Growth Hacking Uber

The “U” stands for User. How Customer Optimization fuels Growth.

It started as a black car service in San Francisco in 2009, but has now expanded to more than 30 countries worldwide. As of August 2013, the company is valued at $3.76 Billion and received a significant investment of $258M from Google, which gives the company coveted access to Google’s technology — self-driven Ubers anyone? Groceries delivered by Uber?

As Uber investor and advisor Shervin Pishevar says, “Uber is building a digital mesh—a grid that goes over the cities,” Pishevar says. “Once you have that grid running, in everyone’s pockets, there is a lot of potential for what you can build as a platform.”

Affordable, safe transportation should be available to all. Think about areas of the world where it is unsafe for women to travel alone (see here) and imagine how an UberX could be a solution.

Many have written on how Uber has grown and I will not add to the literature here, instead I want to focus on how optimization can equal growth. Uber grows users in two ways: Customer acquisition and Customer optimization. I highly recommend you read Bill Gurley’s post on Conversion. It is important to note that Gurley is a General Partner at Benchmark Capital, and to date Benchmark is the largest investor in Uber. Bill has sat on its Board of Directors for the past 3 years.

Customer Acquisition

Acquisition can include such things as: AdWords, Facebook and a varied number of channels. User acquisition is what everyone assumes causes growth. This can be seen in the fetishization of the term “Growth Hacker” across the valley. One cannot talk to a start-up founder without them exclaiming they need a “Growth Hacker”, when in actuality they need to focus on product development. Don’t ask for fuel when there is nothing to start the fire.

Customer Optimization

Customer optimization on the other hand is a heads down focus on owned properties (mobile apps, website, landing pages). It is the ability to convert a user effectively in an app or website. It involves significant testing of each step of the user-flow and optimizing at each stage. This means that design decisions need to be made not because of clairvoyance of a designer, but in conjunction with analytical support provided by marketing or engineering. A hundred plus changes may be needed before a lift on conversion is achieved, but it is so worth it. Here is why.

Let’s say Uber employs two models: Model A and Model B. Model A is pure acquisition, for example, on Adwords alone. Model B relies on conversion optimization. Let’s say you were to improve conversion rate from 10% conversion to 11%, and in the chart below that’s a delta of 10% in new customers. Then lets say that it costs $1 to acquire a customer via AdWords acquisition. If you acquire 100 customers per 1,000 visitors, with acquisition alone I would spend $10 to acquire one customer. But with optimization, I would acquire 110 customers per 1000 visitors. Which means my costs to acquire a customer is down to $9.09 and for every dollar I spend, I will get to keep $2.20 versus $2.

example is illustrative only

With a user base like Uber’s, to state that the implications of the above analysis on revenue are significant, is an understatement. In a week in 2013, given the data available, Uber sees approximately 1.2 Million requests with 887K users completing the requests. Revenue for that week is approximately $22.1 Million dollars. If we increase conversion rate by a measly 1% ( ie. going from 74% to 75%) that’s an additional $312K in revenue per week and results in an additional $16.2 Million revenue for that year.

Customer Optimization Can Accelerate Uber’s Growth

Uber already has a great user acquisition model, but I would argue that there are some improvements to be had on the customer optimization side. In a industry which is increasingly crowded with multiple competitors (Lyft, SideCar, Flywheel) vying to own the market, every user matters.

There are two areas I looked at when thinking about customer optimization for Uber: the driver-to-customer experience and the in-app experience.

Driver-to-Customer Experience

First and foremost, I would look to see if there are any opportunities in customer on-boarding through service experience. Two that I have found cause friction because of a driver’s inability to arrive at the customer’s location seamlessly for pick-up and drop-off at the destination desired.

No matter how amazing a driver’s personality is, or how many perks (water, candy, gum, chargers) are offered — if the minimum bar of getting from point A to point B is not met, then the experience results in failure.

Recommendation 1: Don’t call a user when the established standard is text

Confirm with the user via text of user’s location. An automated text via Twilio that notes: “Confirming user’s location is at 355 Sansome St. Reply “yes” or “no”?” The driver can then send such a text if he is unclear as to the location of the customer.

Recommendation 2: On-board drivers with tools to increase efficiency

On board drivers to Waze. This Google owned app will ensure that drivers are aware of real-time traffic and road information. This will ensure the most efficient and time saving route is selected.

In-App Experience

Let’s first look at Uber’s customer on-boarding flow from download in the app store to app installation.

Uber’s User Login Screens,, Image Credit: Author

Recommendation 3: Use a Social Login

There are two paths to signing up for a Uber account. Either you want to use an existing login or you can create a new one. One way to optimize this process could be to use social logins such as Facebook or Twitter. However, there are reasons as to why you want to be the ultimate holder of User Login information, which using a third party system would be sub-optimal.

idea for social login

After a User downloads the app, the following image describes the user flow from ride request to completion.

Uber in-app user flow screens,, Image Credit: Author

Recommendation 4: Design better loading screens and progress bars

Optimizing conversion rates , Gurley notes, requires ruthless dedication to analytics at each stage of a user’s interaction within an app or website. Given that Uber is an analytical company, they probably collect data for in-app behavior. My hypothesis is that screens labeled with 1, 2, 3 and 4 cause the highest friction for a user.

Screens labeled 1 and 3 are areas where a user is kept waiting. No one likes to be kept waiting. If the loading bar on screens 1 and 3 were designed for effect, they need to be removed. If they in fact do denote time of backend data retrieval, then the design needs to be modified such that a user feels that they have not been waiting that long. It is possible to make progress bars feel faster to users by increasing the ribbing and pulsation of a progress bar. Noted reading here and here.

Recommendation 5: Use Google Maps to serve up map search results instead of Foursquare

The screen labeled 2 offers a significant opportunity for improvement. In this screen, a user’s intention is to search for their current location for pick-up. In the example screen, I searched for 290 Utah St. But the query results that were reported back were not even within 2 blocks. In fact, they were at least 5 miles away. A user, in frustration, may exit and drop out of the app and choose a competitor such as Lyft to request a ride. This is what screen labeled 2 looks like for Lyft and is executed more effectively.

Uber should not use Foursquare to serve up map results. Foursquare works if a customer is at a public and popular location, but works less optimally when the customer is at a friend’s house, or other less know places. Perhaps, switching to Google Maps to serve up results would be a better solution.

Recommendation 6: Analyze trends so that demand and supply are balanced

Finally, a source of pain for a potential customer occurs when a car cannot be found at a given point in time, due to an increase in demand and lack of availability. An example of this is shown on screen 4, where a text notes, “Check back soon, or try another type of car”. This can be remedied by looking at user request data over time and seeking trends by: time of day, day of the week and month of the year, proximity to holidays and local events. When a user gets texts denoting a lack of cars, the user will switch to a competitor to find a ride. If this happens enough times, then it will result in the loss of a customer.


Optimize for the user and growth will follow. It will be interesting to see which of these optimizations, if any, Uber will implement to further their growth strategies.

*I don’t work for or represent Uber. I’m a passionate user exploring ways to make things better. I love making good products great, and introducing great products to users. Currently, I’m working on growth marketing @Tradecraft until March 28th. ❤

Into Tech & high-impact, high-velocity marketing.