The other day, I saw an advertisement for an on-demand dog walking service. Not to offend the founders, investors, or fans of Rover the on-demand dog walking service, but the idea seemed preposterous to me. Does this count as innovation? Is Rover actually a technology startup? Do people realize that tacking an “on-demand” onto the front of a business model doesn’t necessarily make it better? I went to Google, typed “on-demand” into the search bar, and then had some fun going through the alphabet to see how many on-demand businesses I could find. On-demand accounting, art prints, barbershops, book printing, bike repair, car washing… it was a long list. After enjoying a healthy scoff, I acknowledged that I was being unfair and the issue deserved more consideration. Why wasn’t I taking Rover seriously? What qualifies a business as a startup? What separates the good startup ideas from the bad, and the great from the good?

In this post I borrow heavily from Paul Graham and Bill Gurley. They’re both incredible writers with an unparalleled understanding of the venture capital and startup ecosystem. Their deliberate style of thinking shines through in the quality of their ideas, and I’d recommend reading their essays if you’re interested in learning more.

What makes a business a startup?

There’s more to being a startup than having a name like “QuirkE”, a modern logo, and office space with inspirational quotes painted on the walls. More than 600,000 small businesses are started each year, but the majority of them wouldn’t be considered startups. In the simplest terms, there are three factors that make a company a startup: (1) it provides real value to consumers, (2) it has a large addressable market, and (3) it is able to deliver the real value to the addressable market incredibly quickly.

Hopefully every small business has factor (1); if they don’t, they won’t be around much longer. Good businesses will have either (1) and (2), or (1) and (3). It is very rare, however, to combine all three. The companies that beat the odds and manage to capture each of the ingredients will share a certain characteristic–they will all grow incredibly quickly.

Growth can be used as an identifier of sorts. Successful startups grow faster than a rabbit farm. They define, track, and obsess over growth, and focus all of their efforts on increasing their growth rate. So if a company is growing incredibly quickly, it’s probably a startup. However, growth should be treated as an indication of something more important. It should be the result of combining the three aforementioned factors, and should not be mistaken for the end goal. It’s an important distinction to make; we’ll discuss later how you can fake growth in an attempt to disguise a faulty foundation.

Just as growth is a result of our three factors, many of the characteristics that we associate with startups are a result of their focus on growth. Startups need to be agile in order to get products to market quickly, so they usually have small teams. Many are software companies because of the negligible costs of scaling. In an attempt to motivate employees and inspire the creativity needed for rapid growth, startups often have quirky office spaces. What’s important to remember is that these are tertiary identifiers, not core characteristics. If you see a company with beanbag chairs, ping pong tables, and free beer for breakfast, check and see how fast they’re growing. If they’re growing quickly, try to determine if they’re creating real value, if they have a large addressable market, and if they can quickly and consistently deliver the value to market. Only once they’ve demonstrated that they have our three factors should they be classified as a startup.

What separates the good from the bad, and the great from the good?

It’s my opinion that if you meet the qualifications of a startup, you’re already a good startup. In a sense, there is no such thing as a bad startup; you’re either good, or you’ve failed. Bad startups, if we can call them that, are companies that will eventually fail because they are gunning for growth without all three of the necessary factors.

Keeping things simple, we could say that great startups are ones that provide more value, to a larger market, faster. But if we want to dive deeper, we can identify a number of other characteristics that separate the best. Similar to our original three factors, these elite characteristics are more general principles than they are specific details. They form a framework that applies to companies of the past, and will continue to apply to ventures in the future.

This general framework is important because our idea of what a startup looks like will fluctuate wildly as time goes on. Identifying specific traits such as software platforms, B2B solutions, audio interfaces, or blockchain applications will undoubtedly prove shortsighted. If we want to identify the commonalities of successful startups–the factors that will hold true regardless of changing environments and improving technology–we need to avoid the temptation to simply list the characteristics of well-known products. We need to focus on ideas that result in sustained value creation.

Great startups do supposedly impossible things

Paul Graham says that startup ideas come to you when you operate on the fringe of the possible. When your interests and hobbies lead you to uncharted territory, when you become an expert in your field, when you question convention: then great ideas will start to come naturally. The best startups occur when one of these new discoveries creates a powerful wave in the marketplace, rippling through established products and services and changing the way that we see the world.

This characteristic is better thought of in terms of science or technology than in terms of design or innovation. Think SpaceX or Tesla rather than Airbnb or Facebook. These types of startup ideas are the most difficult (obviously: they’re problems that people think are impossible) but they also spark the imagination and inspire fiercely loyal customers.

So do stuff that’s supposed to be impossible. Blow people’s minds. That’s what truly great startups do. It makes your customers happy, it makes fundraising a heck of a lot easier, and it makes for great headlines.

Great startups have defensible positions

The best businesses have a defensible position, meaning it is difficult for new entrants to compete with them. This is classically referred to as a moat, and is not a new concept by any means. Moats have taken a number of forms over the years: control of scarce resources, such as precious metals or convenient waterways; established infrastructure, such as telephone lines or railroad tracks; brand domination, which can fend off superior products while giving companies time to respond; and exclusive technology that competitors can’t reproduce. Whatever the form, moats have always been important for defending the gains of a company.

Great startups also have moats, they just look different than the ones we’re used to seeing. A new moat that has been on the rise for the past few years–and that I think we’ll continue see frequently–is the network. A network of dedicated users may seem essentially the same as brand dominance, but networks are far more powerful. Each user, by participating in the network, is able to increase the value of the service for other users. Facebook, LinkedIn, and Uber are all examples of great startups using the power of networks to their benefit. The system of connections increases the value proposition of the company, and as the network grows the moat between the startup and the competition gets wider.

Networks don’t need to be based on consumer interaction, like Facebook or LinkedIn. I think that we will start to see more and more creative applications of the network moat that don’t have anything to do with users talking or connecting with each other. For example, a startup that collects data from consumers can use that network of data to refine their products and services. This in turn helps them attract more customers, which gives them access to more data, which allows them to further refine their offerings. A startup that operates in the IoT space would generate more value for each active device they had connected to the internet of things. Depending on their specific product, having more devices in the marketplace could mean providing consumers with more accurate weather predictions, better wifi, or a safer car ride. As the network grows, so does the value provided. It becomes less and less likely that a competitor will be able to successfully adopt the same strategy.

Great startups create new markets

This characteristic of a great startup–creating new markets–is closely related to the first, being unprecedented. If you’re doing something that was previously impossible, chances are you’re creating a new market. However, I think there’s enough of a distinction to have two separate sections.

The best way to describe the distinction is with an example. We’ll use a small, relatively obscure young firm called Uber to illustrate this idea of creating new markets.

When people talk about Uber, many mistakenly credit them with the first characteristic of a great startup: doing something that used to be impossible. They talk about Uber as if the company was the first with an on demand service, or as if hailing a car from your mobile device was some sort of unprecedented new technology. Uber didn’t do anything “on the fringe of the possible”, at least not in the sense mentioned earlier. The technology that they launched with was (relatively) straightforward and was used by any number of companies before them. What made Uber so wildly successful was their creation of a new market.

By changing the nature of the ride hailing industry, Uber blew the roof off of the market’s size. This idea of market creation becomes obvious when you think about what the ride hailing industry used to look like: taxis were only widely used in a few cities, and were often very expensive–the majority of cities had a limited number of cabs that consumers used infrequently. Uber didn’t just capture this preexisting market; that is, they didn’t just convince people that already used cabs to switch to Uber. Their improvements to pricing, payment, coverage-density, pick-up times, and accountability created a new market and changed the way that entire generations of people think about car ownership. It’s made them one of the highest valued startups in history.

I’d like to note that a fair argument could be made that Uber has the first characteristic of a great startup, and that applying old technologies in new ways is still doing something that was previously impossible. I think that this is a very compelling argument–and it may be a more accurate depiction–but for the sake of a clearer framework I decided to split the idea into two characteristics, each with a unique focus.

Profitability and the shortcomings of growth as an indicator

I mentioned earlier that startups could, in some cases, “fake growth” in order to hide a faulty foundation. When I say that companies are faking growth, I don’t mean that they aren’t actually acquiring more customers. I’m referring to the practice of spending money to grow, burning through cash to try to capture market share despite not being profitable.

The venture capital and fundraising ecosystem that surrounds startups is what makes this possible. With investors committing billions of dollars to chasing the next big startup, companies often have access to cash that they can use to acquire customers. And this isn’t necessarily a bad thing. As part of a larger strategy, burning through cash to acquire users early on can be very effective. It only becomes a problem if the startup isn’t actually providing enough value to customers for the company to ever become profitable.

So growth isn’t a perfect indicator for successful startups, in part because it doesn’t factor in profitability or even the potential for profitability. Many assume that if a company is growing quickly it must be providing real value to its customers, and that that real value can be monetized fairly easily. That can be a dangerous assumption to make, though, as the practice of spending investors’ money to acquire customers becomes more common. Growth should only be used as an initial indicator, and a deeper dive into the characteristics detailed above is necessary to determine whether a startup idea is good, great, or destined for failure.

What about Rover?

Is on-demand dog walking a good startup idea? It definitely deserves more consideration than I originally gave it. There’s quite a bit of positive overlap with Rover and the characteristics that we’ve reviewed above. Approximately 75 million dogs are owned in the United States alone, so it seems that they have a large addressable market. Their app forms a network of dog walkers and dog owners, creating added value and defensibility.

However, I still think that there are plenty of reasons to question the legitimacy of the model. My skepticism of on-demand dog walking stems from three things: the nature of the network they’re forming, their ability to create new markets, and the amount of real value they’re providing to users.

The growth of Rover’s network connecting dog walkers to dog owners runs into some geographical limitations. Networks that require in-person interaction thrive in dense cities, where there is a higher concentration of people and more opportunities for them to meet. However, in crowded, busy urban environments, people are less likely to own a dog. They don’t have the time, resources, or open space that they need for a dog. In the suburbs, where the ownership percentage is higher, people are more likely to have the resources and flexibility to take care of their pets without Rover’s help.

On-demand dog walking doesn’t create new markets. It’s highly unlikely that I will go out and buy a dozen more dogs now that I can easily get someone else to walk them for me. It’s similarly unlikely that the number of times I walk my dog would drastically increase. At best, Rover is hoping to capture most of the service market that is currently occupied by family, friends, and neighbors.

Finally, I’m skeptical of how much real value on-demand dog walking provides. It definitely provides some amount of value: I’m sure many people use and love the service. But does it provide enough real value to be profitable? Can they grow their user-base and monetize them enough to cover the costs of developing, staffing, advertising, and maintenance? The combination of limited real value and limited market growth gives me pause.

Having the external trappings of a successful startup doesn’t mean that you have a great idea. Heck, it might not even qualify you to be a startup. Founders and investors need to be careful when considering the merits of an idea, and when considering if businesses have the credentials to play in the startup space. Great ideas should be judged by their revolutionary nature, the defensibility of their position, and their ability to create new markets. Startup businesses should be classified as such only if they can provide real value to a large market quickly and efficiently.