How to create a winning pitch deck

The title of this post makes it sound like it has a huge potential audience. But its contents are actually pretty niche.

Or, how to buy a winning ax, which is mostly unrelated to making pitch decks, but is probably more fun.

Almost ten years ago, we made this slide for one of our fundraising decks:

I thought that this slide said that Mode, which we primarily designed for analysts and other ambiguous "data professionals,” was a product built for a huge, $16 billion market. I thought it said that even that figure understated our eventual potential, because, once we entered our more sinister second phase—selling to the enterprise, obviously—we would be a company with $60 billion of economic opportunity. I thought that this slide told investors that we had done our research: We weren’t quixotic dreamers who had become enamored with our own toys and were tilting at technological windmills; we were shrewd capitalists who had been inspired by a hopeful vision for a better future, and by the tables in the appendix of the International Data Corporation’s Worldwide Semiannual Big Data and Analytics Software Tracker®. I thought it said that we had read the canon, and knew that markets were the most important determinant of a startup’s success. Though I knew the numbers were rough, I thought that this slide said that we were a serious business with serious ambitions to sell to serious customers.

In hindsight, I suspect that it said nothing of the sort. And what it said instead—which was a lot—wasn’t particularly good:

The slide said that we don't know who our product was for, or who we wanted to sell to. “Analysts and data scientists” is a notoriously vague set of buyers, and “knowledge workers” is even worse. If that was the best definition we could come up with for our market, it said that we hadn’t done the work to understand who might buy what we were building—or worse still, we didn’t even know who wanted to build something for. It said we were planning on casting a uselessly wide and prohibitively expensive net to find customers, that we’d probably end up making a product for a bunch of loosely related but ultimately disjointed buyers, and that our product would eventually get drawn and quartered by a user base that would pull us in a dozen different directions. 

The slide said that we didn’t know how startups actually work. I thought that it pitched Mode as a clever two-step plan: Analysts, and then everyone else. A wedge, and then the rest of the ax head. But not only were we misplaying that move—“analysts and data scientists” isn’t a sharp blade; it’s a blunt hammer—those sorts of plans never really work anyway. Step two is a delusion and a distraction.

The slide said that we were naive, and didn’t understand how hard it is to build enduring businesses. It takes years, if not decades, to make a meaningful dent in a market that big.¹ By the time we were pushing on the boundaries of our $16 billion circle, that market—to say nothing of our early-stage startup and our untested product—would likely have evolved into something else. We were a middle schooler trying to convince our AAU basketball coach to put us on their team by showing them a game plan for defending the Boston Celtics: We were skipping a whole lot of steps. 

The slide said that we weren’t all that ambitious. On the surface, it seems to—$16 billion! $60 billion! Big numbers!² We’re already making plans to play against the Celtics! But what it really says is that we wanted to build a revolutionary product without revolutionizing the market. The most ambitious companies are those that redefine old market boundaries. If we wanted to tell a story about our ambition, showing the size of the current market is just the first chapter. We should’ve also talked about how we expected the market to change, and how we would nudge it along that evolutionary path. That’s the space that the most ambitious companies fill: They don’t have to be category creators in a traditional sense, but they often expand or blur the lines around existing paradigms. Notion, for example, didn’t really create a new category, but blended together parts of Google Docs, to-do lists, internal documentation software, and static website builders. Redshift wasn’t a new category—like Snowflake, it started as a simple database that was uniquely fast and cheap and easy to manage—but made analytical databases accessible to companies that couldn’t afford or maintain older generations of products. In cases like these, market maps and Gartner forecasts are the setup. The punchline is how they’re going to change.³

And most of all, the slide said that we hadn’t thought about any of this for very long. Because it was in our pitch for the same reason it’s in everyone else’s fundraising presentation: We Googled "how to make a pitch deck," found the same Sequoia template that everyone finds,⁴ made a slide called “Market Opportunity” that said something about TAMs and SAMs and SOMs and CAGR, and threw as gaudy of a number on it as we could muster.⁵ 

The initial addressable market

I suppose you could say that that's the point. Pitch decks, one might argue, should follow a numbingly predictable structure—a problem slide with bewildered customers, this market sizing slide, two-by-two grids, “a world class team of builders” with little logos of their previous companies under each headshot—for the same reasons resumes should be full of bot-baiting buzzwords: It helps the automatons who read them—VCs, overworked recruiters, and the 24-year olds who are working as frontline filters for both, Tinder-swiping through resumes and decks while watching YouTube in the background—quickly scan each document in 2.3 minutes.

Which, ok, maybe. Maybe we should write these decks for the pattern-matching AIs that will inevitably read them. Still, if I were to do it all over again, I would’ve done it differently—if not for the venture associates that were skimming our deck, but for ourselves.

When early-stage startups think about their market, they tend to wonder, as we did, if the global market is big enough for the product they plan to build. Are there enough analysts in the world to sell $100 million worth of SQL editors and charts? (No.) Can a company that hosts online conferences be worth a billion dollars? (Yes, but not for long.) These are the markets we research, and the ones we use to convince people of our potential.

But they aren’t the markets that most startups are actually constrained by. Most startups die because the first thing they build—a CRM with one feature, a half-baked forecasting tool for finance teams, a cloud-based collaborative analytics platform for ad-hoc analysis—appeals to some tiny idiosyncratic subset of the billion-dollar Gartner blob. That’s the “addressable market” that really matters—not the one that would conceivably buy a product that will likely never exist, but the one that will buy the runt of an MVP that’s currently on the shelf. 

We should’ve run those numbers. How big could we get if we sold only to SaaS companies in Silicon Valley between 100 and 500 people that had data teams, were comfortable using cloud software, and would likely be reevaluating their BI tools over the next two years? Are there enough of those buyers that we could find some fraction of them fast enough to prove that the more ambitious thing we wanted to create was worth building? 

Often, if you ask these questions honestly, the answers are troubling. They may tell you that you have to capture 50 percent of your initial market to get to a few million dollars in revenue. That’s bad—but good to know! Because that’s how a lot of startups die: They plan to make something that people want, build an incomplete version of it in two years, not many people want that thing, and then they run out of money groping around their huge hypothetical TAM for the handful of people who are willing to buy their barren and buggy prototype. But better to know that—and to make a different plan—than to bank on the fantasy of starting with “data workers” and moving to “all workers.”⁶

Aim small, miss small

One of the other slides that people say you should put in your pitch deck is your path to $100 million in revenue. I always thought this was another dumb bit of busy work. What’s the point of extrapolating our five “design partners” into a financial forecast with 5,000 customers? 

The point, it turns out, wasn’t the projection; it was the plan. It was to force you to explicitly design the type of business you wanted to build. And it was to make you ask what would have to happen for the startup to reach that scale—“we have to match our best week of signups for 31 straight weeks to be on pace to hit our goal”—and to question if those assumptions are remotely plausbile. Any company can hand wave their way to an IPO; it’s much harder to stare down the actual milestones that are required to get you there.

Market sizing is the same exercise, more or less. It can be useless, if you treat it as a mechanical assignment to find a big number in a Forrester report, say “we can sell to five percent of this,” and feel good about your eventual potential. Or it can be helpful, if you use it as a way to figure out—and more importantly, define, for the sake of your own focus—the very narrow market you’ll actually sell to first. 

That’s the only two-step plan that works: Start with some painfully small set of buyers—data teams of a particular type, college students at Ivy League schools, cheap startups who need video conferencing with screen sharing—and expand into the same market but with fewer qualifiers.⁷ And market sizing slides should explain that strategy. 

Admittedly, I have no idea if that’s a pitch that VCs will get excited about. But it’s the pitch they should get excited about, because it’s a real plan. It shows that the company has thought about how to sell a crude beta and not just an aspirational ambition. And it shows that they know who they can actually sell to. Aim for a big market, and miss big by building something for nobody. But aim for a small one, and you might crack it open.

1. SAS, which makes about $3 billion a year, is one of the biggest products in the “analyst and data scientist” market that isn’t owned by Microsoft or Oracle, and it’s 50 years old.

2. Loud noises!

3. To be clear, you don’t have to make all these claims about reinventing markets and creating new categories, and too many companies do. Not every startup needs to be revolutionary; some could just make nicer versions of existing stuff. But not many startups want that to be their narrative, especially when they’re pitching VCs who are addicted to the criminally ambitious founder.

4. A few weeks ago, Nan Yu, Linear’s head of product, gave a banger of a talk about org charts for product and engineering teams. He pointed out that nearly every blog post about the subject references a famous Spotify paper about “tribes, squads, chapters & guilds.” It’s become the single load-bearing column underneath an entire library of online startup scholarship.

This Sequoia deck is the same thing. Because it’s been copied by hundreds of bloggers and corporate content marketers, because recut versions of it are scattered everywhere, and because ChatGPT and Claude now remix it in real time, it’s almost impossible to ask the internet how to make a deck without getting recommended that outline. If you do any research on how to put together a fundraising pitch, you will inevitably be advised to make some derivative of a derivative of a derivative of Sequoia’s original template. Hence this post’s shameful clickbait title—to put a crack in the monopoly of outdated advice that that one deck has over Google’s algorithms.

5. We didn’t even do a good job of that! Surely we could’ve done better than $60 billion. I recently saw a startup that sells scrubs and hospital gowns claim that their market is “all eight billion people around the world,” because, I assume, everyone might one day go to a hospital? Which, first, bold strategy, let’s see if it pays off for them. And second, why stop there? Ten billion people are going to be born over the next 75 years! And lots of them will be born in hospitals! VCs don’t want to back people swinging for singles and doubles, who are making products for only eight billion people. They want home run hitters, who are selling to 18 billion people!

6. In the data world, the people that they plan on making something for are almost always “people who use Excel.” If I had a dollar for every startup that initially thought their market was spreadsheet users, I’d have tens of dollars. If I lost five million dollars for every startup that initially thought their market was spreadsheet users, I’d be a venture capitalist.

7. If your two-step plan is, “We will build an initial product for some people and use that success to build a new product for other people,” that’s a bad plan. If your two-step plan is, “We will build a product for a tiny market with lots of qualifiers and use that success to slowly strip off those qualifiers,” that’s a better plan.