Building a Unicorn

Why creating "Uber" could cost you $1MM--but doesn't have to.

A question I get a lot from bright-eyed, aspiring entrepreneurs is, “How much would it cost to build X,” where X is always a unicorn startup. I both love and dread this question, mostly because it makes a lot of assumptions, and also forces me to make a lot of assumptions.

As an avid Uber customer, I spend a lot of time in the back seat of an UberX. Being friendly, I often engage with drivers who inevitably ask what I do for a living. When I tell them I am a consultant who helps entrepreneurs launch successful lean startups, most of them ask for a quote for their idea for the next Facebook, Uber, or other unicorn startup. However, sometimes they have more specific questions about what it takes to launch a product.

A few weeks ago, I had a very long conversation with my Uber driver who said, “My friend told me it would cost a million dollars to make the Uber app. Is this true?”

Explaining how real startups are created.

The first step to answering this question is to explain Uber's story to the casual observer. Many people look at a company like Uber, see their $65 billion valuation, and think that it is just an app. The general confusion comes from the highly publicized early days of the Apple App Store. We all remember the 15-year-old from Iowa who made $100,000 and paid for college with his $0.99 fart jokes app.

While Apple did the work of bringing the 15-year-old entrepreneur customers and handling distribution, he was one of the rare lucky ones. The more seasoned entrepreneur knows that big ideas like Uber take big money and a visionary team of product designers, technology architects, and marketing gurus.

Starting small and simple

An important point to make is that it could cost hundreds of millions to make Uber the company, but it could cost less than $50,000 to make Uber the app as it exists today. One of the very important differences to keep in mind is the scale at which Uber operates. The more customers you have, the more you have to worry about infrastructural scaling for people and architecture.

But Uber didn’t always work at this scale–keep in mind, they started with an SMS based product in San Francisco only. The product that you know of today is backed by an unprecedented $9 billion in funding. That’s right–billion. And while the growth and funding happened relatively quickly, keep in mind that it didn’t happen overnight.

Think of what the smallest minimally viable product that could compete in the same space as Uber might be: A radio dispatch company. You could call a phone number, they would find a driver via radio, and then call you back to confirm the ride.

Obviously in today's landscape, you'd have to offer something really unique to pry them away from the instant gratification Uber provides. But in the early days of Uber, the company did almost exactly that. The area of disruption Uber focused on was allowing you to book a black car via text message, rather than having to call a car service.

In other words, they set out to solve the two major problems standing between the average Joe and private car services: content discovery and accessibility. If I had 15 minutes to get a car in 2008, I wouldn't even think to hire a black car. For starters you'd actually need to know the name and number of a driver or dispatch company, and then you'd need to hope they were conveniently available or near you. It was previously really something people would do a couple of hours in advance. In short, the best part of Uber’s model was pairing users with available drivers close to them.

Concierge as a Service

In 2009, you'd text Uber an intersection, and they'd text you back the information about your driver and when they could pick you up. (Imagine a $65 billion company starting with an SMS plan!) For a reasonably small enough user base, it is something one guy with a prepaid phone and Google could run. Here’s how it works:

Text comes in, look at a list of black car provider's, call each one to schedule a pickup, then text the customer back.

Being a concierge is a cheap way to validate a business model. You can start to gain a customer base with a tight feedback loop, allowing you to validate your ideas before spending money on technology no one wants. It will also humanize your product to your early adopters.

Scale slowly and then accelerate over multiple iterations

One problem first-time entrepreneurs face is the million user problem. They all seem to think they need to build a solution that scales to a million users at launch. This often leads them to building over-engineered solutions that solve problems they don't yet have. Hiring a bigger labor force and brute forcing it for the first couple of years is a great way to buy scale. Labor is also much better at reflecting on and fixing issues itself than technology is.

“Should our Management dashboard have the ability to refund, or can we have someone issue refunds manually via Stripe's dashboard?” It’s an additional step, and you have to train your service team, but if you only give out 1 refund a month, did you really need the feature to be easy?

The fear of failure

One of the reasons people build overly complicated systems with every feature they can imagine is so that they can blame their technology for their failure instead of themselves. If they say, “We will be responsible for X to keep technology costs lower,” and the product fails and people point to X, then they failed. If they can shuffle the blame to the system doing a bad job at X, or even that they built X so they don't know why the business failed, they feel validated that they did their best.

Worse than this is investor pressure. If you think your investors will sue you for not building every feature in your deck, then you either put together a very misleading deck or have no idea what you are doing. Investors should be investing in you, not features you promised. A failed startup will always be your fault, not your products'. You are your product and your product is you.

Check out a recent post on (Failing Fast)[].

Companies, not products, become unicorns.

Success comes from having the right team and meeting your sales goals. It’s important to remember that products are just a way to connect with customers. It is merely anecdotal that, in a world with so many entrepreneurs and companies, users gravitate to companies with the best product. Following this thinking, we can derive that our most successful salesman is our product. Let your product sell you. Using this thinking, we can shift the paradigm when thinking about what product needs to be.

The focus of a startup should be on creating a User Experience that people love, not one that makes your life easier. Put an emphasis on bigger picture items like growth and marketing. Don't micromanage your app for months trying to get it perfect. Put it in the hands of users fast, get feedback, and iterate. And keep in mind how big the market is–bad feedback early is better than bad feedback 6 months later.


There is a time and a place for building every feature, having a 1000 person team, and spending millions of dollars on your app. Ideally these things should happen after your product is getting traction and earning positive reviews. If you are getting good traction and feel like there is huge market potential, but you don't have enough cash to grow fast enough to keep up with market, then it might be a good time to take on investment. This is how you spend a million dollars building a company like Uber, but it should come after hustling your (MVP)[] into a profitable small business.


MVPs should prove a few things: They should prove the market's actual need and response to your idea, but they should also prove if you are any good at running a company that solves X. You might have the best idea for a product, but if you can't make it work at a small scale with a minimal scope, then it might not be the right business for you. Start small and save yourself and your investors a ton of cash.

And as always, let us know if you need some help!