Andreessen Horowitz partner Chris Dixon recently wrote a piece on the concept of the full-stack startup. In it, he details how companies like Uber, Lyft, Munchery, Sprig, Shyp, Warby Parker and Airbnb are part of a trend in technology built on the notion that fixing broken industries by wrapping a software layer over incumbent services just isn't enough. Instead, to truly revolutionize a space, you need to reinvent the entire experience by building an end-to-end service enabled by software.

In reading Dixon’s piece and subsequent interview on the Andreessen blog, I was struck by how much of what he described mirrored my own experiences as a technology entrepreneur. 

My first startup, Taxi Magic, was the first company to use mobile technology to give users access to on-demand services in local markets. Instead of launching an end-to-end service the way that Lyft ultimately went on to do in 2012, we tried to improve the end user experience by adding a layer of technology on top of the taxi industry.

In its early days, Taxi Magic was a huge hit--it got incredible usage in many cities and was magical for consumers. However, at the end of the day, having the technology (Taxi Magic) and the operations (the taxi companies) as separate entities didn’t work. 

At the same time this was happening, Lyft’s founders were working on an early version of their service (then known as Zimrides), and initially they wanted to use the existing ground travel infrastructure to enable ride sharing between two people. (Indeed, they had approached us to see if they could use our black car partners in New York for this purpose.) In the course of that experience, it’s likely that they came to the conclusion that Dixon wrote about: only a transformation of the entire product could create the kind of magical customer experience they were after. 

The rest, as they say, is history. Lyft took off in San Francisco and soon, Uber followed suit with UberX. Both companies have turned the taxi industry on its head and have spawned a wave of full-stack startups that are transforming everything from eyewear to groceries. It's worth noting that not everyone agrees with Dixon in calling Uber and Lyft as full-stack companies because they don't actually own the cars. I side with him in classifying them as such because they control the user experience, set the pricing, determine the kinds of cars that drivers can use and can boot drivers who don't meet the minimum star rating.

My experience founding Taxi Magic, and watching the subsequent full-stack revolution replace the industry we were trying to fix, has been pivotal for me. It has greatly influenced the way we’ve built my current company, Shift, which is completely re-thinking the experience of car ownership, beginning with the highly inefficient way used cars are currently bought and sold. I am more convinced than ever that truly broken industries simply can’t be innovated from the inside. You have to bypass incumbents and build something completely new from the ground up. This is obviously much easier said than done, and I’ve learned a lot of lessons along the way. Below are the four I’ve found most valuable lessons that I’m currently employing at Shift to pursue a very different trajectory for growth and customer experience.

Hire a Rockstar Operations Team 

The prevailing wisdom in Silicon Valley is that building a great technology startup rests almost solely on the strength of your engineering team. The reality is that for a full-stack startup, technology and operations go hand in hand, and you have to make operations a priority from the get go in order to achieve the kind of amazing consumer experience that will help you succeed. Your operations team can’t just be sales people -- it has to be the heart of your business.

Invest in a Unified Consumer Experience 

While they may not get credit for it, the inventor of the full-stack approach is Apple. They control the entire value chain for their products, from retail to hardware to apps and branding. And of course, no one has mastered the consumer experience quite the same way that Apple has, with its incredible customer service and highly knowledgeable Genius Bar employees. From the moment you walk into an Apple Store, you know you’re going to get a premium experience from top to bottom. For all these reasons, full stack entrepreneurs would do well to take a page out of Apple’s book. Even though an incumbent might provide a subpar experience, it’s likely one that consumers are familiar with. Gaining awareness and changing consumer behavior can be extremely challenging, which is why it’s essential to have a wow-worthy experience that solves multiple customer pain points.

Test & Learn 

The single biggest lesson I learned at Taxi Magic is that you have to get your product in front of users as soon as possible, and then iterate based on their feedback. We did essentially the opposite. We had a very strong vision for the product and spent nearly a year and half perfecting it based on our assumptions about what users wanted. When we finally released it into the wild, we tried to launch into too many markets at once, and with way too many features and integrations. We built a sophisticated, fully fledged payment system (this was before the days of Braintree and other credit card processing platforms), because we thought that this was a critical part of the experience. Turns out that very few people actually used the app to pay for the ride. They were far more interested in getting real-time updates on where their taxi was. In hindsight, we should have released a bare bones version of the service early on, and then made incremental changes along the way, instead of trying to build it in a vacuum.

Empower Your Engineering Team 

The last thing companies want their engineers doing is dealing with meetings and internal debates, when they could be coding. Many startups rely on their product managers to manage their engineers. I take a different approach, one that is likely more controversial, but can lead to stronger products. It centers around the belief that great engineers have a vision for the product and don’t need excessive product management. If you have a strong operations team, the product manager role can be shared by engineers and ops folks alike. The best thing you can do for engineers is empower them, create an environment that fosters learning and leave them to do what they do best. 

There are many more industries in desperate need of change, and many proposed solutions, technology layers and services vying for customer love and mass adoption. The next wave of technology change will continue to come from building companies from ground up that rethink the entire user experience, not just some layer of it. This can be very hard, yet the payoff and pain is worth it, if you seek to become the next essential magical product. I feel lucky to have a chance to do this with an amazing team to the largest retail segment of our economy.