Three things I learned during 4+ years at Uber

I recently joined Greylock Partners as an Entrepreneur-in-Residence, where I’m exploring opportunities to start a new marketplace company. With the rise of the ‘uncollared’ worker and the growth of enabling technologies such as instant background checks (Checkr) and payments (Stripe/Adyen), we’ve only begun scratching the surface of these opportunities.

During my four years at Uber, I learned a lot about what it takes to build a massive marketplace. I wanted to share three somewhat counterintuitive aspects about marketplaces that are not always discussed, but are critical to get right.

1. May the supply be with you

Photo credit: Hero Complex

Most marketplaces are supply-constrained. In the early phases, you might struggle to generate demand but if a marketplace solves a pain point for consumers, demand usually outstrips supply pretty quickly.

Despite this fact, many marketplace companies tend to over-emphasize demand. As users, employees gravitate toward what they understand best, which is typically the demand side of the product. When was the last time you drove with Uber, hosted on Airbnb, or started a project on Kickstarter? This can create a gap in empathy regarding the supply side of the marketplace that must be consciously managed. Otherwise, you run the risk of slowing growth rates or even worse… creating an opportunity for a competitor that does a better job of tapping new sources of supply.

Once your marketplace starts to click, don’t be seduced by the dark side by skewing heavily towards demand… supply is the engine that fuels a great marketplace.

2. Don’t listen, show them

Steve Jobs, Apple founder

One of my favorite quotes about designing products is this from Steve Jobs:

“It’s really hard to design products by focus groups. A lot of times, people don’t know what they want until you show it to them.”

This is a truth for marketplace companies. When designing for multiple user bases, there’s one pool of “value” and it’s up to the company to decide how to divvy up the pie. This means when building marketplace products you have to be careful about listening to users, because what’s best for one side of the marketplace isn’t necessarily best for the overall health of the system.

A great example of this was how we implemented surge pricing at Uber. As recapped in this post, during the early days of Uber Boston, there just wasn’t enough supply (cars) to meet demand (riders) late night. Since last call is at 2am, there tends to be a huge spike in demand for rides between 1:30–2:30am. Furthermore, because of the late hour, many drivers headed home early. As an experiment, we decided to increase payouts to drivers (no changes occurred for riders) during late night and eureka!… drivers stayed out longer. Dynamic pricing makes Uber more efficient, allowing ETAs to stay short and drivers to complete more trips.

However, riders and the media have historically called for Uber to turn off or limit surge pricing:

3. Easy to play, hard to master