What is a startup?

“A startup is a company designed to grow fast”. 

A simple but powerful description of what a startup is meant to be.

It comes from Paul Graham, the legendary startup builder, investor and Co-founder of Y Combinator. He wrote about the purpose of startups in 2012 on his blog, in a post titled startup = growth

At Dealroom, where our bread and butter is discovering new companies, we’ve found this the most useful explanation of the term startup. But why do we need a definition for startups in the first place? What does “fast growth” even mean? And what’s the relationship between startups, tech and innovation?

Discover new startups
Startups ≠ tech

Startups and tech are often talked about interchangeably. Indeed, most startups are heavily tech enabled. But so are lots of other companies. In fact, many traditional SMEs and mature companies work more deeply on technology than startups do (e.g. industrial conglomerates).

When we talk about “tech startups”, what we really mean is information-age technology (internet, software, AI, algorithms) as David Galbraith writes. Even with these types of technologies though, “normal” companies are catching up. It’s been said that “every company is now a tech company”, or at least needs to be.

Paul Graham also makes this point:

“Being newly founded does not in itself make a company a startup. Nor is it necessary for a startup to work on technology, or take venture funding, or have some sort of “exit.” The only essential thing is growth. Everything else we associate with startups follows from growth.”

This framing makes it easier to understand why a company like WeWork was called a startup – it was a real estate company like many others, but it was built to grow fast. Or why Ocado is considered a startup, but Tesco grocery delivery isn’t – because Ocado is built to grow fast. Or why Tails.com is a startup – they are not the first company to make dog food, but… you get the idea.

Technology may enable a high-growth, scaleable business, but so might an innovative business model. Tech is not a prerequisite of a startup.

The same applies to being “innovative”. For example, was three-star restaurant El Bulli not innovative? Of course it was, but its innovation was applied to molecular gastronomy. Yet Deliveroo is a startup, because their innovation is focused not on the end product, but on growth. 

“I know a startup when I see one”

So how do you spot a company that’s designed to grow fast?

For many investors, the answer to that might be: “I know a startup when I see one”. 

It’s a fair point, startups vary so much that strict criteria would be impossible, indeed the best startups break the mould anyway. But that’s not super useful if, say, you’re asking governments to support startups. They like to know things like, how many startups there are, how many people they employ, how fast they’re growing, and their impact on the economy.

Well that’s where we come in. Dealroom combines machine learning, data engineering and unique partnerships to provide trusted data to dozens of governments, and help people find the world’s most promising companies. We also have a robust verification process, by which we mean “we know a startup when we see it”.

When does a startup stop being a startup?

We often get asked questions like “how many startups are there in Berlin today?” Seems simple enough. Maybe you would think to look at companies with a maximum age of 10. For Berlin though, that would mean Delivery Hero and Zalando are excluded, but N26 and HelloFresh are included. That seems kind of random. And what if a startup has gone public or been acquired? That would mean HelloFresh is out too. All these companies definitely have been startups, do they still count?

If the real goal is understanding how Berlin’s startup ecosystem has evolved to this point, a better way of looking at it is with time-series data. We might wonder how many startups are founded/funded each year, and how that has developed. Instead of a snapshot, we can gain more meaningful insights by comparing different cohorts (vintages) of startups: how do startups from the last five or 10 years compare with older ones? What percentage made the actual startup journey of fast growth?

This way, picking an arbitrary cut-off age is no longer needed. Of course a company doesn’t stay a startup forever, but for us to learn from them, all that matters is they were once.

How fast is fast growth?

The OECD says a high-growth firm is one growing +20% per annum. Now that percentage might be up for a bit of debate (more in a sec), but the trajectory of startup growth does tend to follow a similar pattern.

Startups will each have their own unique growth journey of course, but this is the kind of “S-curve” pattern people (founders, investors, governments) are looking for.

First there’s an initial phase of experimentation, where over the course of maybe a couple of years, a startup seeks product-market fit. Then once things (product, customers and distribution) click into place, high growth begins.

NB. deeptech and biotech startups have their own twist on the S-curve, but we’ve got a whole report dedicated to exploring that.

The year of deep tech report

Y Combinator, one of the most successful, and perhaps most growth-obsessed startup programs in the world, tend to talk in weekly growth rates at this point rather than annual ones. 

“A good growth rate during Y Combinator is 5-7% a week. If you can hit 10% a week you’re doing exceptionally well. If you can only manage 1%, it’s a sign you haven’t yet figured out what you’re doing.”

Whether measuring users or revenues, 5-7% growth per week works out as 13-34x growth per year! Most startups never achieve this level of hypergrowth and the ones that do cannot sustain it for very long. But Y Combinator have backed 43 unicorns including Airbnb, Dropbox, and Stripe so far, so they do know a thing or two.

Being this obsessed with growth is certainly one approach, but it’s not exactly T2D3* or bust. As Point Nine Capital Managing Partner Christoph Janz points out, for many not-Stripe-but-still-really-really-good startups, the journey might not be quite as neat.

* T2D3 = triple, triple, double, double, double, revenue year on year from Y2.

As Jason Lemkin points out, compound out 60% growth from $10M ARR, and you get to unicorn status pretty quick too.

So somewhere between OECD and Y Combinator level growth gets you into a reasonable high-growth startupy ballpark, and from there, anything can happen. 

There are many things that a startup is not. A startup is not any newly founded company. It’s not necessarily a tech company. Not a ping pong table, a hoodie or a culture. And definitely not “a state of mind”. 

There’s only one thing a startup is. It’s “a company designed to grow fast”.

Discover new startups