After my recent post about the startup escape path, a comment on HN complained complained about my overly narrow definition of the word "startup". This is kind of funny, because if anything, my definition of "startup" is much wider than the usual HN definition.
So, considering this site's title is "swombat.com on startups", let's explore this topic a bit. We'll start by covering some other people's definitions.
The naive definition
Before people get stuck into the tricky art of coming up with a scientific definition of what is a startup and what isn't, they have a simpler approach, dubbed "I know it when I see it".
By that definition, most people will agree (today, in late 2011) that Path, Instagram or Duedil are startups. They look, smell, behave like startups. Most people will also agree that IBM, Apple or Walmart are not startups.
The problem with this approach is it leaves an enormous grey area. Is 37signals a startup? Maybe, depending on your point of view. What about Twitter and Facebook? Probably. Google? Probably not. This is a grey area the size of the pacific ocean, if it fits companies all the way from FreeAgent to Twitter. In trying to resolve this grey area, people come up with all sorts of arbitrary distinctions, which just muddy things up.
"Startup" should have a relatively simple definition in one sentence. Especially "startup". Of all the words in the dictionary, this one certainly deserves a decent elevator pitch.
What about that dictionary?
The next possible solace is our good old friend the dictionary. No luck there, though. Most dictionaries (and spell-checkers) seem to regard "startup" as an alternate and possibly incorrect spelling of "start-up", and they have completely useless definitions like:
noun: the act or fact of starting something; a setting in motion.
Seems like startups literally need to throw away the dictionary. At least if they want a good definition.
The technological approach
"Startups are fresh, new, innovative companies trying to do cool stuff with technology."
This a common instinctive definition, trying to put the naive definition into words. It's terribly restrictive, however, and full of vague words that promote misunderstanding. Saying that startups are exclusively technological is missing the point, too. In many startups, the innovation may be supported by technology, but it's largely business innovation - and it's that innovation and the growth it brings which makes the business a startup, not the fact that it has a website built in Ruby on Rails. We can do better.
There have been a number of government-blessed efforts to promote "startups". However, their definition of "startup" tends to basically include any newly formed company. Whether you're creating a restaurant, a design consultancy, or the next Google, those guys think you're a startup.
I actually like this definition, because it's inclusive. It's arrogant to think that the only interesting new companies are high-tech disruptive startups, and much of the advice given to one group applies to the other as well. The skills for creating a successful brick-and-mortar business have a lot of overlap with those needed to create a successful online business.
The main problem with this over-inclusive definition seems to be the backlash it generates in the "traditional" startup community. When StartUp Britain launched, it faced a backlash. Soon after, I had a chat with Oli Barrett, one of the founders of StartUp Britain, and he confirmed that in his mind, "startups" included all sorts of non-technological businesses. The advice on sites like StartUp Britain is not geared towards startups as the average HN member understands them, but towards all new businesses.
People in tech circles feel, perhaps rightly, that this definition dilutes a word which they depend on. I agree that it's probably too expansive, though it does serve a purpose. Let's move on and look for another definition.
Steve Blank has perhaps one of the best definitions for startups as the tech community understands them:
A startup is an organization formed to search for a repeatable and scalable business model.
This is a good one. I like it because it doesn't focus on the technological element. Instead, it highlights the uncertain nature of startups. When you create a restaurant, you're pretty certain what your business model will be. When you create a startup, though, you have to be open to the idea that it will turn out very differently from what you initially thought (if you're not open to that idea, you will usually fail).
What about Nokia? I've argued before that Nokia is a startup, but in serving the goal of finding a good definition, I'm willing to recant. What about BlackBerry? They're in a similar situation, but somehow they don't feel like a startup. Does facing similar challenges make them startups? Not really.
Of course, you could argue that Nokia doesn't count as a startup by Steve Blank's definition, because it wasn't formed to search for a business model. It found it already, and then lost it.
This word is the crux of my problem with this definition. What about organisations which weren't explicitly formed with the goal of finding new business models, but which stumbled on one anyway, and grabbed the opportunity to grow.
37Signals, for example, was not formed to search for new business models. It started its life as a consultancy, and then developed a product, which then took off and enabled them to drop the consultancy work and focus on product work. As far as we know, 37signals' explicit goal was never to find a new business model, and it certainly wasn't formed for that purpose. And yet I'd argue it's definitely a startup. It looks like a profitable, successful, growing startup, by my internal compass.
The swombat.com definition
I propose the following definition:
A startup is a business which has ambitions and plans to grow by a large factor (10x or more) over the next few (1-5) years.
Let's break it down.
A startup is a business: As much as people talk about social enterprise and other nifty and charitable concepts, I believe a startup is a business, not a charity. If your aim isn't to make money with it, it ain't a startup. This rules out the slew of "check out my new startup, it's a plugin for Firefox" efforts which, while perhaps very useful and worthwhile, are not startups, just pet projects. Note that making money via selling the business itself is a valid business model by this definition.
which has ambitions: Beyond the need for discovery of business models, I think startups are primarily about ambition. If you want to build a chain of restaurants to take over the UK in the next few years, in my book you deserve the moniker of "startup". Building something like this business involves a lot of ambition and uncertainty. Snog deserves to be called a "food startup".
and plans: This is important, because if you ask most small business owners whether they'd like to grow by a factor of 10 in the next 3 years, most will say yes (though many will say no). But if you ask them how they're going to do it, most don't have a plan, or any idea how to get there - in fact, they probably think (rightly) it's not possible. Even the most uncertain startup has some kind of idea how they'll get all that growth. "We're going to build something awesome and get VC investment" is a plan, albeit very Silicon-Valley-centric. If there are no plans (correct or false), just a general wish or hope, it's not a startup.
to grow by a large factor (10x or more): A lot of arguments about "what is a startup" focus on "lifestyle businesses", whose aim is to provide income for a handful of people at most, vs "scalable businesses", which aim to grow. I would argue that although Tim Ferriss's "muses" and other similar businesses are very interesting, fun and worthwhile, they're not startups, and the same is true for other, less scientific efforts to build a small business. With no intention to grow, those are simply cool small businesses. The ambition to grow much larger than you are today is a defining characteristic of startups.
over the next few (1-5) years: of course, that ambition needs a timescale. If your ambition has no timescale attached to it, it's probably not a real ambition, just a dream.
Applying the recipe
So, with that in mind, let's revisit some of our earlier candidates.
Is 37signals a startup? Based on bits of available analysis, they seem to be growing their revenues by a factor of 2 or so every year, so in terms of actual growth they're within the 10x in 5 years. Was that their ambition? Yes, I think Jason Fried and DHH have had that ambition for quite some time. Their growth is definitely deliberate, and they probably were aiming even higher. So yes, 37signals is (or was) a startup. Whether they are one now would require knowing what their current ambitions are.
Are Twitter and Facebook startups? Maybe. They're still growing fast, and aiming to continue growing fast, to justify their lofty valuations. Certainly they were startups before. Their growth ambitions may still qualify them as startups.
Google? Google has grown nice and fast, but in 2005 their revenues were about $6b, and in 2011 it's looking like about $37b. I would say that Google long ago reached the scale where it's simply not possible to grow that fast anymore. The same is true for, say, Microsoft. At that scale, it's very hard to grow that fast. This rules out most mega-companies, like Nokia, HP or RIM, regardless of what uncertainties they may face.
What about GrantTree, my own productised service company? With the incredible insight I have into the mind of the founders of GrantTree (one of them is sitting in my chair right now!) I can tell you that our ambition is indeed to grow the company by about 20x in the next 5 years, and we have planned out how to achieve that, so I would call us a startup by the swombat.com definition.
I think this definition works relatively well for most situations - though you often require insight into the founders' minds to fully resolve it - but I welcome counter-examples. No doubt someone will post this on HN. If there are some good counterexamples there, I will link to them or discuss them in future posts.
Let me be the first to throw a spanner in the works, though. What about Apple?
Apple may not have done it in the 5 years timeframe, but their growth has been almost startup-like, from about $5b/year in 2003 to $50b/year in 2010 (and presumably even more in 2011) (source). Did they have the ambition to do this? You bet. Did they have plans? Assuredly.
I'll call Apple an outlier, and not a startup. Any definition is going to have some gap. I hope my definition has fewer gaps than others, but we'll see.
The swombat.com definition of "startup" is:
A startup is a business which has ambitions and plans to grow by a large factor (10x or more) over the next few (1-5) years.
Updates: These have been added on 21st December, based on the comments on HN.
The Eric Ries definition
A startup is a human institution designed to deliver a new product or service under conditions of extreme uncertainty.
I have the same qualms with it as with Steve's definition: the "designed to" part. I think a lot of companies stumble on the startup model after having been designed for something else, and redesign themselves along the way. Many others are simply not designed, and just flailing around for a way to survive, and stumble on a way to organise themselves that will suit that objective (or fail to). Both of those categories deserve the name "startup".
I think Eric's definition largely exists to serve the needs of his excellent Lean Startup methodology - i.e. it is more an answer to "who will find the Lean Startup methodology useful?" than "what is a startup?". There's nothing wrong with that, but I prefer my pragmatic definition for those reasons.
The Paul Graham definition
...essentially a startup is a new business designed for scale.
Here's another one:
Startups usually involve technology, so much so that the phrase "high-tech startup" is almost redundant. A startup is a small company that takes on a hard technical problem.
The first has the same problem as Eric and Steve's definitions. The second one, imho, was not a serious attempt at defining the generic term "startup", and so should not be criticised too harshly, but limiting startups to hard technical problems seems unfair.
I'd love to hear (and add) a considered definition by Paul Graham, though.
Many other definitions
Scaling is the key here. Startups are able to scale, whether at 5x, 10x, or 20x really doesn't matter. The point is that they can scale much faster than other types of businesses. This usually involves technology and automation, but not necessarily.
That's close to my definition, but with a wider range of scaling factors. I guess this definition would include Apple as a startup. I think it misses the "has the ambition" part too.
A startup is a 1-5 year old company that has a potential to grow up its enterprise value by 50X within the next 5 years and has a business that can scale up quite easily.
I've discussed the problems with focusing on the age of the company already. Any company that can realistically aim to grow 10x in less than 5 years can become a startup.
I'm not sure this part works: "to grow by a large factor (10x or more)".
Most businesses I come across that describe themselves as startups and fit the rest of the definition, haven't yet got themselves to a point where a 10x improvement over the next year would make them that significant. I'm specifically thinking of participants in accelerator programes like YC, TechStars, Start-Up Chile and SeedCamp.
Actually, I'd argue that those companies have a non-zero value already. YC typically takes 6% for $15k, or thereabouts, valuing the company at $250k. Pretty much every company YC takes on has the ambition to grow that by 10x in the next six months, let alone the next 5 years, which is why they do qualify as startups.
I like gghootch's take on it:
A startup is a business which aims to repeatedly grow by a large factor (5x or more) in the first few (1-5) years.
This does seem to improve on my definition, though again, why the focus on the "first few years"? Tweaked to say the "next few years", it seems like a great definition.
My personal definition of a startup is a company that is not yet a business. Which then brings up the definition of business. To me a business is a company that produces a stable profit, maybe 6-12 months profit.
Personally i think calling very well established businesses like 37signals a startup is quite insulting, as to me, its saying they're not successful, which they clearly are.
"A company that is not yet a business" seems too inclusive. Most home businesses are not really businesses - are they startups? I don't think so. Also, by this definition, RIM is a startup, since it's not profitable right now. As for 37signals being offended by the "startup" label, I don't see where in my definition of "startup" it suggests lack of success.
The key differentiator addressed by both Steve Blank and Eric Ries in their definitions, is that a startup is innovating. A company taking an existing, proven business model and attempting to build a business based on it (whether in the tech space or elsewhere), does not qualify as a startup, even if it is aiming for hyper growth.
By this definition, none of the SaaS startups out there qualify, since their business model is largely proven. I disagree with that.
And finally, a question from billpatrianakos:
Wouldn't you also agree that a true startup needs to have the requirements you listed but at the same time start from zero. Companies that just start out, have the growth you talk about in the time frame you talk about would definitely be startups. Companies that have been around for a while and suddenly have explosive growth and hit the other requirements only resemble startups.
No, I think a company could be plodding along for years or decades (for example as a consulting company) and then decide to try out something new. The age of the company is, imho, as irrelevant as the original intent of the company. What matters is their current ambition.
If you read this far, you should follow my RSS feed here.