow in 3 steps. What's live today versus what's roadmap —and make the difference visible. If pre-product, show the proof: interviews, a letter of intent, a pilot, a waitlist with corporate emails, not a landing page of the curious.

Doesn't belong: ten screenshots, a video that won't open in the PDF, or mockups sold as product. Lying here gets discovered in the first demo.

7. Business model

Who pays, how much, how often, how you collect. In LATAM it isn't a given: do you invoice in dollars or local currency? Who buys (the owner, the accountant, IT)? Self-serve or three months of committee?

One price and an estimated gross margin are worth more than a canvas. If you don't charge yet, state the price and who you validated it with.

Doesn't belong: five revenue lines "for diversification". An early startup with marketplace + SaaS + ads + services + data is an agency with PowerPoint. Pick the line that has to work first.

8. Traction

The most-read slide. A chart of what matters for your stage, with dates and definition. "MRR 18k USD, August 2026, net of discounts, 40 paying clients" can be discussed. "Explosive growth" can't.

If there's no revenue, don't invent a vanity proxy. Show concrete demand: 40 calls with the right profile, 8 who would pay price X, 3 who already used a prototype.

Doesn't belong: vanity (followers, downloads, "reach"), or a chart that only goes up because you changed the definition halfway through.

9. Competition

A simple matrix: you vs. 3–5 real alternatives on the axes the client uses to choose. In LATAM the winning alternative is almost never another startup. It's Excel, WhatsApp, the bank, the accountant, or "do nothing". Put them in. If your only competitor is "traditional solutions", you didn't do the work.

Doesn't belong: "we have no competition". That phrase doesn't sound like blue ocean. It sounds like you didn't talk to the client. A map with 20 logos to prove you read Crunchbase doesn't belong either.

10. Team

Why you can build this: problem domain, ability to build, to sell. A founder who ran foreign trade for 8 years and a CTO who has already shipped product to production beat three MBAs and a famous advisor.

If the team is in LATAM and the client in the United States (or the other way around), explain it: timezone, language, how you sell, who's full-time.

Doesn't belong: 12 advisors who don't answer, the future org chart, or having someone who isn't a founder present. The investor is buying the person who stays when the quarter goes bad.

11. Ask (the round)

How much, in what instrument, for how many months, and where the money goes. An ask of 400k USD that's "used to grow" isn't an ask. One of 400k USD, 16 months of runway, 2 engineers, 1 salesperson and the rest in infrastructure and legal, is.

This slide is also where it shows whether you understand what dilution costs you. Raising more "for margin" dilutes early. Raising less "to hold equity" leaves you without the milestone the next round demands. The number buys a verifiable milestone, not a feeling of safety.

Doesn't belong: a deck without an ask. "We're open to talking" forces the other side to do your job. Neither does an aggressive valuation without traction, or marketing as a large percentage without a channel.

12. Close (optional)

The cover line again, the ask in one line, and how to contact you. If you already covered that, you don't need a "thank you" slide with a landscape.

Pre-seed vs seed: not the same file

The skeleton is the same. The weight changes.

Pre-seed. Buys problem + insight + team, and a signal the world reacts. Traction can be qualitative if it's specific. The product can be ugly if someone already used it. The market is a beachhead, not a continent. The ask covers 12–18 months for a milestone that unlocks seed: first revenues, early retention, or a repeatable channel. Eighteen slides of financials in pre-seed are a costume.

Seed. The story no longer suffices. You need a curve (even a small one), why it repeats, and unit signals: what acquisition costs, what's left after collection, how fast they leave. The product gets demonstrated. The ask scales what already works a little, not "finish the MVP". If your traction slide is a list of plans, you're sending a pre-seed deck with another name.

Rule of thumb: if the first question is "has anyone paid?" and you can't answer with a number, you're not at seed. Don't fake it on the cover.

Mistakes that close the file

Inflated TAM. Latin America isn't your market because the problem "exists in 33 countries". Your market is where you can sell with the channel you have. A huge number without a beachhead signals you haven't sold.

25 slides. Nobody asks for them. Personal history, values, repeated screenshots and three versions of the same chart go to the appendix.

No ask. Ending on "next steps" without amount, instrument or use of funds forces the other side to guess whether you need 80k or 2 million. That doubt doesn't resolve in your favor.

A non-founder presenting. Sending the head of marketing, the advisor, or the friend who "speaks better" is a signal. The one raising is the one who'll be on the board. If your English isn't native, present anyway: clear and slow beats a spokesperson who can't answer product questions.

"We have no competition." There's always a current way to solve the problem, even if it's bad. If you don't name it, the other side assumes the client is fine —or that you didn't make the calls.

The projection that only goes up. A 5-year curve with no assumptions, no churn and no flat quarter doesn't read as ambition: it reads as template. If you project, show the 3–4 assumptions (price, conversion, retention, CAC) and what happens if one fails.

Others that repeat: mockup sold as product, five revenue models, vanity metrics, fashionable valuation without numbers, and copying a US unicorn's deck. That file was written for a different capital density.

How to build it without losing two weeks

One afternoon for the skeleton in 11 titles. Another to fill in with numbers you already have. A third to cut: each slide, one sentence out loud; if you need two, split it or delete it. Then five people of the right profile and one question: what didn't you understand? Rewrite that. Not the logo color.

Format: PDF, 16:9, typography readable on a phone. The file is called Company_preseed_2026.pdf, not deck_final_v27_NEW.pdf.

Outside the PDF, ready if asked: simple cap table, MRR, burn, runway, margin if applicable, and how you get to the next milestone.

Frequently asked questions

How many slides should a pitch deck have?

Between 10 and 12 to send or present. An appendix of 4–6 slides (unit economics, pipeline, product) is fine if it doesn't mix with the main story. More than 15 in the body is, almost always, fear of choosing.

Do I need a financials slide?

In pre-seed, not as a 5-year projection. Runway, burn and use of the ask are enough. In seed, one slide with real history plus assumptions for the next stretch. The full model lives in the spreadsheet you send if they ask.

Can I use the same deck for a LATAM angel and a US fund?

The skeleton, yes. Currency, round size, whether the beachhead is local or pays in USD, and which comparables the other side understands change. An angel in Bogotá doesn't need you to explain SPEI. A Bay Area fund does need to know why your collector isn't Stripe.

Does a Sequoia or YC template work?

The order works. The phrases don't. Copying "we are building the OS for X" doesn't replace a line one of your clients would sign. Use the structure, ditch the copy.

The good deck is understood in three minutes, defended for an hour, and doesn't promise what the quarter can't pay. Keep it short, put the ask in, say who really competes with you, and let the business —not the template— do the rest.