How dogfooding their own product helped Uber Health increase ride completion by 20%?
Context: These experiments were conducted in the context of Uber Health within Uber for Business. Uber Health serves healthcare organizations and needs to facilitate better patient health outcomes by reducing the barrier to care by providing rides to and from appointments. The main KPI for the product was ride completion rate, which was disproportionately challenging because the product targets a non-tech-savvy audience. Kavan and his team dogfooded the product from all angles by having different team members call rides, go to healthcare facilities, and experience the challenge of completing a successful ride themselves. This led to multiple great experiments based on the challenges they faced.
Experiment #1: Automated Phone Calls for Improved Patient Engagement increased ride completion rates by 15%
Problem: In the healthcare segment, the completed-to-request ratio was drastically lower (~65%) as compared to the consumer segment (>90%). Part of the problem was the ride was requested by the hospital/healthcare provider on behalf of the patient as compared to directly by the patient like how consumers order Uber today. This caused communication challenges and given the patients tended to be older generation (and thus non-tech savvy), texting isn’t their primary mode of communication like other generations.
Experiment: In this experiment, an automated phone call was introduced to patients when the driver was 2 minutes away from the pick-up location. This made it less likely for patients to miss the notification compared to a text message.
Impact: The experiment led to a 15% relative increase in ride completion rates, and it required only 1-2 days of engineering work.
Learnings: The experiment highlighted the importance of adapting to the needs of a less tech-savvy audience in healthcare, where phone calls were more effective than text messages.
Experiment #2: Enhanced Mapping for Precise Location Information for healthcare facilities increased ride completion by 20%
Problem: Uber Health product team dogfooded their product at hospital facilities to realize the challenges of locating and meeting drivers in parking lots and drop off/pick up locations. The generic maps provider didn’t cover specific pickup points within these areas. This experiment was carried out to improve mapping precision and location information within Uber Health.
Experiment: The experiment focused on launching core mapping enhancements to convey precise GPS information in the text messages sent to patients. This improvement aimed to help patients connect with the driver more effectively. Ex: rather than just having the big hospital facility as the destination, it had specific pick up location within the big facility.
Impact: The experiment led to a 20% increase in the completed-to-request (C/R) ratio, indicating a significant improvement in patient engagement and ride completion rates. The development cost for this experiment was 4 weeks (as it was integrating a technology that another team at Uber had built).
Learnings: Providing patients with exact GPS information in the text message improved their ability to connect with the driver, resulting in a higher C/R ratio.
Experiment #3: Driver-Passenger Communication Enhancement
Introduction: This experiment addressed the need for improved communication between drivers and passengers, particularly when passengers required additional assistance ex: older age patients.
Experiment: In this experiment, Uber implemented a feature where the driver received a message indicating that the passenger might require extra help and instructing them to call the passenger at a specified number. This enhanced communication by the driver to call the patient improved patient satisfaction and willingness to wait for the passenger as they talked to them on the phone. This made the experience more human and made them less likely to cancel/disappoint them by canceling the ride on either side.
Impact: The experiment resulted in a 10% increase in the completed-to-request (C/R) ratio, demonstrating the positive impact of improved human connection between the driver and the patient. The development cost for this experiment was 1 week.
These experiments were designed to address specific challenges and improve patient engagement, ride completion rates, and overall user experience within the Uber Health segment.
If you enjoy GrowthInsider's Newsletter, it would mean the world to us if you can share it.
If you refer 3 people, you will get access to our PM course about frameworks to communicate and solve problems.
If you refer 5 people, you will be invited to our private PM webinar where we share recent successful experiments.
If you refer 10 people, we will offer a private, 1-hour growth consultation session for your product.
May this year be your best year yet and have fun experimenting 🚀