How Uber and Airbnb understood their user's motivations to increase revenue

Today’s experiments come from Uber and Airbnb.

Experiment 1: Simplifying Pricing Display Increased Rides per User by 10%

Thanks to Zach for sharing these successful experiments with us.

Context

When the rider is about to book a ride, they want to know how much they would pay for the ride. In an effort to better manage unknown traffic, Uber experimented by showing a price range so Uber won’t lose money on any ride ($14-$20). While well-intentioned, this approach led to unintended consequences.

Problem

Psychologically, users tended to anchor on the higher end of the price range, feeling uncertain about the final cost. This uncertainty caused a decline in ride bookings, as users hesitated to commit to a potentially higher fare.

Solution

The product team pivoted from displaying a price range to showing a singular estimated price. This adjustment aimed to provide users with a clear and confident expectation of the ride cost, reducing decision paralysis and enhancing trust in the pricing.

Impact

Switching to a singular price estimate resulted in a double digit increase in rides per user. This change helped users feel more confident in their decision to book a ride, as it removed the ambiguity associated with the previous price range.

Learning

This experiment underscored the importance of user perception and trust. Providing a clear and singular price estimate helped users feel more secure in their purchase decision. The counterintuitive learning here is that sometimes more information (like a price range) can lead to greater uncertainty, whereas a single, clear piece of information can drive better user outcomes.

Experiment 2: Segmenting Users by Intent Improved Profits by millions for Uber Eats

Context

Uber Eats users fall under two personas. First, people who haven’t decided what they want to order and thus discover which restaurant to pick. Second group of users know exactly what they want to order and from which restaurant. 

Surge pricing on Uber Eats is a dynamic pricing strategy where the supply (drivers) and demand (eaters) had to be balanced or the end consumer would either see an increased delivery fee, long wait times, or a limited selection of restaurants. If demand significantly exceeded the number of available drivers, UberEats would temporarily hide restaurants from the browse and search experience to mitigate the demand. 

Problem

The challenge was to balance the availability of popular restaurants with the need to manage delivery logistics with long wait times. Initially, users who searched for a specific restaurant that was blocked would see an error message, which negatively impacted their experience.

Solution

The product team segmented users by their intent. High-intent users, who searched for specific restaurants, would still see these restaurants in the search results, even if they were blocked. For users who were browsing, only non-blocked restaurants were displayed. This approach tailored the visibility of restaurants based on user intent, enhancing the experience for both high-intent and exploratory users.

Impact

This targeted approach led to a single digit increase in profits, translating to millions added to the bottom line. The development cost was two sprints, but the payoff in user experience and retention was substantial.

Learning

The experiment revealed the importance of understanding user intent and personalizing the experience accordingly. A counterintuitive insight was that showing blocked restaurants to high-intent users, rather than hiding them, actually improved satisfaction and high orders. It demonstrated that acknowledging user preferences, even when limitations exist, can enhance trust and engagement, leading to significant financial gains.

Experiment 3: A/B Testing Email Campaigns for Airbnb to increase the # of referrals

Context

In early days, Airbnb aimed to boost the effectiveness of its referral program by optimizing the messaging in promotional emails. This experiment sought to determine which value proposition resonates best with Airbnb users.

Problem

Airbnb needed to identify the most compelling way to communicate the benefits of its referral program in promotional emails. Specifically, the goal was to understand whether users respond better to self-interested messaging or altruistic messaging.

Solution

Airbnb conducted an A/B test with two different value propositions for the same product:

1. Self-Interested Approach:

   • The email emphasized that the user can earn $25 for inviting a friend.

2. Altruistic Approach:

   • The email emphasized that the user is sharing $25 with their friend.

Impact

The altruistic email outperformed the self-interested email, demonstrating that users responded better to the message of sharing a benefit with a friend rather than just receiving a reward themselves. 

Learning

Messaging that highlights altruism (sharing a benefit with a friend) tends to perform better than self-interested messaging to increase #referrals. Source: growsurf