Most pitch decks are written for the wrong reader. Founders build them as if an investor will sit down, work through forty slides in order, and reach a considered judgment at the end. What actually happens is that someone opens the file between two meetings, skims it in a few minutes, and forms an opinion long before the appendix.
That gap explains most bad decks. They are structured like a report when they need to work like an argument that survives skimming — where the point of each slide is visible in a second and the story holds together even when half of it is read out of order.
This is a guide to writing a deck for how it is genuinely consumed: the sequence that works, what belongs on each slide, the sections founders consistently get wrong, and the design decisions that make the difference without requiring a designer.
How a deck is actually read
Three facts about the reading process should shape everything you build.
First, the first pass is fast. Someone forms a rough view from the title, the problem, the traction, and the team, then decides whether to slow down. That means your best material cannot be buried in the second half.
Second, it gets forwarded. A deck that requires you to narrate it will be misread by everyone you are not in the room with, and the people who matter most in an investment decision are often the ones you never meet. Write it to stand alone.
Third, the decision is not made in the deck. The deck’s only job is to earn a meeting. Trying to close an investment in a document leads to bloat, and bloat is what makes decks unreadable.
A deck does not have to answer every question, only enough of them that someone wants to ask the rest in person.
The sequence that works
There is a conventional order, and the reason to follow it is not tradition. It is that experienced readers know where to look for each piece, and a deck that reorders things arbitrarily makes them hunt. Save your originality for the content.
- Title. Company name and one line stating what you do, in plain language.
- Problem. Who suffers, how often, what it costs them.
- Solution. What you built and how it removes that cost.
- Product. What it actually looks like. Screens, not descriptions.
- Traction. Evidence that the market is responding.
- Market. Who the buyers are and how many of them exist.
- Business model. Who pays, how much, how often.
- Competition. The alternatives and why you win against them.
- Team. Why these specific people.
- The ask. How much, for what, to reach which milestone.
Ten to fifteen slides. Anything beyond that goes in an appendix after the ask, where interested readers will find it and uninterested ones will not be slowed by it.
One structural exception is worth making: if you have strong traction, move it earlier — straight after the problem. Traction is the most persuasive thing in any deck, and holding it until slide nine assumes a patience your reader does not have.
The slide most founders waste
The title slide usually contains a logo, a company name, and a tagline that means nothing to anyone outside the company. It is the first thing seen and the least used.
Put a single sentence on it that a stranger could repeat accurately: what you do, for whom, and what changes as a result. “Scheduling software for home care agencies that cuts unfilled shifts” beats “Reimagining the future of care operations” every time.
The test is whether someone who reads only that slide could describe your company to a colleague. If they cannot, the rest of the deck starts from confusion, and confusion is very hard to recover from in a document nobody is reading carefully.
Writing the slides that get skipped
Problem
The failure here is abstraction. “Businesses waste time on manual processes” describes nothing and could introduce a thousand companies. Name the specific person, the specific moment, and the specific cost.
Better problem slides read like a scene: an operations manager at a mid-size distributor spends the first two hours of every Monday reconciling three systems by hand, and errors from that process cost the company real money in wrong shipments. A reader who can picture the moment can evaluate your solution.
Traction
Show the shape of a curve, not a single number. Growth over time is informative; a cumulative total is a number that can only go up and therefore says little.
Be honest about scale. Small numbers presented plainly are far more credible than small numbers dressed up with percentages that hide the base. Sophisticated readers can tell, and the attempt to obscure costs you more than the number would have.
If you have no revenue yet, show engagement, retention, pilot commitments, or waitlist conversion. If you have nothing, say what you have learned instead. A pre-traction deck that is straightforward about being pre-traction reads better than one padded with vanity metrics.
Market size
The market slide is where credibility is most often lost. Enormous top-down numbers derived from an industry report and a percentage assumption are treated as noise, because everyone has seen the same slide a hundred times.
Build it from the bottom instead: the number of potential customers who match your profile, multiplied by a realistic annual contract value. The result is smaller and infinitely more persuasive, because it demonstrates you understand who your buyer is. Do not invent the inputs — if you cannot source a number honestly, describe the segment qualitatively and say how you would size it.
Competition
Two versions fail. The first claims no competitors, which reads as either naivety or a market with no demand. The second is a feature grid where you have every checkmark and everyone else has few, which nobody believes.
The credible version names the real alternatives, including the manual process or spreadsheet that most customers actually use today, and states plainly what each one is good at. Then it explains the specific wedge where you are clearly better and why that wedge matters to your buyer. Acknowledging a competitor’s genuine strength makes your claim about your own strength more believable, not less.
Team
Logos of former employers are the weakest possible version of this slide. What matters is the connection between who you are and why this problem is solvable by you specifically.
Write one line per founder tying background to the problem: eight years running operations at the exact kind of company you sell to, or the person who built the system this product replaces. If a founder’s background does not connect, say what they own instead. And name the gaps — a deck that identifies the role it needs to hire next reads as self-aware rather than incomplete.
The ask
Be specific and be quiet about the rest. How much you are raising, roughly how it splits across hiring, product and go-to-market, and what milestone the money reaches.
The milestone is the part founders leave out and readers care about most, because it answers the real question: what will be true when this money runs out that would justify the next round. “Eighteen months of runway to reach a defined revenue level and a repeatable sales motion” is an answer. “To accelerate growth” is not.
Design rules that do not require a designer
- One idea per slide. If a slide has two arguments, it is two slides. Density is not thoroughness.
- Headlines that state the point. Write “Retention has held above launch levels for six months,” not “Retention.” The headline should carry the message even if nothing else is read.
- Charts over tables. A reader absorbs a shape instantly and a table not at all. If a number is important, make it the largest thing on the slide.
- Real product screens. An actual screenshot, even of a rough interface, is more convincing than a diagram of boxes and arrows.
- Text large enough to read on a phone. Many first reads happen on a small screen. If you shrink the font to fit more words, cut words instead.
- Consistency over decoration. Same fonts, same colors, aligned elements. Restraint reads as competence; visual noise reads as a company that has not decided what matters.
None of this requires design skill. It requires deciding what each slide is for and deleting everything that does not serve it, which is a writing problem rather than a visual one.
The failure patterns worth checking for
Explaining the market for four slides. Founders in unfamiliar industries over-explain context and run out of attention before reaching their own company. Two sentences of context, then get to what you do.
The deck that needs narration. If a slide is meaningless without you talking over it, it will fail in every forwarded read. Test it by sending the deck to someone unfamiliar with the business and asking what they understood.
Hedging everything. Decks written to be unfalsifiable end up saying nothing. Make specific claims. A specific claim that a reader disagrees with starts a conversation; a vague one ends it.
Hiding the weakness. Every early company has an obvious problem — no revenue, a missing co-founder, a well-funded competitor, a churn number that is not good yet. Readers will find it. Naming it yourself, with what you are doing about it, converts a discovered weakness into evidence of clear thinking.
Appendix as a dumping ground. An appendix is for answering predictable follow-up questions in depth. It is not a place to move slides you could not bring yourself to delete.
Test the narrative before you polish it
Before touching layout, write the deck as ten sentences — one per slide — in a plain document. Read them in sequence and check whether they form an argument that someone could disagree with at a specific point.
A working narrative sounds roughly like this: a defined group has an expensive problem, they currently solve it badly, we built something that solves it well, these people are using it and continuing to use it, there are many more like them, they pay us this way, here is why we beat the alternatives, we are the right people to do this, and this amount of money gets us to that milestone.
If the sentences do not connect — if slide five does not support slide four — no amount of design will fix it. Most decks that feel wrong are narrative problems being treated as visual ones, and the ten-sentence version exposes that in twenty minutes.
Frequently Asked Questions
How long should a pitch deck be?
Ten to fifteen slides for the main deck, with an appendix for detail. The constraint is useful: forcing yourself into that range makes you decide what your actual argument is. If a section will not compress, it is usually because you have not yet decided what it is claiming.
Should I send the deck before the meeting?
Generally yes. Investors usually expect to read something in advance, and a deck that only works when narrated is a weakness you want to discover early anyway. Keep the version you present sparser than the version you send, since a slide that reads well alone is often too dense to speak over.
What if I have no traction to show?
Show whatever evidence you have — customer conversations and what you learned, pilot commitments, a working prototype, waitlist behavior, or relevant depth in the problem. Then be explicit that you are pre-traction and state what the raise is meant to prove. Pretending otherwise is the more damaging option, because the truth surfaces in the first serious conversation.
Write it for the reader who is not in the room
The most useful reframe is that your deck’s real audience is rarely the person you sent it to. It is whoever they forward it to, read in a spare ten minutes, without context and without you there to explain.
Write for that person and everything else improves. Headlines carry the argument because they have to. Slides make one point because a distracted reader cannot follow two. The weakness gets named because a stranger will find it anyway. The story is compressed to ten sentences because that is all the attention available.
Decks that succeed are rarely the most beautiful. They are the ones that survive being skimmed by a stranger and still leave a clear, specific, arguable idea behind — which is the only thing a document can realistically achieve, and enough to get you into the room where the actual decision happens.
