Most Common Questions We Get About Shift (From Engineers)

When interviewing with Shift, we give you plenty of time throughout the process to ask questions. But to live up to our value of transparency, we wanted to get the most common ones answered honestly and publicly up front. Here they are:

The Mission

Wait, what is Shift again?

We’re the modern way to buy and sell used cars! Have you bought a car recently? It’s a pretty terrible experience: unfair, slimy, often discriminatory. At Shift, we aim to fix that. We think buying and selling a car should be convenient, transparent, and fair.

Why is this important?

Cars are a critical part of life for the vast majority of people in this country. They’re how we get to work, take our kids to school, buy groceries, take vacations, or earn a living. They can even feel like a secondary home (depending on your time in traffic). Access to transportation matters.

Buying a used car is financially smart and better for the environment than buying a new one. Unfortunately, the current process to buy is intimidating and unpleasant. We think it should be simple, easy, and enjoyable!

Ok, so how does Shift’s service work?

If you want to buy a car, you just browse our inventory on shift.com. We try to be as transparent as possible, giving customers the full out-the-door pricing with no negotiating. You can also view a free copy of the Carfax report, detailed photos of all the wear and tear on the car, and the results of the inspection we performed on the car – with details on everything we found and fixed.

When you find a car you like, you can schedule a free test drive and we’ll bring the car to you! You do a quick test drive with our non-pushy, non-commissioned Concierges. If you like the car, you can buy it on the spot! We offer transparent, competitive financing and additional warranties. We’ll help you complete all the paperwork in under 30 minutes from the comfort of your driveway using our iPad app. Keep the car and we’ll leave in a Lyft. Every car comes with a 5-day return policy and a 30-day powertrain warranty. Instead of spending your whole day at a dealership, you spend an hour with us from the comfort of your home.

Want to sell your car? It’s just as easy! Request an instant estimate on our website. If you like the estimate, schedule an evaluation and we’ll come out to you for brief inspection and provide a finalized offer based on your car’s condition. If you like the offer, we’ll initiate a payment direct to your bank account, collect and process the DMV paperwork, and leave with your car. Depending on your bank, your money arrives in your account in 3-5 business days. We also accept cars with loans or leases on them.

The Business

I don’t know much about the used car market, what’s the opportunity here?

Most of our current engineers didn’t know much about the used car market before they joined, either! But we’ve learned a lot. There are $750 billion worth of used cars sold in the US every year. That’s more than the total amount of smartphones sold globally ($450 billion). If the US automotive industry was its own economy, it would be the 15th largest in the world – and it’s incredibly fragmented. Carmax has a 2% market share and is the biggest used car dealership in the US. They’re worth $12 billion. The top 100 used vehicle dealers combined only have 7% market share.

What are the high-level plans for the company in the next year?

The next year is all about growth. We did about $140M in total sales in 2018 and we plan to double that in 2019. This will mean growing our existing markets as well as launching new ones. Currently, we serve the Bay Area, LA, and San Diego, which covers about 80% of people in California. But we’re opening our first non-California markets as of… now.

Five years?

Our goal is to become the platform for buying and selling cars. Used and new. Additionally, we aim to provide a great service for ongoing vehicle maintenance and to become the most reliable source for used car pricing and information. Also, maybe IPO.

IPO?! Tell me more about that…

Companies in this space tend to IPO earlier than other tech startups. That’s because it’s a very capital intensive business and the public markets are a cost-effective way to raise lots of capital. We have to buy cars, fund new markets launches, lease real estate, and hire engineers. Our closest competitor, Carvana, is 2 years older than us and IPO’d last year at about a $2B valuation with around $300M in total sales.

What about the competition?

The market is so fragmented that competition is a lot less relevant here than in many other businesses. It’s not a winner-take-all market. It comes down a lot more to execution and persistence than luck. Pretty much all our competitors could be successful and it wouldn’t really hurt us. If we can get to even 0.25% market share, we’d most likely be a deca-corn (a $10B+ valuation company).

But we still think about our competitors, at least as a benchmark, and they fall into three main camps: new startups like Carvana and Vroom, listing websites like cars.com, CarGurus, and Autotrader, and traditional dealerships, like CarMax and AutoNation.

For the first two groups, our big advantage is our operations. Online listing sites let you browse lots of cars, but customers still have to visit an old-school dealership to test drive and buy them. With Carvana and Vroom, customers can buy the car, but can’t get test drives. Customers have to buy it sight unseen and then the car is delivered it to you. User research (and common sense) shows that a vast majority of customers want (some might say need) to test drive a used car before they buy it.

Lastly, compared to traditional dealerships, our advantage is in our technology. Traditional dealerships typically rely on third-party software – which is an outdated strategy. Software gives us an incredible advantage in customer experience, efficiency, and scale.

So how do you make money?

Two primary ways. First, on the spread between the price we pay for cars and the price at which we sell cars. This is highly dependent on our ability to accurately and fairly price cars to match supply and demand. Most traditional dealerships rely on humans to do this because it’s a hard, messy problem. But as a technology company, we use software and machine learning to automate nearly all of it. This is a hard, technical challenge.

The second way we make money is by selling financing, warranties, and Gap insurance. These are useful add-on products that are often hard to get on used cars. A loan application can be messy and complex for a customer, so we’ve put a lot of work into making it easy and convenient. In the future, there are also many other potential revenue streams that our proprietary technology and data will enable, like licensing our platform or selling data products.

But isn’t it expensive to deliver test drives?

This was somewhat counter-intuitive to us at first, but it turns out, it’s actually cheaper. It costs less than half as much for us to serve a test drive as it does for a traditional dealership. That’s mainly because we don’t need to have expensive real estate/showrooms. Additionally, we serve a much higher volume of test drives much more efficiently with our sophisticated logistics software and better labor utilization. This logistics software is a fundamental competitive advantage, and a super hard technical problem. You can imagine the difficulty of efficiently coordinating thousands of cars, hundreds of employees with different schedules and capabilities across dozens of parking lots to offer at-home on-demand test drives.

The Engineering Team

Alright, let’s talk about the engineering team.

Let’s! At the end of the day, we’re a technology company. Our business doesn’t exist without technology. We’ve put together a world-class engineering team, with top talent from all over. Currently, we’re about 40 engineers, but we’re looking to double that in the next year.

What’s the engineering process like?

Each team has the cross-functional members they need to own a problem area and deliver on it end-to-end. Each team has its own product manager, designer, engineering manager, tech lead, and a team of engineers. We believe in pushing decision making down as far as possible, so each team is responsible for coming up with their own roadmap, getting buy-in from leadership, and delivering on it. This means individual engineers have a lot of ownership, autonomy, and impact! The startup trifecta!

We loosely follow a sprint methodology, but our work is largely focused around delivering high-impact projects. We know startups are all about trying things, learning, and iterating, and as we double in size our processes will have to scale as well. Every major project built by our engineering team goes through a design doc and architecture review process, and each feature is thoroughly code reviewed. We use quarterly OKRs for longer-term planning and Phabricator for the day-to-day.

What are the teams?

Since our teams are primarily cross-functional, most of our teams are full-stack product teams. This means there are interesting frontend and backend problems on nearly all of our teams. Teams are focused around problem areas, and in pursuit of solving those problems and creating a better customer experience they will span the full stack and use whatever technologies are necessary. Currently, since we’re growing so fast, we re-evaluate our teams once or twice a year to determine if we need to create new, increasingly focused teams. Here’s the current list, as of December 2018:

  • Seller: Making our seller experience great and acquiring a bunch of cars.
  • Local: Owns the tools that our field teams use, our complex scheduling systems for appointments, as well as the iPad app for completing tasks in our warehouse and with customers.
  • Shopping: Owns the top-of-the-funnel experience for someone who first visits Shift and is shopping for a car. The goal: to create a shopping experience for cars that is as educational and engaging as Amazon’s is for…everything else.
  • Appointments: Helps customers schedule, prepare for, and complete test drives and pickups.
  • Purchase: Product and tools team that’s responsible for checkout, including paperwork, financing, and warranty integrations. Also, customer support team tools.
  • Pricing: The team that’s responsible for pricing cars, managing the prices of cars in inventory, data infrastructure, and the tools we use to acquire auction cars (about 10% of our inventory).
  • Infrastructure: Backend and frontend infrastructure team that’s responsible for site reliability, DevOps tooling, and developer productivity.

What technical challenges are there?

We’ve got a ton! A lot of the challenge stems from the inherent complexity of buying and selling cars and there’s too many to get into all of them here, but here’s three examples:

  • Pricing: How do you accurately price assets that are all unique with messy, incomplete data and then explain to customers why the price is fair?
  • Appointments: How do you efficiently offer appointment slots to customers to test drive cars at their home when the cars are scattered across dozens of parking lots with variable traffic and staffing?
  • Paperwork: How do you complete loan applications and sale paperwork in a customer’s driveway in under 30 minutes, when it takes three hours and a half dozen specialized people to do it manually in a traditional dealership?

What’s the tech stack?

We use Go on the back end, running on ECS with PostgreSQL as our main database. On the frontend, we use Typescript, React, and Redux. And on mobile, Swift. For a company our size, we have pretty robust developer tooling, deploy tools, continuous integration, and production infrastructure.

What’s the culture like?

Friendly, helpful, transparent, and pragmatic! We believe in fostering an environment where everyone feels safe to take risks and be vulnerable (psychological safety). We take the time to listen, learn and teach. We understand that engineering is about trade-offs, so we like to identify the ideal solution, but balance that with the most practical one. We strive to have solid tooling and stable systems, to take time to clean up tech debt, and to leave code better than we found it. As with our product, we believe in transparency on the team. We have open communication from leadership to the team with monthly town halls, weekly AMAs, and open sharing of sensitive information, like company financials and fundraising. We believe in being thoughtful and introspective about who we are and how we work.

This isn’t limited to engineering. We’ve also got amazing design, PM, data science, and business operations teams with whom we work super closely! We’re a small but tight-knit group with a wide range of backgrounds and ages. We’ve got a strong contingent of new grads, but also a bunch of experienced engineers with families. We have a number of clubs, including board games, surfing, running, and biking! We like to eat lunch together and do escape rooms. Everyone on the team is invested in making this a great place to work for people from all backgrounds and at all stages in life.

Wow, sounds great! What’s the catch?

No catch! Well, our biggest challenge is that we don’t have enough engineers to do all the things we need to do! But you can help us change that — we’re hiring!

How I Used Machine Learning To Price Used Cars

Andrew Carman is a software engineer at Shift.

At the core of any market is pricing. If you get pricing wrong, your marketplace doesn’t work. Our success at Shift—as a marketplace that makes it easy for anyone to buy and sell used cars—hinges on accurately and competitively pricing cars. 

Dealerships traditionally have a team of experts that price trade-ins for customers and manage the prices of their vehicles. They haggle over those prices with buyers that come to their dealerships, and often take advantage of consumers who have less access to car price data.

Our goal is to make selling and buying a car a fun, fair, and accessible experience by using technology to disintermediate what car dealerships do poorly. So we build software instead of back-office sales teams to price our cars, which increases efficiency, levels the playing field for our customers, and lets us systematically improve the accuracy of our pricing over time.

Read on to see exactly how we did it, and pass this along to the folks who might be interested in our approach.

Continue reading “How I Used Machine Learning To Price Used Cars”