Uber is facing a challenge with the increase in cancelled rides especially during peak hours. As product manager, how would you solve this situation?

Cancellation is being observed everywhere during peak hour (defined here as 09-11 AM and 06-08 PM) for all ride type with majority of cacellation reason being given as not listed.

Since the reason is not lised for most cancellation, it becomes open ended to see what could be the problem here.

Who are the major user of the product?

Office going people- for commuting to and fro

College student- not major audience, since they prefer cheap ride as bus. Fe opt for share and uber auto ride

Family outing- to go for restaurant and outings. not frequent rider

User Journey:

Open app

Select from and to destination

Check available ride, ETA and cost

Book a ride type

Communicate with driver (location guide, mode of payment, ETA etc.)

Pick the ride

Complete and review

If we see the journey funnel, the scope of cancellation happening is between ride type book to ride pick up and one major step here is communication with driver.

What could possibly go wrong here? Major pain point?

1. Driver generally serve mutiple competitor- An uber serving driver also serves other competitor or do leaisur activity. Once they get ride detail, they tend to compare it with competitor in terms of distance and money offered. A lesser deal leads to quick cancellation.

2. A far off/hectic ride in mroning/evening does not seem viable for drivers as they often end up in traffic rendering loss of money (more time, less ride, loss of money)

3. Distant pick up location in case of share ride often lead to cancellation as it's loss bearing. Far off location involves fuel burn, ton of traffic and less money

4. Driver not able to figure out location creates bad expereince for rider leading to cancellation at times.

5. A higher ETA for rider is mostly a negative sign and they tend to cancel before even starting communication with driver

6. A bad communication call/text between rider and driver often leads to cancellation

Now that we are aware of possible reason for cancellation, let's try to see which are the major factor contributing to cancellation and can be tactfully handled.

1st reason is the pressing issue. When a person cater to multiple competitor for same service, they have options to choose. just like rider has. In this case,a lucrative offer is always a deal breaker. A keen eye on hwo the pricing strategy need to be defined becomes viable

2nd issue can be addressed by offering nearby rides or rides near to final destination of driver (home) so that they can attend to the ride willingly without opting for cancellation. Simillarly, in morning, the driver can be suggested with option to choose the route to begin with

3rd issue can be handled by making most of gmap. Assigning share ride in route of a journey or final destination is always smart practice to earn multiple customer and make most profit out of it

4th issue is not a pressing issue as at most cases driver are able to find out location or rider help them with. A very bad case only leads to cancellation here.

5th issue is frequent occuring. A simple solution to this would be to select driver nearest to the location for assigning. Anything above 15 min should not be allocated and simply should show as "no ride available". As a rider, nobody is ready to wait for more than 15 mins and are likely to look for alternate option in different apps

6th issue is not frequent but also at times not avoidable. Having a police check on text/call would be good. A simple message saying your call is being recorded for quality purpose will put both driver and rider to choose words wisely while talking.

As a PM, I would start by implementing Issues which leads to most cancellation to be addressed first which would be 1st, 2nd, 3rd and 5th. 1st is long shot and would need ton fo study for revising pricing structure but is unavoidable considering the spiking no of cancellation.

Ideally, I would lay out a roadmap as well for 3,6 and 9 month here.

Metrics I would look at to measure if the problem is getting addressed or not:

0. No of ride request daily

1. No of cancellation daily

2. No of cancellation at peak hours (as defined above)

3.No of cancellation by rider

4. No of cancellation by driver

5. No of ride cancellation by different reasons listed

pavan roy · Answer 2 · 2022-01-28T09:53:38+0000

Ride cancellation = Ride requested + Cancelled

Clarifying Questions

Do you mean driver cancellations or rider cancellations? Both

Are we looking at shared rides also? Yes

Is there an increase in booking ride also? Yes

Is it sudden increase or a gradual one? Gradual

User Journey

a. Open the app.

b. Confirm your pickup spot and select your destination.

c. Choose a vehicle type.

d. Confirm your ride

e. Uber allocate the request to a nearby Uber driver

f. Uber driver accepts the ride

g. Uber driver pick up the passenger (if the driver couldn't meet the passenger, it will cancel the ride)

h. Uber driver arrives at the destination and mark the ride as completed.

i. Some passengers might leave a review and /or tip the driver.

Factors impacting this behaviour

Internal:

· Change in the way Cancellation is measured

· End of Marketing Campaigns

· was there a recent software update?

External:

· was there a recent announcement by us due to that reputation has changed

· a feature release by competitor recently

· New Competitor entered into market

· Is due to Covid pandemic in shared ride?

· Covid fear has ended in the area

· Interruption to the transportation during peak hours

· Seasonality

Hypothesis:

Change in the way Cancellation is measured

1. Uber has added limitation of wait time after which driver/ rider cancel without penalty or negative feedback.

Assumption: If without communicating, rider or driver has to wait more than 5 minutes, they can Cancel.

2. Actual wait time is longer than estimated wait time for Rider

3. Waiting time is high for driver

4. High volume of request of rides are coming

Solution

1. Uber has added limitation of wait time after which driver/ rider cancel without penalty or negative feedback: After 5-minute wait time, uber charge for waiting time.

Assumption: If without communicating, rider or driver has to wait more than 5 minutes, they can Cancel.

Rider: During peak hours, rider has not much time to wait and no lack of alternate transportation.so, they cancel the ride.

Driver: During Peak Hours, Driver get lot of rides request.so, they don’t want to wait much for a single ride.so, they cancel the ride.

For reducing the cancellation, Rider and driver should communicate each other with accurate wait time so that cancelling chance is lowered.

2. Actual wait time is longer than estimated wait time for Rider: In peak hours, due to live traffic, it is not easy for Prediction Model to adjust the accuracy of the wait time estimates for the arrival of vehicle. In case of unexpected delay, inform the rider and offer uber credits as a concession to remediate the negative customer experience which will help to lower the ride cancellations.

3. High Waiting time for Driver: During peak hours, it is not always possible for the driver to find a parking space where the driver can safely wait for a long time. For this assess the parking situation using the GPS and share that information with the rider so that rider can get to the car faster and cancellation can be lowered.

4. High Volume of request of rides are coming: Rider made request for ride but ride is cancelled by driver before ride request was confirmed. It is possible that drivers are cancelling ride request received more than usual in the window allotted for accepting and confirming ride request. For this, hire more car or tie up with other Rental Car services for Peak hour for providing the services so that Cancellation can be lowered.

Metrics

# Of unique passengers for the day

# of Active drivers

# of Ride requests

# of Rider cancellations

# of Driver Cancellations